Computing systems including hardware, software, communication, and networks are becoming increasingly large and heterogeneous. In short, they have become in- creasingly complex. Such complexity is getting even more critical with the ubiquitous permeation of embedded devices and other pervasive systems. To cope with the growing and ubiquitous complexity, autonomic computing (AC) focuses on self-manageable computing and communication systems that exhibit self-awareness, self-configuration, self-optimization, self-healing, self-protection and other self-* properties to the maximum extent possible without human intervention or guidance. Organic computing (OC) additionally addresses adaptability, robustness, and con- trolled emergence as well as nature-inspired concepts for self-organization.

As the computing and communication infrastructure continues to expand and diversify, the need for adaptability in software is growing. Adaptability is especially important to pervasive computing, which promises anywhere, anytime access to data and computing resources. The need for adaptation in pervasive computing applications is particularly evi- dent at the “wireless edge” of the Internet, where software in mobile devices must balance several conflicting concerns, including quality-of-service, security, fault-tolerance, and en- ergy consumption. We say that an application is adaptable if it can change its behavior dynamically (at run time). Developing and maintaining adaptable software are nontrivial tasks, however. Even more challenging is to enhance existing programs so that they execute effectively in new, dynamic environments.

The need for adaptability in software is growing, driven in part by the emergence of pervasive and autonomic computing. In many cases, it is desirable to enhance existing programs with adaptive behavior, enabling them to execute effectively in dynamic environments. In this paper, we propose a general programming model called transparent shaping to enable dynamic adaptation in existing programs. We describe an approach to implementing transparent shaping that combines four key software development techniques: aspect-oriented programming to real- ize separation of concerns at development time, behavioral reflection to support software reconfiguration at run time, component-based design to facilitate inde- pendent development and deployment of adaptive code, and adaptive middleware to encapsulate the adaptive functionality. After presenting the general model, we discuss two specific realizations of transparent shaping that we have developed and used to create adaptable applications from existing programs.

Keywords: Transparent Shaping, Software Adaptation, Dynamic Reflection, Mobile Computing.

The fields of Grid, Utility and Cloud Computing have a set of common objectives in harnessing shared resources to optimally meet a great variety of demands costeffectively and in a timely manner. Since Grid Computing started its technological journey about a decade earlier than Cloud Computing, the Cloud can benefit from the technologies and experience of the Grid in building an infrastructure for distributed computing. Our comparison of Grid and Cloud starts with their basic characteristics and interaction models with clients, resource consumers and providers. Then the similarities and differences in architectural layers and key usage patterns are examined. This is followed by an in depth look at the technologies and best practices that have applicability from Grid to Cloud computing, including scheduling, service orientation, security, data management, monitoring, interoperability, simulation and autonomic support. Finally, we offer insights on how these techniques will help solve the current challenges faced by Cloud computing.

Keywords: Cloud Computing and Grid Computing.

Web services are gaining acceptance as the predominant standards-based approach to building open distributed systems. Business Process Execution Language (BPEL) allows for the composition of existing Web services to create higher-level Web services. There is a need to deliver reliable service compositions with precise Quality of Service (QoS) attributes covering functional correctness, performance and dependability, especially since current BPEL standard provides limited constructs for specifying exceptional behavior and recovery actions. In this chapter, we present a language-based approach to transparently adapting BPEL processes to improve reliability. This approach addresses reliability at the business process layer (i.e., the language layer) using a code generator, which weaves fault-tolerant code to the original code and an external proxy. We also explain the software patterns present in this approach. These patterns constitute abstract reusable concepts that will facilitate rapid model-driven development of adaptive service compositions that can be easily configured for a range of situations.

Keywords: Web service composition, BPEL, Concept Reuse, Design Patterns, Adaptability, Dependability, Business Process.

In this chapter, we propose a distributed multimedia data management architecture, which can efficiently store and retrieve multimedia data across several nodes of a Grid environment. The main components of the proposed system comprises of a distributed multidimensional index structure, a distributed query manager handling content-based information retrievals and a load balancing technology. The proposed distributed query manager embeds the high-level semantic relationships among the multimedia data objects into the k-NN based similarity search, thus bridging the semantic gap and increasing the relevance of query results manifold. This research has two major usabilities. First, it models a web environment where each node of the Grid can be considered as the nodes or sources of data in the world-wide-web. This should help to investigate and understand the challenges and requirements of future search paradigms based on content of multimedia data rather than on text annotations, as used currently. Second, it provides the foundation to develop content-based information retrievals as a possible Grid service. Extensive experiments were conducted with varied data sizes and different number of distribution nodes. Encouraging results are obtained that makes this endeavor a potential architecture to manage complex multimedia data over a distributed environment.

Keywords: Grid Computing, Cloud Computing, Distributed Multimedia, and Data Management.

Abstract. The Latin American Grid (LA Grid) joint research program fosters collaborative research across eleven universities and IBM Research with the objective of developing innovative grid technologies and applying them to solve challenging problems in the application areas of bioinformatics and hurricane mitigation. This paper describes some of these innovative technologies, such as the support for transparent to the application expert grid enablement, meta-scheduling, job flows, data integration, and custom visualization, and shows how these technologies will be leveraged in the LA Grid infrastructure to provide solutions to pharmagenomics problems and hurricane prediction ensemble simulations.

Keywords: Meta-scheduling, job flows, data integration, transparent grid enablement, custom visualization, bioinformatics, hurricane mitigation

The ever-increasing complexity of computing systems has been accompanied by an increase in the complexity of their management. Contributing factors include the increasing size of individual networks and the dramatic growth of the Internet, the increasing heterogeneity of software and hardware components, the deployment of new networking technologies, the need for mobile access to enterprise data, and the emergence of pervasive computing. In this chapter, we focus on the management complexity resulting from integrating existing, heterogeneous systems to support corporate-wide, as well as Internet-wide, connectivity of users, employees, and applications.

Keywords: Grid Computing, Interoperability, Meta-Schedulers.

Keywords: Clould Computing, Federation, Interoperability.

This paper proposes and describes a methodology developed to port complex scientific applications originally written in FORTRAN to nVidia CUDA. The significance of this lies in the fact that, despite the performance improvement and programmer-friendliness provided by CUDA, it presently lacks support for FORTRAN. The methodology described in this paper addresses this problem using a multiple step process that includes identification of software modules that benefit from being ported, familiarization with the code, porting, optimizing, and verifying the ported code. It was developed and carried out by porting an existing module of a weather forecasting application written in FORTRAN. Using this approach, we obtained a functional prototype of the ported module in approximately 3 months, despite our lack of knowledge of the theory of the weather code. Considering the relevance of this application to other scientific applications also written in FORTRAN, we believe that the proposed porting methodology described can be successfully utilized in several other existing scientific applications.

Keywords: gpu, programming, cuda, weather modeling

Grid applications composed of multiple, distributed jobs are common areas for applying Web-scale workflows. Workflows over grid infrastructures are inherently complicated due to the need to both functionally assure the entire process and coordinate the underlying tasks. Often, these applications are long-running, and fault tolerance becomes a significant concern. Transparency is a vital aspect to understanding fault tolerance in these environments.

Keywords: Distributed Workflows, Grids, Managing Faults.

The development of collaborative multimedia applications today follows a vertical development approach, where each application is built on top of low-level network abstractions such as the socket interface. This stovepipe development process is a major inhibitor that drives up the cost of development and slows down the innovation pace of new generations of communication applications. In this paper, we propose a network communication broker (NCB) that provides a unified higher-level abstraction for the class of multimedia collaborative applications. We demonstrate how NCB encapsulates the complexity of network-level communication control and media delivery, and expedites the development of applications with various communication logics. We investigate the minimum necessary requirements for the NCB abstraction. We identify that the concept of user-level sessions involving multiple parties and multiple media, is critical to designing a reusable NCB to facilitate next-generation multimedia communications. Furthermore, the internal design of NCB decouples the user-level sessions from network-level sessions, so that the NCB framework can accommodate heterogeneous networks, and applications can be easily ported to new network environments. In addition, we demonstrate how the extensible and self-managing design of NCB supports dynamic adaptation in response to changes in network conditions and user requirements.

Keywords: CVM, NCB, User-Centric, Network Communication Broker, and Multimedia Collaborative Computing.

The increasing demand for high performance computing resources has led to new forms of collaboration of distributed systems, such as grid computing systems. Moreover, the need for interoperability among different grid systems through the use of common protocols and standards has become increased in the last few years. In this paper we describe and evaluate scheduling techniques for multiple grid scenarios. In particular, they consist on the proposed “bestBrokerRank” broker selection policy and two different variants. The first one uses the resource information in aggregated forms as input, and the second one also uses the brokers average bounded slowdown as a dynamic performance metric. From our evaluations performed with simulation tools, we state that, although the aggregation algorithms lose resource information accuracy, the broker selection policies using aggregated resource data do not penalize their performance significantly. Moreover, we show that the best performance results are obtained with the coordinated policy using dynamic performance information, in addition to aggregated resource information. Therefore, we conclude that delegating part of the scheduling responsibilities to the underlying scheduling layers promotes separation of concerns and is a good way to balance the performance among the different grid systems.

Keywords: Grid Computing, Broker Selection, Job Scheduling, Resource Management, Resource Data Aggregation.

Increasingly, software systems are constructed by integrating and composing multiple existing applications. The resulting complexity increases the need for self-management of the system. However, adding autonomic behavior to composite systems is difficult, especially when the constituent components are heterogeneous and they were not originally designed to support such interactions. Moreover, entangling the code for self-management with the code for the business logic of the original applications may actually increase the complexity of the systems, counter to the desired goal. In this paper, we address autonomization of composite systems that use CORBA, one of the first widely used middleware platforms introduced more than 17 years ago that is still commonly used in numerous systems. We propose a model, called Adaptive CORBA Template (ACT), that enables autonomic behavior to be added to CORBAapplications automatically and transparently, that is, without requiring any modifications to the code implementing the business logic of the original applications. To do so, ACT uses “generic” interceptors, which are added to CORBA applications at startup time and enable autonomic behavior to be introduced later at runtime. We have developed ACT/J, a prototype of ACT in Java. We describe a case study in which ACT/J is used to introduce three types of autonomic behavior (self-healing, self-optimization, and self-configuration) to a distributed surveillance application.

We define adaptability as the capacity of software in adjusting its behavior in response to changing conditions. To list just a few examples, adaptability is important in pervasive computing, where software in mobile devices need to adapt to dynamic changes in wireless networks; autonomic computing, where software in critical systems are required to be self-manageable; and grid computing, where software for long running scientific applications need to be resilient to hardware crashes and network outages. In this paper, we provide a realization of the transparent shaping programming model, called TRAP.NET, which enables transparent adaptation in existing .NET applications as a response to the changes in the application requirements and/or to the changes in their execution environment. Using TRAP.NET, we can adapt an application dynamically, at run time, or statically, at load time, without the need to manually modify the application original functionality-hence transparent.

Web services paradigm is allowing applications to electronically interact with one another over the Internet. The business process execution language (BPEL) takes this interaction to a higher level of abstraction by enabling the development of aggregate Web services. However, the autonomous and distributed nature of the partner services in an aggregate Web service present unique challenges to the reliability of the composite services. In this paper, we present an approach where existing BPEL processes are automatically instrumented, so that when one or more of their partner services do not provide satisfactory service (e.g., because of a service being overwhelmed, crashed, or because of a network outage), the request for service is redirected to a proxy Web service, where the failed or slow services are replaced by substitute services.

The convergence of data, voice and multimedia communication over digital networks, coupled with continuous improvement in network capacity and reliability has resulted in a proliferation of communication technologies. Unfortunately, despite these new developments, there is no easy way to build new application-specific communication services. The stovepipe approach used today results in rigid technology, limited utility, lengthy and costly development cycle, and diculty in integration. In this paper, we introduce Communication Virtual Machine (CVM) that supports rapid conception, construction and realization of new application-specific communication services through a model-driven, user-centric approach. We present the concept, architecture, modeling language, prototypical design, and implementation of CVM in the context of a healthcare application.

This paper describes Service Clouds, a distributed infrastructure designed to facilitate rapid prototyping and deployment of adaptive communication services. The infrastructure combines adaptive middleware functionality with an overlay network substrate in order to support dynamic instantiation and reconfiguration of services. The Service Clouds architecture includes a collection of low-level facilities that can be invoked directly by applications or used to compose more complex services. After describing the Service Clouds architecture, we present results of experimental case studies conducted on the PlanetLab Internet testbed alone and a mobile computing testbed.

Keywords: Adaptive communication, self-managing system, overlay network, service composition, mobile computing, autonomic computing, quality of service.

This paper describes the internal architecture and operation of an adaptable communication component called the MetaSocket. MetaSockets are created using Adaptive Java, a reflective extension to Java that enables a component's internal architecture and behavior to be adapted at runtime in response to external stimuli. This paper describes how adaptive behavior is implemented in MetaSockets, as well as how MetaSockets interact with other adaptive components, such as decision makers and event mediators. Results of experiments on a mobile computing testbed demonstrate how MetaSockets respond to dynamic wireless channel conditions in order to improve the quality of interactive audio streams delivered to iPAQ handheld computers.

Keywords: Mobile computing, quality-of-service, computational reflection, adaptive middleware, real-time audio-streaming, event-based systems.

Compositional adaptation enables software to modify its structure and behavior dynamically in response to changes in its execution environment. A review of current technology compares how, when, and where recomposition occurs.

Keywords: adaptive software, compositional adaptation, aspect-oriented programming, reflection, component-based design, middleware.

This paper describes the design and operation of a composable proxy infrastructure that enables mobile Internet users to collaborate via heterogeneous devices and network connections. The approach is based on detachable Java I/O streams, which enable proxy filters and transcoders to be dynamically inserted, removed, and reordered on a given data stream. Unlike conventional Java I/O streams, detachable streams can be stopped, disconnected, reconnected, and restarted. As such, they provide a convenient method by which to support the dynamic composition of proxy services. Moreover, use of the I/O stream abstraction enables network distribution and stream adaptability to be implemented transparently with respect to application components. The operation and implementation of detachable streams are described. To evaluate the composable proxy infrastructure, it is used to enhance interactive audio communication among users of a Web-based collaborative computing framework. Two forward error correction (FEC) proxylets are developed, one using block erasure codes and the other using the GSM 06.10 encoding algorithm. Separately, each type of FEC improves the ability of the audio stream to tolerate errors in a wireless LAN environment. When composed in a single proxy, however, they cooperate to correct additional types of burst errors. Results are presented from a performance study conducted on a mobile computing testbed.

Keywords: Adaptive middleware, heterogeneous collaborative computing, mobile hosts, wireless local area networks, forward error correction, interactive audio streams, component-based design.

In cloud computing, virtual containers on physical resources are provisioned to requesting users. Resource providers may pack as many containers as possible onto each of their physical machines, or may pack specific types and quantities of virtual containers based on user or system QoS objectives. Such elastic provisioning schemes for resource sharing may present major challenges to scientific parallel applications that require task synchronization during execution. Such elastic schemes may also inadvertently lower utilization of computing resources. In this paper, we describe the elasticity constraint effect and ripple effect that cause a negative impact to application response time and system utilization. We quantify the impact using real workload traces through simulation. Then, we demonstrate that some resource scheduling techniques can be effective in mitigating the impacts. We find that a tradeoff is needed among the elasticity of virtual containers, the complexity of scheduling algorithms, and the response time of applications.

Keywords: Efficiency Assessment, Parallel Workloads, Virtual Resources.

Keywords:

Low upfront costs, rapid deployment of infrastructure and flexible management of resources has resulted in the quick adoption of cloud computing. Nowadays, different types of applications in areas such as enterprise web, virtual labs and high-performance computing are al- ready being deployed in private and public clouds. However, one of the remaining challenges is how to allow users to specify Quality of Service (QoS) requirements for composite groups of virtual machines and en- force them effectively across the deployed resources. In this paper, we propose an Infrastructure as a Service resource manager capable of al- locating Distributed Ensembles of Virtual Appliances (DEVAs) in the Cloud. DEVAs are groups of virtual machines and their network connec- tivities instantiated on heterogeneous shared resources with QoS speci- fications for individual entities as well as their connections. We discuss the different stages in their lifecycle: declaration, scheduling, provision- ing and dynamic management, and show how this approach can be used to maintain QoS for complex deployments of virtual resources.

Keywords: DEVA, Cloud Computing, Virtual Appliances

Cloud computing represents a solution for applications with high scalability needs where usage patterns, and therefore resource requirements, may fluctuate based on external circumstances such as exposure or trending. However, in order to take advantage of the cloud’s benefits, software engineers need to be able to express the application’s needs in quantifiable terms. Additionally, cloud providers have to understand such requirements and offer methods to acquire the necessary infrastructure to fulfill the users’ expectations. In this paper, we discuss the design and implementation of an Infrastructure as a Service cloud manager such that non-functional requirements determined during the requirements analysis phase can be mapped to properties for a group of Virtual Appliances running the application. The discussed management system ensures that expected Quality of Service is maintained during execution and can be considered during different development phases.

Keywords: DEVA, Could Computing, Non-Functional Requirements, Software Engineering

We present our work on modeling distributed ensembles of virtual appliances (DEVAs) on Infrastructure as a Service (IaaS) clouds. Designing solutions on IaaS providers require a good understanding of the underlying details such as the software installation or the network configuration. We propose the use of DEVAs, a modeling approach built on top of the notion of virtual appliances, that allows easy-to-compose and ready-to-use cloud application architectures that are IaaSagnostic, and that abstract away unnecessary details for web application developers. In this paper, we extend the definition of a DEVA from previous work by presenting an underlying metamodel and how that metamodel can be transformed to an actual deployment. We also present a case study where we model a web application architecture and we discuss how we can instantiate it in an IaaS cloud. We argue that the DEVA modeling approach is suitable for typical cloud use cases.

Keywords: DEVA, Could Computing, Meta Model, Software Engineering

We present our ideas for modeling groups of interdependent virtual machines in the cloud. We call these models virtual environments. This abstraction is built on top of virtual ap- pliances and the services they provide. We discuss previous attempts in this domain and present our motivations for working on an uncomplicated model for non-expert users of cloud computing such as Web developers and CS stu- dents. Visual and internal representations of the model are presented. Early work on a prototype implementation is de- scribed. We argue that easier to use models such as ours are needed for today’s and tomorrow’s distributed applications.

Keywords: virtual environment, virtual appliance, flexible modeling, cloud computing.

Large-scale applications, in the form of workflows, may require the coordinated usage of resources spreading across multiple administrative domains. Scalable solutions need a decentralized approach to coordinate the execution of such workflows. At runtime, adjustments to the workflow execution plan may be required to meet Quality of Service objectives. In this paper, we provide a decentralized execution approach to large-scale workflows on different resource domains. We also provide a low overhead, decentralized runtime adaptation mechanism to improve the performance of the system. Our prototype implementation is based on standard Condor DAGMan workflow execution engine and does not require any modifications to Condor or its underlying system.

Keywords: Application workflow, resource domain, execution, decentralization, distributed system.

This paper proposes and describes a developed methodology to port complex scientific applications originally written in FORTRAN to the nVidia CUDA. The process was developed and validated by porting an existing FORTRAN weather and forecasting algorithm to a GPU parallel paradigm. We believe that the proposed porting methodology described can be successfully utilized in several other existing scientific applications.

The distributed nature of services poses significant challenges to building robust service-based applications. A major aspect of this challenge is finding a model of service integration that promotes ease of dynamic reconfiguration, in response to internal and external stimuli. Centralized models of composition are not conducive for data-intensive applications such as those in the scientific domain. Decentralized compositions are more complicated to manage especially since no service has a global view of the interaction. In this paper we identify the requirements for dynamic reconfiguration of data-intensive composite services. A hybrid composition model that combines the attributes of centralization and decentralization is proposed. We argue that this model promotes dynamic reconfiguration of data-intensive service compositions.

Keywords: Service Composition Models, Scientific Workflow, Adpaptability, Dynamic Reconfiguration, Choreography, Orchestration.

Scientific workflows are often composed by scientists that are not particularly familiar with performance and faulttolerance issues of the underlying layer. The inherent nature of the infrastructure and environment for scientific workflow applications means that the movement of data comes with reliability challenges. Improving the reliablility scientific workflows in distributed environments, calls for the decoupling of data staging and computation activities, and each aspect needs to be addressed separately In this paper, we present an approach to managing scientific workflows that specifically provides constructs for reliable data staging. In our framework, data staging tasks are automatically separated from computation tasks in the definition of the workflow. High-level policies can be provided that allow for dynamic adaptation of the workflow to occur. Our approach permits the separate specification of the functional and non-functional requirements of the application and is dynamic enough to allow for the alteration of the workflow at runtime for optimization.

Keywords: Data Staging, Scientific Workflow, and Distributed Systems.

Due to the widespread use of web services by enterprises, the need to ensure their reliability has become crucial. There are several standards that intend to govern how web services are designed and implemented, including protocols to which they must adhere. These standards include the WS-Reliability and WS-Reliable Messaging standards that define rules for reliable messaging. We present here patterns for these standards which define how to achieve reliable messaging between entities. We compare their features and use.

Keywords: Web Services, Reliability, and Patterns.

The lack of access to sufficient computational, storage, and networking resources in the past three years has proven to be the major hurdle in the rate of discovery for our GCB research projects. The TeraGrid Pathway Fellowship Program has helped us address this problem. In this presentation, we will show how this program has helped us enhance the syllabus and contents of the GCB course with the existing TeraGrid educational and training materials (e.g., the CI Tutor) so that the students taking the GCB course become able to utilize the TeraGrid resources to accelerate the rate of their findings and to be able to submit their research papers for publication within the two semesters of the GCB program.

Keywords: Global CyberBridges, TeraGrid, and High-Performance Computing.

Grid meta-broker is a key enabler in realizing the full potential of inter-operating grid computing systems. A challenge to properly evaluate the effectiveness of meta-brokers is the complexity of developing a realistic grid experimental environment. In this paper, this challenge is addressed by a unique combination of two approaches: using compressed workload traces to demonstrate the resource matching and scheduling functions of the meta-broker, and using emulation to provide a flexible and scalable modeling and management for local resources of a grid environment. Real workload traces are compressed while preserving their key workload characteristics to allow exploration of various dimensions of meta-broker functions in reasonable time. Evaluation of round-robin, queue-length, and utilization based meta-broker scheduling algorithms shows that they have different effects on various workloads.

Keywords: Grid Computing, Meta-Broker, Job Scheduling, and Experimental Evaluation.

Parallel applications have a pressing need for the utilization of more and more resources to meet user performance expectations. Unfortunately, these resources are not necessarily available within one single domain. Grid computing provides a solution to scaling out from a single domain; however, it also brings another problem for some applications: resource heterogeneity. Since some applications require having homogeneous resources for their execution, virtualizing the resources is a noble and viable solution. In this paper, we present two parallel applications, namely WRF and mpiBLAST and report the results of different runs scaling them out from 2 to 128 virtual nodes. Later, we analyze the effects of scaling out based on the application’s communication behavior.

Keywords: Large Scale Computing, Virtualized Resources, and Experimental Study.

In this paper, we describe the use of a tiled-display wall platform for use as a general purpose collaboration and learning platform. The main scenario of emphasis for this work is online learning by users in different countries. We describe the general efficacy of this platform for our purposes and describe its shortcomings for this purpose empirically. We discuss its advantages and also the shortcomings that we found. We also describe an enhancement made to make it more viable for our target usage scenario by implementing an interface for a modern human interface device.

Keywords: Cyberinfrastructure, interdisciplinary, collaboration, e-learning.

This Partnership for International Research and Education (PIRE) is a 5-year long project funded by the National Science Foundation that aims to provide 196 international research and training experiences to its participants by leveraging the established programs, resources, and community of the Latin American Grid (LA Grid, an international academic and industry partnership designed to promote research, education and workforce development at major institutions in the USA, Mexico, Argentina, Spain, and other locations around the world). In return, PIRE will take LA Grid to the next level of research and education excellence. Top students, particularly underrepresented minorities, are engaged and each participant will receive multiple perspectives in each of three different aspects of collaboration as they work with (1) local and international researchers, in (2) academic and industrial research labs, and on (3) basic and applied research projects. PIRE participants will engage not only in computer science research topics focused on transparent cyberinfrastructure enablement, but will also be exposed to challenging scientific areas of national importance such as meteorology, bioinformatics, and healthcare. During the first year of this project, 18 students out of a pool of 68 applicants were selected; they participated in complementary PIRE research projects, visited 7 international institutions (spanning 5 countries and 4 continents), and published 9 papers.

According to Computing Research Association, between 2003 and 2007 each year fewer than 3in computer science and computer engineering are Hispanic or African American and fewer than 20under-representation not only compromises the competitiveness of the US economy, but also deepens the divide and imbalances in our society. It is therefore imperative that undergraduate institutions introduce students to graduate school at an early stage of their academic careers and to provide them with the tools necessary for the successful conduct of research in graduate programs. The School of Computing and Information Sciences (SCIS) at Florida International University (FIU) has been working to strengthen the pipeline of underrepresented students to graduate work in computer science by hosting an NSF Research Experiences for Undergraduates (REU) site for the last three years. Our REU site provided this opportunity to 30 undergraduate students, 23 of them were underrepresented including 7 females, 16 Hispanics, and 4 African Americans, who published 13 technical papers. Six of ten students who have already graduated, have started their graduate studies.

Nowadays, many grid applications are developed as job flows that are composed of multiple jobs. The execution of job flows requires the support of a job flow manager and a job scheduler. Due to the long running nature of job flows, the support for fault tolerance and recovery policies is especially important, and yet complicated due to the sequencing and dependency of jobs within a flow, and the required coordination between workflow engines and job schedulers. In this paper, we describe the design and implementation of a job flow manager that supports fault tolerance. First, we identify and label job flow patterns within a job flow during deployment time. Next, at run time, we introduce a proxy that intercepts and resolves faults using job flow patterns and their corresponding fault recovery policies. Our design has the advantages of separation of job flow and fault handling logic, requiring no manipulation at the modeling time, and flexibility in fault resolution at run time. We validate our design by a prototypical implementation based on the ActiveBPEL workflow engine and GridWay Metascheduler, and Montage application as the case study.

The execution of job flow applications is a reality today in academic and industrial domains. Current approaches to execution of job flows often follow proprietary solutions on expressing the job flows and do not leverage recurrent job-flow patterns to address faults in Grid computing environments. In this paper, we provide a design solution to development of job-flow managers that uses standard technologies such as BPEL and JSDL to express job flows and employs a two-layer peer-to-peer architecture with interoperable protocols for cross-domain interactions among job-flow mangers. In addition, we identify a number of recurring job-flow patterns and introduce their corresponding fault-tolerant patterns to address runtime faults and exceptions. Finally, to keep the business logic of job flows separate from their fault-tolerant behavior, we use a transparent proxy that intercepts job-flow execution at runtime to handle potential faults using a growing knowledge base that contains the most recently identified job-flow patterns and their corresponding fault-tolerant patterns.

With Web services, distributed applications can be encapsulated as self-contained, discoverable software components that can be integrated to create other applications. BPEL allows for the composition of existing Web services to create new higher-function Web services. We identified that the techniques currently applied at development time are not sufficient for ensuring the reliability of composite Web services In this paper, we present a language-based approach to transparently adapting BPEL processes to improve reliability. This approach addresses reliability at the Business process layer (i.e the language layer) using a code generator, which weaves fault-tolerant code to the original code and an external proxy. The generated code uses standard BPEL constructs, and therefore, does not require any changes to the BPEL engine.

In the last few years we have witnessed a number of devastating hurricanes around the world. It is believed that the global climate change is fuelling an increase in the magnitude and also in the average number of hurricanes and tropical storms. Therefore, there is a pressing need to provide a range of users with accurate and timely formation that can enable effective planning for and response to potential hurricane landfalls. The Weather Research Forecast (WRF) code has been adopted worldwide by meteorological services. The numerical model employed by WRF demands a large amount of computing nodes. Such demands can increase dramatically if the WRF is used to model a large geographical area with a high resolution level (e.g. < 1 km). Although WRF can be run in homogeneous clusters, it was not designed for grid environments which can potentially offer a higher amount of computing resources. The transparent Grid enablement of WRF includes carrying out intelligent brokering and scheduling. This is needed to ensure that a run of WRF will take an acceptable amount of time and to optimize the Grid resource usage. Resource usage prediction is required to achieve such brokering and scheduling. However, current approaches to resource prediction tend to address parts of the problem by either focusing on a specific application, or a specific platform, or a small subset of system resources. In this paper, we present our research on Grid enablement of WRF by leveraging our work on resource usage prediction, metascheduling and job-flow management. We report on our experience on the design and development of the La Grid WRF Portal to provide a comprehensive, but customized, Web-based user interface for meteorologist to conduct their hurricane research and to forecast hurricanes in near real-time. We pay a special focus on our approach for modelling application resource usage in a platform independent manner enabling prediction of resource usage on unseen platforms.

Keywords: Grid Enablement, Scientific Applications, WRF, Portal, Meta-Scheduling, Job Flow Management, Modeling, and Profiling.

Host intrusions prevention systems for both servers and end-hosts must address the dual challenges of accuracy and performance. Researchers have mostly focused on addressing the former challenge, suggesting solutions based either on exploit-based penetration detection or anomaly-based misbehavior detection, but yet stopping short of comprehensive solutions that leverage merits of both approaches. The second challenge, however, is rarely addressed; doing so comprehensively is important for practical usability, since these systems can introduce substantial overhead and cause system slowdown, more so when the system load is high. We present Rootsense, a holistic and real-time intrusion prevention system that combines the merits of misbehaviorbased and anomaly-based detection. Four principles govern the design and implementation of Rootsense. First, Rootsense audits events within different subsystems of the host OS and correlates them to comprehensively capture the global system state. Second, Rootsense restricts the detection domain to root compromises only; doing so reduces runtime overhead and increases detection accuracy (root behavior is more easily modeled than user behavior). Third, Rootsense adopts a dual approach to intrusion detection – a root penetration detector detects activities that exploit system vulnerabilities to penetrate the security perimeter, and a root misbehavior detector that tracks misbehavior by root processes. Fourth, Rootsense is designed to be configurable for overhead management allowing the system administrator to tune the overhead characteristics of the intrusion prevention system that affect foreground task performance. A Linux implementation of Rootsense is analyzed for both accuracy and performance, using several real-world exploits and a range of end-host and server benchmarks.

Keywords: Operating systems, security, rootsense.

Current weather forecast and visualization systems lack the scalability to support numerous customized requests for weather research and forecasting, especially at the time of natural disasters such as a hurricane landfall. Most of these systems provide somewhat generic forecasts for different types of users including meteorologists, business owners and emergency management officials. Such forecast while may be relevant to some specific group of users; to others it may not provide any useful information apart from the prediction of impending weather hazards. In other words, one size does not fit all. Weather data and its visualization indicating inclement weather conditions such as snow or ice storm, tornadoes and hurricanes need to be customized for the different type of users using such systems; thus, assisting them in ensuring effective preparatory and meticulous recovery plans. In this paper, we propose a self-configurable, user specific on-demand weather research and forecasting system that utilizes Grid computing to facilitate scalable weather forecast data analysis and prediction.

Keywords: Web-based portal, weather forecasting, WRF, self-configuration, ensemble forecasting.

The execution of job flow applications is a reality today in academic and industrial domains. In this paper, we propose an approach to adding self-healing behavior to the execution of job flows without the need to modify the job flow engines or redevelop the job flows themselves. We show the feasibility of our non-intrusive approach to self-healing by inserting a generic proxy to an existing two-level job-flow management system, which employs job flow based service orchestration at the upper level, and service choreography at the lower level. The generic proxy is inserted transparently between these two layers so that it can intercept all their interactions. We developed a prototype of our approach in a real Grid environment to show how the proxy facilitates runtime handling for failure recovery.

Keywords: job-flow management, meta-scheduler, generic proxy, fault-tolerance, job-flows.

Grid computing supports workload execution on computing resources that are shared across a set of collaborative organizations. At the core of workload management for Grid computing is a software component, called meta-scheduler or Grid resource broker, that provides a virtual layer on top of heterogeneous Grid middleware, schedulers, and resources. Meta-schedulers typically enable end-users and applications to compete over distributed shared resources through the use of one or more instances of the same meta-scheduler, in a centralized or distributed manner, respectively. We propose an approach to enabling autonomic meta-scheduling through the use of a new communication protocol that –if adopted by different meta-schedulers or by the applications using them— can improve the workload execution while avoiding potential chaos, which can be resulted from blind competition over resources. This can be made possible by allowing the metaschedulers and/or their applications to engage in a process to negotiate their roles (e.g., consumer, provider, or both), scheduling policies, service-level agreement, etc. To show the feasibility of our approach, we developed a prototype that enables some preliminary autonomic management among three different meta-schedulers, namely, GridWay, eNANOS, and TDWB.

Keywords: meta-scheduler, grid resource broker, grid interoperability, autonomic workload management.

Grid computing supports the harness of computing resources from cooperating organizations or institutes in the form of virtual organizations. At the core of matching the resource requests for jobs is a resource brokering middleware, commonly known as a meta-scheduler or a meta-broker. The recent advances in meta-scheduling capabilities are broadened to resource matching across multiple virtual organizations, not limiting to a single one. Different architectures have been proposed for these interoperating meta-scheduling systems. In this paper, we present a hybrid approach, combining hierarchical and peer-to-peer architectures for flexibility and extensibility of these systems. We also define a set of protocols to allow different meta-scheduler instances to communicate using Web Services. In our experiments, three remote organizations using different scheduling technologies (namely, IBM, BSC, and FIU) interoperate using the communication protocols.

Keywords: meta-scheduler, resource broker, interoperable scheduling protocol.

In a Grid computing environment, resources are shared among a large number of applications. Brokers and schedulers find matching resources and schedule the execution of the applications by monitoring dynamic resource availability and employing policies such as first-come- first-served and back-filling. To support applications with timeliness requirements in such an environment, brokering and scheduling algorithms must address an additional problem - they must be able to estimate the execution time of the application on the currently available resources. In this paper, we present a modeling approach to estimating the execution time of long-running scientific applications. The modeling approach we propose is generic; models can be constructed by merely observing the application execution “externally” without using intrusive techniques such as code inspection or instrumentation. The model is cross-platform; it enables prediction without the need for the application to be profiled first on the target hardware. To show the feasibility and effectiveness of this approach, we developed a resource usage model that estimates the execution time of a weather forecasting application in a multi-cluster Grid computing environment. We validated the model through extensive benchmarking and profiling experiments and observed prediction errors that were within 10experience, we believe that our approach can be used to model the execution time of other time-sensitive scientific applications; thereby, enabling the development of more intelligent brokering and scheduling algorithms.

Keywords: High-Performance Computing, Profiling, Behavior Modeling, Weather Research and Forecasting.

The impact of hurricanes is so devastating throughout different levels of society that there is a pressing need to provide a range of users with accurate and timely information that can enable effective planning for and response to potential hurricane landfalls. The Weather Research and Forecasting (WRF) code is the latest numerical model that has been adopted by meteorological services worldwide. The current version of WRF has not been designed to scale out of a single organization's local computing resources. However, the high resource requirements of WRF for fine-resolution and ensemble forecasting demand a large number of computing nodes, which typically cannot be found within one organization. Therefore, there is a pressing need for the Grid-enablement of the WRF code such that it can utilize resources available in partner organizations. In this paper, we present our research on Grid enablement of WRF by leveraging our work in transparent shaping, GRID superscalar, profiling, code inspection, code modeling, meta-scheduling, and job flow management.

Keywords: Grid Enablement, Scientific Applications, WRF, Portal, Meta-Scheduling, Job Flow Management, Modeling, and Profiling.

Today’s communication-based applications are mostly crafted in a stovepipe development paradigm, which is inflexible to be used by various domain-specific applications and costly in the development phase. In a previous paper [1], we proposed a new design called CVM (Communication Virtual Machine) to overcome these problems by having a high-level API which can be reused and extended easily for user-centric applications in any domain. Within CVM framework, we came across a practical issue, which is actually the case for any end-to-end multimedia communication, namely the NAT-traversal (network address translation) problem that limits the reliability and availability of CVM and variants of CVM. In this paper, we explain about the necessity of self-configuration for the NAT-traversal problem in end-to-end communications, and propose a solution within the core CVM framework.

Keywords: Communication Virtual Machine, CVM, Sele-Configuration, NAT-Resolution.

With the rapid emergence of new communication software and hardware tools and the improvement of telecommunication infrastructures, a new collaboration paradigm is on the horizon that allows researchers around the globe to expand their loop of collaborators to cross geographical and cultural boundaries. However, much needs to be learned from the user experiences not only to improve the quality of the collaboration facilities, but also to develop new social protocols for distributed human interactions. In this paper, we try to analyze the usage of cyberinfrastructure in remote collaboration among researchers. For that, we draw on survey data and interviews with members from different collaborative projects, and we analyze how our current communication tools meet the needs of collaborative research activities. Then, we articulate a series of key challenges and requirements that contemporary teams are facing. In the end, we present ideas on what sorts of collaborative tools need to be built in order to fulfil the distributed and interdisciplinary collaboration projects. Our findings shed light on the factors that drive the use of cyberinfrastructure and the effectiveness in the success of cross-national and interdisciplinary research collaboration and distance learning, and suggest further research topics.

Keywords: e-Science, formative assessment, group collaboration, distributed collaboration, distance learning.

Grid computing aims to create an accessible virtual supercomputer by integrating distributed computers to form a parallel infrastructure for processing applications. To enable service-oriented Grid computing, the Grid computing architecture was aligned with the current Web service technologies; thereby, making it possible for Grid applications to be exposed as Web services. The WSRF set of specifications standardized the association of state information withWeb services (WSResource) while providing interfaces for the management of state data. Key to the realization of the benefits of Grid computing is the ability to integrate WS-Resources to create higher-level applications. The Business Process Execution Language (BPEL) is the leading standard for integrating Web services and as such has a natural affinity to the integration of Grid services. In this paper, we share our experience on using BPEL to integrate, create, and manage WS-Resources that implement the factory pattern. We use a Bioinformatics application as a case study to show how BPEL can be used to orchestrate Grid services. The execution environment for our case study comprises the Globus Toolkit as the Grid middleware and the ActiveBPEL as the BPEL engine. To the best of our knowledge, this work is among the handful approaches that successfully use BPEL for orchestrating WSRF-based services and the only one that includes the discovery and management of instances.

Keywords: BPEL, Grid Computing,WSRF, OGSA-DAI, Service Composition.

A Grid Resource Broker, or also called meta-scheduler, is a component used for matching work to available Grid resources. The Grid resources usually have a local resource management system with a particular scheduler belonging to different IT centers or institutions. These centers or institutions may have different policies or requirements on how the resources should be used. This situation causes two main problems: the user uniform access to the Grid is lost, and the scheduling decisions are taken separately while they should be done in coordination. These problems have been observed in different efforts such as the HPC-Europa project but it is still an open problem. In this paper we discuss how to achieve a new approach in global brokering with new scheduling techniques through meta-brokering. As the result of the discussion on requirements for meta-brokering, we propose a design in two different contexts: as an extension of HPC-Europa on top of different meta-schedulers, and as a distributed model for the LA Grid meta-brokering project.

High performance computing (HPC) is gaining popularity in solving scientific applications. Using the current programming standards, however, it takes an HPC expert to efficiently take advantage of HPC facilities; a skill that a scientist does not necessarily have. This lack of separation of concerns has resulted in scientific applications with rigid code, which entangles non-functional concerns (i.e., the parallel code) into functional concerns (i.e., the core business logic). Effectively, this tangled code hinders the maintenance and evolution of these applications. In this paper, we introduce Transparent Grid Enabler (TGE) that separates the task of developing the business logic of a scientific application from the task of improving its performance. TGE achieves this goal by integrating two existing software tools, namely, TRAP/J and GRID superscalar. A simple matrix multiplication program is used as a case study to demonstrate the current use and capabilities of TGE.

We define adaptability as the capacity of software in ad-justing its behavior in response to changing conditions. To list just a few examples, adaptability is important in pervasive computing, where software in mobile devices need to adapt to dynamic changes in wireless networks; autonomic computing, where software in critical systems are required to be self-manageable; and grid computing, where software for long running scientific applications need to be resilient to hardware crashes and network out-ages. In this paper, we provide a realization of the trans-parent shaping programming model, called TRAP.NET, which enables transparent adaptation in existing .NET applications as a response to the changes in the applica-tion requirements and/or to the changes in their execution environment. Using TRAP.NET, we can adapt an applica-tion dynamically, at run time, or statically, at load time, without the need to manually modify the application original functionality-hence transparent.

The 'CyberBridges' pilot project is an innovative model for creating a new generation of scientists and engineers who are capable of fully integrating cyberinfrastructure into the whole educational, professional, and creative process of their respective disciplines. CyberBridges augments graduate student education to include a foundation of understanding in Advanced Networking and Grid Infrastructure for High Performance Computing, and bridges the divide between the information technology community and diverse science and engineering disciplines. CyberBridges is increasing the rate of discovery for science and engineering faculty by empowering them with cyberinfrastructure, fostering inter-disciplinary research collaboration, improving minority graduate education, and institutionalizing this change process. We demonstrate the effectiveness of CyberBridges by providing four case studies with graduate students of Physics, Bioinformatics, Chemistry, and Biomedical Engineering. Groundwork has begun to extend the outreach of CyberBridges for international research and education collaborations.

Multimedia communication services today are conceived, designed, and developed in isolation, following a stovepipe approach. This has resulted in a fragmented and incompatible set of technologies and products. Building new communication services requires a lengthy and costly development cycle, which severely limits the pace of innovation. In this paper, we address the fundamental problem of automating the development of multimedia communication services. We propose a new paradigm for creating such services through declarative specification and generation, rather than through traditional design and development. Further, the proposed paradigm pays special attention to how the end-user specifies his/her communication needs, an important requirement largely ignored in existing approaches. We argue that for the domain of user-centric multimedia communication services, the proposed approach of automatic generation is not only feasible in terms of the ability to meet a range of communication needs in several domains, but is also desirable for maintaining and improving the pace of innovation in multimedia communication services.

TRAP/BPEL is a framework that adds autonomic behavior into existing BPEL processes automatically and transparently. We define an autonomic BPEL process as a composite Web service that is capable of responding to changes in its execution environment (e.g., a failure in a partner Web service). Unlike other approaches, TRAP/BPEL does not require any manual modifications to the original code of the BPEL processes and there is no need to extend the BPEL language nor its BPEL engine. In this paper, we describe the details of the TRAP/BPEL framework and use a case study to demonstrate the feasibility and effectiveness of our approach.

Keywords: TRAP/BPEL, generic proxy, self-management, dynamic service discovery.

Web services paradigm is allowing applications to interact with one another over the Internet. BPEL facilitates this interaction by providing a platform through which Web services can be integrated. However, the autonomous and distributed nature of the integrated services presents unique challenges to the reliability of composed services. The focus of our ongoing research is to transparently introduce autonomic behavior to BPEL processes in order to make them more resilient to the failure of partner services. In this work, we present an approach where BPEL processes are adapted by redirecting their interactions with partner services to a dynamic proxy. We describe the generative adaptation process and the architecture of the adaptive BPEL processes and their corresponding proxies. Finally, we use case studies to demonstrate how generated dynamic proxies are used to support self-healing and self-optimization in instrumented BPEL processes.

Keywords: RobustBPEL2, dynamic proxy, self-management, dynamic service discovery.

The development of collaborative multimedia applications today follows a vertical development approach, which is a major inhibitor that drives up the cost of development and slows down the pace of innovation of new generations of collaborative applications. In this paper, we propose a network communication broker (NCB) that provides a unified higher-level abstraction that encapsulates the complexity of network-level communication control and media delivery for the class of multimedia collaborative applications. NCB expedites the development of next-generation applications with various communication logics. Furthermore, NCB-based applications can be easily ported to new network environments. In addition, the self-managing design of NCB supports dynamic adaptation in response to changes in network conditions and user requirements.

Keywords: Network communication broker, multimedia, middleware.

The convergence of data, voice and multimedia communication over digital networks, coupled with continuous improvement in network capacity and reliability has significantly enriched the ways we communicate. However, the stovepipe approach used to develop today’s communication applications and tools results in rigid technology, limited utility, lengthy and costly development cycle, difficulty in integration, and hinders innovation. In this paper, we present a fundamentally different approach, which we call Communication Virtual Machine (CVM) to address these problems. CVM provides a user-centric, modeldriven approach for conceiving, synthesizing and delivering communication solutions across application domains. We argue that CVM represents a far more effective paradigm for engineering communication solutions. The concept, architecture, modeling language, prototypical design and implementation of CVM are discussed.

Keywords: Model driven, communication application, multimedia, middleware, telemedicine.

We recently introduced Service Clouds, a distributed infrastructure designed to facilitate rapid prototyping and deployment of autonomic communication services. In this paper, we propose a model that extends Service Clouds to the wireless edge of the Internet. This model, called Mobile Service Clouds, enables dynamic instantiation, composition, configuration, and reconfiguration of services on an overlay network to support self-management in mobile computing. We have implemented a prototype of this model and applied it to the problem of dynamically instantiating and migrating proxy services for mobile hosts. We conducted a case study involving data streaming across a combination of Planet- Lab nodes, local proxies, and wireless hosts. Results are presented demonstrating the effectiveness of the prototype in establishing new proxies and migrating their functionality in response to node failures.

Keywords: autonomic networking, distributed service composition, self-managing system, overlay network, mobile computing, quality of service.

Web services are increasingly being used to expose applications over the Internet. To promote efficiency and the reuse of software, these Web services are being integrated both within enterprises and across enterprises, creating higher function services. BPEL is a workflow language that can be used to facilitate this integration. Unfortunately, the autonomous nature of Web services leaves BPEL processes susceptible to the failures of their constituent services. In this paper, we present a systematic approach to making existing BPEL processes more fault tolerant by monitoring the involved Web services at runtime, and by replacing delinquent Web services. To show the feasibility of our approach, we developed a prototype implementation that generates more robust BPEL processes from existing ones automatically. The use of the prototype is demonstrated using an existing Loan Approval BPEL process.

Keywords: ECommerce, Web Service Monitoring, Robust BPEL Processes.

Web services are increasingly being used to expose applications over the Internet. These Web services are being integrated within and across enterprises to create higher function services. BPEL is a workflow language that facilitates this integration. Although both academia and industry acknowledge the need for workflow languages, there are few technical papers focused on BPEL. In this paper, we provide an overview of BPEL and discuss its promises, limitations and challenges.

Keywords: Web services, workflow language, BPEL, business processes, A2A integration, and B2B integration.

Increasingly, software systems are constructed by composing multiple existing applications. The resulting complexity increases the need for self-management of the system. However, adding autonomic behavior to composite systems is difficult, especially when the existing components were not originally designed to support such interactions. Moreover, entangling the code for integrated selfmanagement with the code for the business logic of the original applications may actually increase the complexity of the system, counter to the desired goal. In this paper, we propose a technique to enable self-managing behavior to be added to composite systems transparently, that is, without requiring manual modifications to the existing code. The technique uses transparent shaping, developed previously to enable dynamic adaptation in existing programs, to weave self-managing behavior into existing applications, which interact through Web services. A case study demonstrates the use of this technique to construct a fault-tolerant surveillance application from two existing applications, one developed in .NET and the other in CORBA, without the need to modify the source code of the original applications.

Keywords: application integration, adaptive middleware, autonomic computing, self-configuration, fault-tolerance, dynamic adaptation, transparent adaptation.

The need for adaptability in software is growing, driven in part by the emergence of pervasive and autonomic computing. In many cases, it is desirable to enhance existing programs with adaptive behavior, enabling them to execute effectively in dynamic environments. In this paper, we propose a general programming model called transparent shaping to enable dynamic adaptation in existing programs. We describe an approach to implementing transparent shaping that combines four key software development techniques: aspect-oriented programming to realize separation of concerns at development time, behavioral reflection to support software recon- figuration at run time, component-based design to facilitate independent development and deployment of adaptive code, and adaptive middleware to encapsulate the adaptive functionality. After presenting the general model, we discuss two specific realizations of transparent shaping that we have developed and used to create adaptable applications from existing programs.

In pervasive computing environments, conditions are highly variable and resources are limited. In order to meet the needs of applications, systems must adapt dynamically to changing situations. Since adaptation at one system layer may be insufficient, crosslayer, or vertical approaches to adaptation may be needed. Moreover, adaptation in distributed systems often requires horizontal cooperation among hosts. This cooperation is not restricted to the source and destination(s) of a data stream, but might also include intermediate hosts in an overlay network or mobile ad hoc network. We refer to this combined capability as universal adaptation. We contend that the model defining interaction between adaptive middleware and the operating system is critical to realizing universal adaptation. We explore this hypothesis by evaluating the Kernel-Middleware eXchange (KMX), a specific model for crosslayer, cross-system adaptation. We present the KMX architecture and discuss its potential role in supporting universal adaptation in pervasive computing environments. We then describe a prototype implementation of KMX and show results of an experimental case study in which KMX is used to improve the quality of video streaming to mobile nodes in a hybrid wired-wireless network.

Keywords: Adaptive middleware, pervasive computing, cross-layer adaptation, universal adaptation, multimedia communication, quality of service, video streaming, wireless network.

This paper describes TRAP/J, a software tool that enables new adaptable behavior to be added to existing Java applications transparently (that is, without modifying the application source code and without extending the JVM). The generation process combines behavioral reflection and aspect-oriented programming to achieve this goal. Specifically, TRAP/J enables the developer to select, at compile time, a subset of classes in the existing program that are to be adaptable at run time. TRAP/J then generates specific aspects and reflective classes associated with the selected classes, producing an adapt-ready program. As the program executes, new behavior can be introduced via interfaces to the adaptable classes. A case study is presented in which TRAP/J is used to introduce adaptive behavior to an existing audio-streaming application, enabling it to operate effectively in a lossy wireless network by detecting and responding to changing network conditions.

Keywords: generator framework, transparent adaptation, dynamic reconfiguration, aspect-oriented programming, behavioral reflection, middleware, mobile computing, quality-of-service.

This paper addresses the energy consumption of forward error correction (FEC) protocols as used to improve quality-of-service (QoS) for wireless computing devices. The paper also characterizes the effect on energy consumption and QoS of the power saving mode in 802.11 wireless local area networks (WLANs). Experiments are described in which FEC-encoded audio streams are multicast to mobile computers across a WLAN. Results of these experiments quantify the tradeoffs between improved QoS, due to FEC, and additional energy consumption caused by receiving and decoding redundant packets. Two different approaches to FEC are compared relative to these metrics. The results of this study enable the development of adaptive software mechanisms that attempt to manage these tradeoffs in the presence of highly dynamic wireless environments.

Keywords: energy consumption, quality-of-service, forward error correction, mobile computing, handheld computer, adaptive middleware.

This paper addresses the design of adaptive middleware to support autonomic computing in pervasive computing environments. The particular problem we address here is how to support self-optimization to changing network connection capabilities as a mobile user interacts with heterogeneous elements in a wireless network infrastructure. The goal is to enable self-optimization to such changes transparently with respect to the core application code. We propose a solution based on the use of the generic proxy, which is a specific CORBA object that can intercept and process any CORBA request using rules and actions that can be introduced to the knowledge base of the proxy during execution. To explore its design and operation, we have incorporated the generic proxy into ACT [1], an adaptive middleware framework we designed previously to support adaptation in CORBA applications. Details of the generic proxy are presented, followed by results of a case study enabling self-optimization for an existing surveillance application in a heterogeneous wireless environment.

Keywords: adaptive middleware, autonomic computing, self-optimization, dynamic adaptation, transparent adaptation, generic proxy, quality-of-service, mobile computing, CORBA.

This paper introduces TRAP/J, a software tool that enables autonomic computing in existing Java programs by generating adapt-ready versions of the original programs at compile time. The generation process is transparent to the original program source code, in which there is no need to modify the source code manually. At run time, new behavior can be introduced to the adapt-ready programs. To reduce overhead, TRAP/J enables the developer to select, at compile time, a subset of classes, constituting an existing program, to be adaptive at run time. To support dynamic adaptation in existing Java programs, TRAP/J benefits from aspect-oriented programming and behavioral reflection. TRAP/J generate specific aspects and reflective classes associated with the selected classes. A case study is presented in which TRAP/J was used to enable an existing audio-streaming application to perform self-optimization in a wireless network environment by adapting to changing conditions automatically.

Keywords: autonomic computing, adapt-ready programs, transparent adaptation, aspect-oriented programming, behavioral reflection, middleware, quality-of-service, mobile computing.

This paper proposes an Adaptive CORBA Template (ACT), which enables run-time improvements to CORBA applications in response to unanticipated changes in either their functional requirements or their execution environments. ACT enhances CORBA applications by weaving adaptive code into the applications' object request brokers (ORBs) at run time. The woven code intercepts and adapt the requests, replies, and exceptions that pass through the ORBs. ACT itself is language- and ORB-independent. Specifically, ACT can be used to develop an object-oriented framework in any language that supports dynamic loading of code and can be applied to any CORBA ORB that supports portable interceptors. Moreover, ACT can be integrated with other adaptive CORBA frameworks and can be used to support interoperation among otherwise incompatible frameworks. To evaluate the performance and functionality of ACT, we implemented a prototype in Java to support unanticipated adaptation in non-functional concerns, such as quality-of-service and system-resource management. Our experimental results show that the overhead introduced by the ACT infrastructure is negligible, while the adaptations offered are highly flexible.

Keywords: middleware, CORBA, dynamic adaptation, interoperability, request interceptor, dynamic weaving, proxy, quality-of-service, mobile computing

This paper describes the internal architecture and operation of an adaptable communication component called the MetaSocket. MetaSockets are created using Adaptive Java, a reflective extension to Java that enables a component's internal architecture and behavior to be adapted at run time in response to external stimuli. This paper describes how adaptive behavior is implemented in MetaSockets, as well as how MetaSockets interact with other adaptive components, such as decision makers and event mediators. Results of experiments on a mobile computing testbed demonstrate how MetaSockets respond to dynamic wireless channel conditions in order to improve the quality of interactive audio streams delivered to iPAQ handheld computers.

This paper investigates the use of programming language constructs to realize adaptive behavior in support of collaboration among users of wearable and handheld computers. A prototype language, Adaptive Java, contains primitives that permit programs to modify their own operation in a principled manner. In a case study, Adaptive Java was used to construct MetaSocket components, whose composition and behavior can be adapted to changing conditions during execution. MetaSockets were then integrated into Pavilion, a web-based collaboration framework, and experiments were conducted on a mobile computing testbed containing wearable, handheld, and laptop computer systems. Performance results demonstrate the utility of MetaSockets to improving the quality of interactive audio streams and reliable data transfers among collaborating users.

Keywords: adaptive middleware, reflection, wearable computing, mobile computing, wireless networks, forward error correction.

This paper proposes the use of programming language constructs to sup­port adaptive self­monitoring and self­reporting software. The methods are particularly well­ suited to wireless mobile devices, where limited resources may constrain the use of certain software audits. An adaptive software ar­chitecture is described that supports run­time transformations on software components, enabling them to report internal details on how they are be­ing used to other parts of the system. Effectively, any component of the system can be turned into an ``informer'' at run time, and the nature of the reported information can be adapted dynamically based on changing conditions or directives from another authority, such as an intrusion detec­ tion system. A prototype implementation is described. The operation of the system is demonstrated through an experiment in which it detects and responds to a malicious host that multicasts ``noise'' packets to a wireless iPAQ handheld computer.

This paper describes the use of programming language constructs to support run-time software adaptation. A prototype language, Adaptive Java, contains primitives that permit programs to modify their own operation in a principled manner. In case studies, Adaptive Java is being used to support adaptation for different crosscutting concerns associated with heterogeneous mobile computing and critical infrastructure protection. Examples are described in which Adaptive Java components support dynamic quality-ofservice on wireless networks, run-time energy management for handheld computers, and self-auditing of potential security threats in distributed environments.

This paper presents an aspect-oriented approach to dynamic adaptation. A systematic process for defining where, when, and how an adaptation is to be incorporated into an application is presented. Specifically, the paper presents a two-phase approach to dynamic adaptation, where the first phase prepares a non-adaptice program for adaptation, and the second phase implements the adaptation at run time. this approach is illustrated with a distributed conferencing application.

Many middleware platforms use computational reflection to support adaptive functionality. Most approaches intertwine the activity of observing behavior (introspection) with the activity of changing behavior (intercession). This paper explores the use of language constructs to separate these parts of reflective functionality. This separation and “packaging” of reflective primitives is intended to facilitate the design of correct and consistent adaptive middleware. A prototype implementation is described in which this functionality is realized through extensions to the Java programming language. A case study is described in which “metamorphic” socket components are created from regular socket classes and used to realize adaptive behavior on wireless network connections.

Keywords: adaptive middleware, reflection, component design, mobile computing, wireless networks, forward error correction.

Introduction: The extent of repeats is a striking feature of genomic DNA sequences. For example, families of repeat sequences account for about one half of the human genome. Finding repeats is useful in finding defective genes and in forensic DNA fingerprinting. Techniques to find repeats can also be used to find signatures for a pair (or a set) of genomes. We consider the problem of finding repeats and signatures in genomic sequences using grid-computing techniques.
Methods: The problem is clearly compute-intensive and creates costly bottlenecks in large-scale analyses. There are a number of programs available for finding repeats, such as the popular RepeatMasker [REF]. We have implemented an algorithm based on suffix arrays for finding repeats, which was then ported to run on MPI clusters with dynamic load balancing. This algorithm also searches for direct repeats and other variants such as inverted repeats and complemented inverted repeats. We have also implemented an algorithm that searches for “signatures”, i.e., subsequences present in one genome but absent from the others.
Results and Conclusions: We have successfully tested our implementation on five strains of the bacteria Pseudomonas aeruginosa, each containing over six million nucleotides. The MPI cluster implementation resulted in near-optimal speedup and balancing.

Keywords: Genome Sequences, Search Engine, Repeats, Signatures, Load Balancing, Fault-Tolerant, Grid Computing.

One of the most striking features of genomic DNA sequences is the extent to which repeated substrings occur in the genome. The problem of identifying subsequences that are repeated or unique has many applications in Bioinformatics. A number of algorithms exist for these problems, but are compute-intensive. We implemented an algorithm based on suffix arrays for finding repeats and unique signatures, which was then ported to run on MPI clusters. It searches for single patterns in a sequence of roughly a few million bases in less than a second. We tested our data on strains of the bacteria Pseudomonas aeruginosa containing over six million nucleotides. The serial implementation completed its task averaging 30 hours for every pair of genomes. The performance of the MPI implementation was compared to that of the serial one. Effectively, the MPI implementation allowed to reduce the amount of time needed to process such tasks.

Keywords: Genome Sequences, Search Engine, Repeats, Signatures, Load Balancing, Fault-Tolerant, Grid Computing.

The main characteristic of genomic data is its large size. For example, the National Institutes of Health (NIH) sequence database has a total of over 65 billion nucleotides. One of the most striking features of genomic DNA sequences is the extent to which repeated substrings occur in the genome. Families of repeat sequences account for about one third of the human genome. The problem of identifying subsequences that are either repeated or unique has many applications in Bioinformatics. Repeat sequences come in many different flavors and are responsible for different functions and diseases. Finding repeats has applications in finding defective genes, and in forensic DNA fingerprinting. It also allows us to find differences between genomes. A number of algorithms and implementations exist for these problems, but are compute-intensive. We have implemented an algorithm based on suffix arrays for finding repeats and unique signatures, which was then ported to run on MPI clusters with good results. This algorithm also searches for direct repeats and other variants such as inverted repeats and complemented inverted repeats. It searches for single patterns in a sequence of roughly a few million bases in less than a second. We tested our data on five strains of the bacteria Pseudomonas aeruginosa, which were acquired from NCBI, each containing over six million nucleotides. The serial implementation completed its task averaging 30 hours for every pair of genomes. The performance of the MPI implementation was compared to that of the serial one. Effectively, the MPI cluster implementation allowed to reduce the amount of time needed to process such tasks.

Keywords: Genome Sequences, Search Engine, Repeats, Signatures, Load Balancing, Faul-Tolerant, Grid Computing.

Weather prediction can save lives and help business owners and emergency responders in the case of inclement weather. The goal of WRF is to improve weather prediction, especially in the case of hurricane mitigation. Optimum forecasting consists of: Accurate and Timely Results and Precise Location Information. Grid-Enablement is the practice of taking existing applications that are currently configured to run on a single machine or cluster and adapting them to run on a non-homogeneous resources connected via the internet with the goal of improving its performance (execution time and resource utilization). The expected benefits of successful grid-enablement to WRF will be: 1) Access to a greater number of compute resources due to the utilization of more domains; 2) Faster results, by virtue of having more compute power (more power = less simulation time); and 3) Improved precision of results (more power = higher domain granularity). The meta-scheduler is the “global” scheduler of the grid environment – above the local schedulers. Its function is to select the best resources for a job to run on by analyzing application and target architecture characteristics to find the best match for optimal resource usage and job return time. Performance prediction provides: 1) the meta-scheduler with accurate prediction run times - “smarter” scheduling; and 2) at local level, improved resource allocation and local scheduling techniques as many users overestimate job execution times for precautionary reasons. Leverage many years of experience from BSC faculty towards these goals and make personal connections that will last throughout our careers.

Keywords: Grid Environment, Application Profiling, Prediction of Execution Time, Meta-Scheduling, WRF.

Homeland Security applications with regard to microbial detection in environmental samples are among the many other possible applications of our project. The main characteristic of genomic data is its large size. For example, the National Institutes of Health (NIH) sequence database has a total of over 65 billion nucleotides. One of the most striking features of genomic DNA sequences is the extent to which repeated substrings occur in the genome. In C. elegans with a genome sequence of 100.2 million nucleotides over 7,000 families of repeat sequences have been identified. Families of repeat sequences account for about one third of the human genome. Repeat sequences come in many different flavors and are responsible for different functions and diseases. Finding repeats has applications in finding defective genes, and in forensic DNA fingerprinting. It also allows us to find differences between genomes.

Keywords: Genome Sequences, Search Engine, Repeats, Signatures, Load Balancing, Faul-Tolerant, Grid Computing.

The purpose of this project is to facilitate remote access to the grid enabled WRF (Weather Research and Forecasting) system, via a web based portal. The portal is used by Meteorologists and provides services facilitating the uploading of WRF configurations, modification of relevant weather variables, and the submission, tracking, and retrieval of jobs.

Keywords: Weather Research and Forecasting, WRF Web-Based Portal, Hurricane Mitigation, Meta-Scheduling, Grid Computing.

Our experience was about developing new avenues as a person and in our research by reaching across cultural boundaries to really experience the meaning of a global living laboratory. The PIRE program offered us a unique opportunity to take on technology from a different perspective than that we receive in our traditional studies. Furthermore, forging working relationships with our international collaborators serves as a big step towards making breakthrough contributions to the research community. Beyond the walls of research, PIRE afforded us the opportunity to develop as individuals, both professionally and personally. We were given the opportunity to really explore the world and take on a trailblazing role in this the first year of this program. As some what of guinea pigs, we were able to set the bar for the years to come in this program and were ambassadors for our home institutions. The experiences we gathered here will be invaluable in the rest of our academic careers as well as our subsequent professional careers.

Keywords: Application Profiling, WRF, High-Performance Computing, Grid Computing

Weather prediction can save lives and help business owners and emergency responders in the case of inclement weather. The goal of WRF is to improve weather prediction, especially in the case of hurricane mitigation. Optimum forecasting consists of: Accurate and Timely Results and Precise Location Information. Grid-Enablement is the practice of taking existing applications that are currently configured to run on a single machine or cluster and adapting them to run on a non-homogeneous resources connected via the internet with the goal of improving its performance (execution time and resource utilization). The expected benefits of successful grid-enablement to WRF will be: 1) Access to a greater number of compute resources due to the utilization of more domains; 2) Faster results, by virtue of having more compute power (more power = less simulation time); and 3) Improved precision of results (more power = higher domain granularity). The meta-scheduler is the “global” scheduler of the grid environment – above the local schedulers. Its function is to select the best resources for a job to run on by analyzing application and target architecture characteristics to find the best match for optimal resource usage and job return time. Performance prediction provides: 1) the meta-scheduler with accurate prediction run times - “smarter” scheduling; and 2) at local level, improved resource allocation and local scheduling techniques as many users overestimate job execution times for precautionary reasons. Leverage many years of experience from BSC faculty towards these goals and make personal connections that will last throughout our careers.

Keywords: Grid Environment, Application Profiling, Prediction of Execution Time, Meta-Scheduling, WRF.

Complex workflows being used more often by scientists. Grid environments are being used to utilize resources in multiple organizations. Heterogeneity and dynamicity of the environment introduces challenges to the execution of workflows. If not handled, these challenges may result in wasted time and resources, and basically unsuccessful execution of workflows. Scientists need reliable tools to successfully conduct their business. Administrators are concerned with the utilization of resources.

Keywords: Workflow, Job Flow, Fault-Tolerant, Design Patterns.

Due to the computational requirements of the Weather Research and Forecast (WRF) numerical simulation models, a large number of computing nodes are needed to reduce the overall wall-clock time of the simulations, so that timely critical decisions can be made. In this project, we use existing and custom-designed profiling and code inspection tools to understand the runtime behavior of WRF on a Grid computing environment. Our goal is to produce a mathematical model to predict the allocation of computing resources, for example, the optimized number of homogenous nodes required for a hurricane path prediction simulation. We will run a number of profiling experiments with different hardware and software configurations. The data gathered through these experiments will help us formulate and validate the mathematical model. This model will enable future self-configured executions of the WRF code on a Grid computing environment.

Keywords: Grid Enablement, Grid Computing, WRF, Profiling, Code Inspection, Modeling.

The Latin American (LA) Grid Weather Research and Forecast (WRF) Portal is aimed to provide a comprehensive Web-based user-friendly interface for meteorologists to conduct their research and forecast, for end users to access and understand the forecast, and for system administrators to configure the underlying computing infrastructure. This way, the meteorologist will spend their time only on the science part of weather research and forecast instead of wasting their time on tedious and error-prone tasks of compiling, installing, and configuring the WRF code. The end users (e.g., business owners, home owners, and emergency response teams) can easily access the forecast that most relates to them and assists their understanding of how to take preventive and recovery actions accordingly. The system administrators can add or remove computational resources (e.g., clusters, supercomputers, and virtual machines) to be allocated dynamically to the different simulation tasks based on some high-level policies. Currently, we are focusing on enabling the meteorologist to set up, automate simulations and organize data results to support ensemble runs of hurricane forecast simulations leveraging grid computing resources.

Keywords: Portal, WRF, Web, Hurricane Mitigation.

The use of local resource management systems to schedule the execution of jobs is extended through many research areas in need of high performance computing. However, while local schedulers manage the rights for a single domain, there is a need to perform a higher level scheduling of resources. A meta-scheduler, or a scheduler of schedulers, is an entity in charge of orchestrating these resources from different organizations. Different architectures have been proposed for meta-scheduling systems. In this poster, we present a hybrid approach, combining hierarchical and peer-to-peer architectures to improve fault tolerance, management and extensibility of these systems. We also define a standard protocol to allow different meta-scheduler instances to communicate using Web Services. In our experiment, three remote organizations using different scheduling technologies (namely, IBM, BSC, and FIU) interoperate using the described protocol.

Keywords: Grid Computing, Meta-Scheduling, LA-Grid.

The execution of job flow applications is a reality today in scientific and enterprise domains, and the LAGrid job flow management project addresses many specific issues related to job flow management in distributed meta-brokering environments. The central idea of the project is a two-level approach that employs job flow based service orchestration at the upper level to control coarse-grain job submissions to the resource management layer, and service choreography at the lower level, or fine-grained interaction with jobs by allocating appropriate computing and data resources. The focus of this project is on experimentation of the span and coordination control of job flow engine and meta-scheduling functions, with respect to distributed (1) job execution, (2) failure management and (3) data handling.

Keywords: Grid Computing, Job Flow Management, Fault-Tolerant.

The lack of an appropriate programming model for grid computing is a major inhibitor in the emergence of grid-enabled applications, which is a focal point of the computer science and engineering portion of the project. In this project, we aim to address this issue by integrating Transparent Shaping developed at FIU, which allows introduction of static and dynamic adaptation in applications, and GRID Superscalar developed at BSC, which allows conversion of a sequential code to a parallel code ready to be run in the grid. This new programming model will include a supporting runtime and middleware that adapts execution of grid-enabled applications as a response to the changes in the grid environment (e.g., the availability of resources such as the number of available computational nodes, the available network bandwidth, and the available data storage). Development work will be carried out first at FIU by selected students and will be supervised by Masoud Sadjadi and then by Rosa M. Badia at BSC.

Keywords: Grid Computing, Prgramming Model, Transparent Shaping, GRID superscalar.

We present a case study on the Autonomic System Manager that we designed and developed in the School of Computing and Information Sciences at Florida International University (FIU-SCIS) as a proof of concept to the usefulness of data mining in autonomic system management. Autonomic System Manager is a software system that helps system and network administrators with their complex job. It automatically monitors Net flow (network flow) data and responds to anomalies in an adaptive fashion. The current focus of this prototype is on network anomaly detection. The purpose of anomaly detection is to detect and categorize situations that deviate from the normal baselines. Network and system administrators receive a torrent of IP requests everyday. As is the case with any network, some events are malicious. In order to secure a network system effectively administrators need to use tools that have the capability to classify malicious activities and respond to threats in real time. The principle challenge in automatically detecting insidious intruders is that the range of anomalous activities is huge, ranging from port scans, denial of service attacks, ping of death, SYN flooding, UDP flooding, worms, viruses, and disconnecting models remotely.

Keywords: Data mining, autonomic computing, self management, self protection, network and system management, anomaly detection.

Rapid improvements in mobile computing devices and wirelessnetworks promise to provide a foundation for ubiquitouscomputing. However, comparable advances are needed in the designof mobile computing applications and supporting middleware.Distributed software must be able to adapt to dynamic situationsrelated to several cross-cutting concerns, includingquality-of-service, fault-tolerance, energy management, andsecurity. We previously introduced Adaptive Java, an extensionto the Java programming language, which provides languageconstructs and compiler support for the development of adaptive middleware.This paper describes the use of Adaptive Java to develop anadaptable communication component called the MetaSocket.MetaSockets are created from existing Java socket classes, buttheir structure and behavior can be adapted at run time inresponse to external stimuli. We are using MetaSockets forseveral distributed computing tasks, including audits of traffic patternsfor intrusion detection, adaptive error control on wirelessnetworks, and dynamic energy management for handheld and wearablecomputers. This paper focuses on the internal architecture and operation ofMetaSockets. In particular, we describe in detail howadaptive behavior is implemented and how MetaSockets interact with other adaptive components, such as decision makersand event mediators. Results are presented from experimentson a mobile computing testbed in which MetaSockets respond to dynamic wireless channel conditionsin order to improve the qualityof interactive audio streams delivered to iPAQ handheld computers.

The increasing demand for resources of the high performance computing systems has led to new forms of collaboration of distributed systems such as interoperable grid systems that contain and manage their own resources. While with a single domain one of the most important tasks is the selection of the most appropriate set of resources to dispatch a job, in an interoperable grid environment this problem shifts to selecting the most appropriate domain containing the requiring resources for the job. In the Latin American Grid initiative, our model consists of multiple domains. Each domain has its domain broker, and the task of scheduling on top of brokers can be called metabrokering or broker selection. In this paper, we present and evaluate the “bestBrokerRank” broker selection policy and its two different variants. The first one uses the resource information in aggregated forms as input, and the second one also uses the brokers average bounded slowdown as a dynamic performance metric. From our evaluations performed with simulation tools, we state that the proposed resource aggregation algorithms are scalable for an interoperable grid environment and we show that the best performance results are obtained with our coordinated policy. We conclude that delegating part of the scheduling responsibilities to the underlying scheduling layers is a good way to balance the performance among the different brokers and schedulers.

Keywords: Grid Computing, Scheduling Strategies, Interoperability.

The increasing demand for resources of the high performance computing systems has led to new forms of collaboration of distributed systems such as interoperable grid systems that contain and manage their own resources. While with a single domain one of the most important tasks is the selection of the most appropriate set of resources to dispatch a job, in an interoperable grid environment this problem shifts to selecting the most appropriate domain containing the requiring resources for the job. In the Latin American Grid initiative, our model consists of multiple domains. Each domain has its domain broker, and the task of scheduling on top of brokers can be called metabrokering or broker selection. In this paper, we present and evaluate the “bestBrokerRank” broker selection policy and its two different variants. The first one uses the resource information in aggregated forms as input, and the second one also uses the brokers average bounded slowdown as a dynamic performance metric. From our evaluations performed with simulation tools, we state that the proposed resource aggregation algorithms are scalable for an interoperable grid environment and we show that the best performance results are obtained with our coordinated policy. We conclude that delegating part of the scheduling responsibilities to the underlying scheduling layers is a good way to balance the performance among the different brokers and schedulers.

Keywords: Grid Computing, Scheduling Strategies, Interoperability.

This technical report details a case study done for the TRAP.NET project (which stands for Transparent Reflective Aspect Programming in Microsoft’s .NET Framework). TRAP.NET provides dynamic adaptation for software programs written in .NET.

Keywords: TRAP.NET, Dynamic Adaptation, and Math Service.

This evaluation of profiling tools started because of a need to examine the behavior of the Weather Research and Forecasting Model (WRF). It is currently written to run on a single cluster; our team wished to explore the options for scaling out WRF to a grid environment. This necessitated understanding how WRF works and what its resource usage patterns look like. In order to do this we required a profiler, which led us to creating this document. We began with a long list of tools, but an in depth investigation for each of them would have been both ineffective and unwarranted. Instead, we broke our evaluation of the field into three passes. In the first pass we discarded programs that did not meet our basic criteria, things like architecture and language support. The second pass was more qualitative. We came up with a list of pros and cons for each tool, and rejected those that did not have features we wanted. Those tools that were still being considered were then given an extensive trial to determine how well they worked for us. We examined things like documentation, ease of installation, whether the tool provided source code correlation and call graphs, etc. It is our hope that this information proves useful to the community, allowing researchers and professionals to learn from our experiences.

Keywords: Profiler Evaluation, Weather Forecast and Reserach, WRF, Fortran, MPI, OpenMP, Grid Computing.

In Summer 2007, we, a group of four undergraduate students and three graduate students under the supervision of Dr. Masoud Sadjadi at School of Computing and Information Science (SCIS) of Florida International University (FIU), started an effort on gridifying the Weather Research and Forecast (WRF) code. During this process, it became apparent to us that we needed a better understanding of the code's functionality before we start the gridification process. As the available documentation on WRF was not targeted for developers like us who would need to modify the code operation to adapt it to a grid computing environment (and not just adding a new physics model, for example), we were pushed to search through lines of the WRF’s FORTRAN and C code to discover how this code was actually functioning; especially, in parts such as domain decomposition and network interactions among the nodes. Due to the large and complex nature of the WRF code, documentation of the program flow proved necessary. With more time and thought, we decided to start a documentation effort to be useful not only for us, but also for other interested in learning the WRF operation in more dept. This guide should help developers understand basic concepts of WRF, how it executes, and how some of its functions branch into different dimensions. We hope that by the time our audience finishes reading this document they will have gained a strong understanding of how WRF operates.

Keywords: Weather Forecast and Reserach, WRF, Fortran, MPI, OpenMP, Grid Computing.

With the advent of mobile, pervasive, and grid computing, software systems must be designed to dynamically adapt to changes that might occur in their runtime environments. Certainly, careful system design and modeling are key factors for systems to be complete. However, as technology changes and new forms of technology continue to emerge, predetermining all possible scenarios in which a system may be running is nothing short of impossible. These issues can be addressed with a tool called TRAP/J (Transparent Reflective Aspect Oriented Programming in Java. However, the first implementation of this tool performs poorly on demanding applications, severely lacked usability, and provided very limited support for adaptation. In this paper, we will be addressing various issues in the first implementation of TRAP/J and we have developed a new version, TRAP/J v2.1, which is aimed at providing better performance and usability over the original TRAP/J. TRAP/J 2.1 is focused on improving the performance of the generation and adaptation phases of Transparent Adaptation and keeping in mind ease of usability. This will allow a decision support system –in our case, a user— to benefit from a user friendly, interactive, web based Composer Interface through which new behavior can be inserted into an application remotely at runtime or startup time. In addition, it has a Generator Interface that allows users to choose which classes they wish to make adaptable.

Keywords: TRAP/J, Dynamic Adaptation, Java, Pervasive Computing, Grid Computing.

Grid computing aims to create an accessible virtual supercomputer by integrating distributed computers to form a parallel infrastructure for processing applications. To enable service-oriented Grid computing, the Grid computing architecture was aligned with the current Web service technologies; thereby, making it possible for Grid applications to be exposed as Web services. The WSRF set of specifications standardized the association of state information withWeb services (WSResource) while providing interfaces for the management of state data. Key to the realization of the benefits of Grid computing is the ability to integrate WS-Resources to create higher-level applications. The Business Process Execution Language (BPEL) is the leading standard for integrating Web services and as such has a natural affinity to the integration of Grid services. In this paper, we share our experience on using BPEL to integrate, create, and manage WS-Resources that implement the factory pattern. We use a Bioinformatics application as a case study to show how BPEL can be used to orchestrate Grid services. The execution environment for our case study comprises the Globus Toolkit as the Grid middleware and the ActiveBPEL as the BPEL engine. To the best of our knowledge, this work is among the handful approaches that successfully use BPEL for orchestrating WSRF-based services and the only one that includes the discovery and management of instances.

Keywords: BPEL, Grid Computing,WSRF, OGSA-DAI, Service Composition.

Grid computing aims to create an accessible virtual supercomputer by integrating distributed computers to form a parallel infrastructure for processing applications. To enable service-oriented Grid computing, the Grid computing architecture was aligned with current Web service technologies. Thereby making it possible for Grid applications to be exposed as Web services. The WSRF set of specifications standardized the association of state information withWeb services (WS- Resource) while providing interfaces for the management of state data. Key to the realization of the benefits of Grid computing is the ability to integrate WS-Resources to create higher-level applications. The Business Process Execution Language (BPEL) is the leading standard for integrating Web services and as such has a natural affinity to the integration of Grid services. In this paper, we share our experience on using BPEL to integrate, create, and manage WS-Resources that implement the factory/instance pattern.

Keywords: BPEL, Grid Computing, WSRF, OGSA-DAI, Grid Service Composition.

High performance computing (HPC) is gaining popularity in solving scientific applications. Using the current programming standards, however, it takes an HPC expert to efficiently take advantage of HPC facilities; a skill that a scientist does not necessarily have. This lack of separation of concerns has resulted in scientific applications with rigid code, which entangles non-functional concerns (i.e., the parallel code) into functional concerns (i.e., the core business logic). Effectively, this tangled code hinders the maintenance and evolution of these applications. In this paper, we introduce Transparent Grid Enabler (TGE) that separates the task of developing the business logic of a scientific application from the task of improving its performance. TGE achieves this goal by integrating two existing software tools, namely, TRAP/J and GRID superscalar. A simple matrix multiplication program is used as a case study to demonstrate the current use and capabilities of TGE.

Keywords: Transparent Grid Enablement, High Performance Computing, TRAP/J, GRID superscalar.

TRAP/BPEL is a framework that adds autonomic behavior into existing BPEL processes automatically and transparently. We define an autonomic BPEL process as a composite Web service that is capable of responding to changes in its execution environment (e.g., a failure in a partner Web service). Unlike other approaches, TRAP/BPEL does not require any manual modifications to the original code of the BPEL processes and there is no need to extend the BPEL language nor its BPEL engine. Furthermore, TRAP/BPEL promotes the reuse of code in BPEL processes as well as in their corresponding autonomic behavior. In this paper, we describe the details of the TRAP/BPEL framework and use a case study to demonstrate the feasibility and effectiveness of our approach.

Keywords: TRAP/BPEL, generic proxy, self-management, dynamic service discovery.

Web services paradigm is allowing applications to electronically interact with one another over the Internet. BPEL facilitates this interaction by providing a platform with which Web services can be integrated. Using RobustBPEL-1, we demonstrated how an aggregate Web service, defined as a BPEL process, can be instrumented automatically to monitor its partner Web services at runtime and replace failed services via a generated proxy. While in the previous work the proxy is statically bound to a limited number of alternative Web services, in this paper we extended the RobustBPEL-1 toolkit to generate a proxy that dynamically discovers and binds to existing services. Further, we present details of the generation process and the architecture of dynamically adaptable BPEL processes and their corresponding dynamic proxies. Finally, we use two case studies to demonstrate how generated dynamic proxies are used to support self-healing and self-optimization (specifically, to improve the faulttolerance and performance) in instrumented BPEL processes.

Keywords: Web service monitoring, BPEL processes, dynamic proxies, self-healing, self-optimization, dynamic service discovery.

Over the years, the advancements in science and technology have led to the increased complexity in computing systems. The systems are thus becoming increasingly more complex with growing number of heterogeneous software and hardware components, increasingly more difficult to monitor, manage and maintain. As a result, it is not a trivial task to provide high performance, high dependability, and high manageability for such computing systems. In this paper, we first present an integrated data-driven architecture for computing system management and then present a case study on Autonomic System Manager: a software system we developed that is currently being used by the system and network administrators of School of Computing and Information Sciences (SCIS) at Florida International University (FIU).

Keywords: Data mining, autonomic computing, self management, self protection, network and system management, anomaly detection.

We recently introduced RobustBPEL [13], a software toolkit that provides a systematic approach to making existing aggregate Web services more tolerant to the failure of their constituent Web services. Using RobustBPEL, we demonstrated how an aggregate Web service, defined as a BPEL process, can be instrumented automatically to monitor its partnerWeb services at runtime and replace failed services via a generated proxy. While in the previous work the proxy is statically bound to a limited number of alternative Web services, in this paper we propose an extension to the RobustBPEL toolkit to generate a proxy that dynamically discovers and binds to existing services. Further, we present details of the generation process, the architecture of the dynamic proxy, and finally use a case study to demonstrate how the generated dynamic proxy is used to support self-healing and self-optimization (specifically, to improve the fault-tolerance and performace) in an instrumented BPEL process.

Keywords: Web service monitoring, BPEL processes, self-healing, self-optimization, dynamic service discovery.

We recently introduced Service Clouds, a distributed infrastructure designed to facilitate rapid prototyping and deployment of autonomic communication services. In this paper, we propose a model that extends Service Clouds to the wireless edge of the Internet. This model, called Mobile Service Clouds, enables dynamic instantiation, composition, configuration, and reconfiguration of services on an overlay network to support self-management in mobile computing. We have implemented a prototype of this model and applied it to the problem of dynamically instantiating and migrating proxy services for mobile hosts. We conducted a case study involving data streaming across a combination of Planet- Lab nodes, local proxies, and wireless hosts. Results are presented demonstrating the effectiveness of the prototype in establishing new proxies and migrating their functionality in response to node failures.

Keywords: autonomic networking, distributed service composition, self-managing system, overlay network, mobile computing, quality of service.

The convergence of data, voice and multimedia communication over digital networks, coupled with continuous improvement in network capacity and reliability has significantly enriched the ways we communicate. However, the stovepipe approach used to develop today’s communication applications and tools results in rigid technology, limited utility, lengthy and costly development cycle, difficulty in integration, and hinders innovation. In this paper, we present a fundamentally different approach, which we call Communication Virtual Machine (CVM) to address these problems. CVM provides a user-centric, modeldriven approach for conceiving, synthesizing and delivering communication solutions across application domains. We argue that CVM represents a far more effective paradigm for engineering communication solutions. The concept, architecture, modeling language, prototypical design and implementation of CVM are discussed.

Keywords: Model driven, communication application, multimedia, middleware, telemedicine.

The development of communication applications today follows a vertical development approach where each application is built on top of low-level network abstractions such as the socket interface. This stovepipe development process is a major inhibitor that drives up the cost of development and slows down the pace of innovation of new generation of communication applications. In this paper, we propose a user-centric communication middleware (UCM) that provides a unified higher-level abstraction for the class of multimedia communication applications. We investigate the minimum set of necessary requirements for this abstraction from the perspective of next-generation communication applications, and provide an API that exemplifies this abstraction. We demonstrate how UCM encapsulates the complexity of network-level communication control and media delivery. Further, we show how its extensible and self-managing design supports dynamic adaptation in response to changes in network conditions and application requirements with negligible overhead. Finally, we argue that UCM enables rapid development of portable communication applications, which can be easily deployed on IP-based networking infrastructure.

Keywords: Multimedia communication applications, user-centric middleware, autonomic computing.

Web services are increasingly being used to expose applications over the Internet. These Web services are being integrated within and across enterprises to create higher function services. BPEL is a workflow language that facilitates this integration. Although both academia and industry acknowledge the need for workflow languages, there are few technical papers focused on BPEL. In this paper, we provide an overview of BPEL and discuss its promises, limitations and challenges.

Keywords: Web services, workflow language, BPEL, business processes, application-to-application integration, and business-to-business integration.

The rapid growth of networking technologies has drastically changed the way we communicate and enabled a wide range of communication applications. However, these applications have been conceived, designed, and developed separately with little or no connection to each other, resulting in a fragmented and incompatible set of technologies and products. Building new communication applications requires a lengthy and costly development cycle, which severely limits the pace of innovation. Current applications are also typically incapable of responding to changes in user communication needs as well as changing network infrastructure and device technology. In this article, we address these issues and present the Unified Communication Model (UCM), a new and user-centric approach for conceiving, generating, and delivering communication applications on-demand. We also introduce a prototype design and implementation of UCM and discuss future research directions toward realizing next generation communication applications.

Keywords: Unified communication model, software generation, on-demand multimedia communication applications, autonomic computing.

Web services are increasingly being used to expose applications over the Internet. To promote efficiency and the reuse of software, these web services are being integrated both within enterprises and across enterprises, creating higher function services. BPEL is a workflow language that can be used to facilitate this integration. Unfortunately, the autonomous nature of web services leaves BPEL processes susceptible to the failures of their constituent services. In this paper, we present a systematic approach to making existing BPEL processes more fault tolerant by monitoring the involved web services at runtime, and by replacing delinquent web services dynamically. To show the feasibility of our approach, we developed a prototype implementation that generates more robust BPEL processes from existing ones automatically. The use of the prototype is demonstrated using an existing loan approval BPEL process.

Keywords: ECommerce, web service monitoring, robust BPEL processes, dynamic service discovery.

In pervasive computing environments, conditions are highly variable and resources are limited. In order to meet the needs of applications, systems must adapt dynamically to changing situations. Since adaptation at one system layer may be insufficient, cross-layer, or vertical approaches to adaptation may be needed. Moreover, adaptation in distributed systems often requires horizontal cooperation among hosts. This cooperation is not only limited to the source and destination(s) of a data stream, but might also include intermediate hosts in an overlay network or mobile ad hoc network. We refer to this combined capability as universal adaptation. We contend that the model defining interaction between adaptive middleware and the operating system is critical to realizing universal adaptation. We explore this hypothesis by evaluating the Kernel-Middleware eXchange (KMX), a specific model for cross-layer, cross-system adaptation. We present the KMX architecture and discuss its potential role in supporting universal adaptation in pervasive computing environments. We then describe a prototype implementation of KMX and show results of an experimental case study in which KMX is used to improve the quality of video streaming to mobile nodes in a hybrid wired-wireless network.

Keywords:

Driven by the emergence of pervasive computing and the increasing need for self-managed systems, many approaches have been proposed for building software that can dynamically adapt to its environment. These adaptations involve not only changes in program flow, but also run-time recomposition of the software itself. We discuss the supporting technologies that enable dynamic recomposition and classify different approaches according to how, when and where recomposition occurs. We also highlight key challenges that need to be addressed to realize the full potential of run-time adaptable software. This survey is intended to be a living document, updated periodically to summarize and classify new contributions to the field. The document is maintained under the RAPIDware project web site, specifically, at http://www.cse.msu.edu/rapidware/survey.

Keywords: adaptive software, compositional adapation, middleware, survey, taxonomy, pervasive computing, autonomic computing, computational reflection, separation of concerns, component-based design, aspect-oriented programming, object-oriented programming.

This paper addresses the energy consumption of forward error correction (FEC) protocols as used to improve quality-of-service (QoS) for wireless computing devices. The paper also characterizes the effect on energy consumption and QoS of the power saving mode in 802.11 wireless local area networks (WLANs). Experiments are described in which FEC-encoded audio streams are multicast to mobile computers across a WLAN. Results of these experiments quantify the tradeoffs between improved QoS, due to FEC, and additional energy consumption caused by receiving and decoding redundant packets. Two different approaches to FEC are compared relative to these metrics. The results of this study enable the development of adaptive software mechanisms that attempt to manage these tradeoffs in the presence of highly dynamic wireless environments.

Keywords: energy consumption, quality-of-service, forward error correction, mobile computing, handheld computer, adaptive middleware

Developing distributed applications is a difficult task due to three major problems: the complexity of programming interprocess communication, the need to support services across heterogeneous platforms, and the need to adapt to changing conditions. Traditional middleware (such as CORBA, DCOM, and Java RMI) addresses the first two problems to some extent through the use of a ``black-box approach, such as encapsulation in object-oriented programming. However, traditional middleware is limited in its ability to support adaptation. To address all the three problems, {\em adaptive} middleware has evolved from traditional middleware. In addition to the object-oriented programming paradigm, adaptive middleware employs several other key technologies including computational reflection, component-based design, aspect-oriented programming, and software design patterns. This survey paper proposes a three-dimensional taxonomy that categorizes different adaptive middleware approaches. Examples of each category are described and compared in detail. Suggestions for future research are also provided.

Keywords: adaptive middleware, taxonomy, computational reflection, component-based design, aspect-oriented programming, software design patterns, static adaptation, dynamic adaptation, quality of service, dependable systems, embedded systems.

Abstract—This paper investigates the energy consumption of forward error correction (FEC) protocols as used to improve quality-of-service (QoS) for wireless handheld devices. Also addressed is the effect on energy consumption and QoS of the power saving mode in 802.11 wireless local area networks (WLANs). Experiments are conducted in which FEC-encoded audio streams are multicast to multiple HP/Compaq iPAQ handheld computers across a WLAN. The results of these experiments help to quantify the tradeoff between improved packet delivery rate, due to FEC, and additional energy consumption caused by receipt and decoding of redundant packets. Moreover, the results enable the development of adaptive software mechanisms that attempt to manage these tradeoffs in the presence of highly dynamic environments.

Keywords: energy consumption, quality-of-service, forward error correction

This paper addresses the design of middleware to support run-time adaptation in pervasive computing environments. The particular problem we address here is how to support adaptation to changing network connection capabilities as a mobile user interacts with heterogeneous elements in a wireless network infrastructure. The goal is to enable adaptation to such changes automatically and, with respect to the core application code, transparently. We propose a solution based the use of generic proxies, which can intercept and process communication requests using rules and actions that can be introduced to the system during execution. To explore their design and operation, we have incorporated generic proxies into ACT [27], a system we designed previously to support adaptation in CORBA applications. Details of ACT-based generic proxies are presented, followed by results of a case study involving adaptation of a surveillance application in a heterogeneous wireless environment.

Keywords: middleware, dynamic adaptation, transparent adaptation, generic proxy, dynamic weaving, quality-of-service, mobile computing

This paper introduces transparent reflective aspect programming (TRAP), a generator framework to support efficient, dynamic, and traceable adaptation in software. TRAP enables adaptive functionality to be added to an existing application without modifying its source code. To reduce overhead, TRAP enables the developer to select, at compile time, a subset of classes to support adaptation through run-time aspect weaving. TRAP uses aspect-oriented programming and behavioral reflection to automatically generate the required aspects and reflective classes associated with the selected types. At run time, new adaptive behavior can be introduced to the application transparently with respect to the original code. TRAP can be applied to any object-oriented language that supports structural reflection. A prototype, TRAP/J, which has been developed for use with Java applications, is described. A case study is presented in which TRAP was used to enable an existing audio-streaming application to operate effectively in a wireless network environment by adapting to changing conditions.

Keywords: dynamic adaptation, aspect-oriented programming, computational reflection, behavioral reflection, adaptive middleware, transparent adaptation, quality-of-service, mobile computing

This paper proposes an Adaptive CORBA Template (ACT), which enables run-time improvements to CORBA applications in response to unanticipated changes in either their functional requirements or their execution environments. ACT enhances CORBA applications by weaving adaptive code into the applications' object request brokers (ORBs) at run time. The woven code intercepts and adapts the requests, replies, and exceptions that pass through the ORBs. ACT itself is language- and ORB-independent. Specifically, ACT can be used to develop an object-oriented framework in any language that supports dynamic loading of code and can be applied to any CORBA ORB that supports portable interceptors. Moreover, ACT can be integrated with other adaptive CORBA frameworks and can be used to support interoperation among otherwise incompatible frameworks. To evaluate the performance and functionality of ACT, we implemented a prototype in Java to support unanticipated adaptation in non-functional concerns, such as quality-of-service and system-resource management. Our experimental results show that the overhead introduced by the ACT infrastructure is negligible, while the adaptations offered are highly flexible.

Keywords: middleware, CORBA, dynamic adaptation, interoperability, request interceptor, dynamic weaving, proxy, quality-of-service, mobile computing

Rapid improvements in mobile computing devices and wireless networks promise to provide a foundation for ubiquitous computing. However, comparable advances are needed in the design of mobile computing applications and supporting middleware. Distributed software must be able to adapt to dynamic situations related to several cross-cutting concerns, including quality-of-service, fault-tolerance, energy management, and security. We previously introduced Adaptive Java, an extension to the Java programming language, which provides language constructs and compiler support for the development of adaptive software. This paper describes the use of Adaptive Java to develop an adaptable communication component called the MetaSocket. MetaSockets are created from existing Java socket classes, but their structure and behavior can be adapted at run time in response to external stimuli. MetaSockets can be used for several distributed computing tasks, including audits of traffic patterns for intrusion detection, adaptive error control on wireless networks, and dynamic energy management for handheld and wearable computers. This paper focuses on the internal architecture and operation of MetaSockets. We describe how their adaptive behavior is implemented using Adaptive Java programming language constructs, as well as how MetaSockets interact with other adaptive components, such as decision makers and event mediators. Results of experiments on a mobile computing testbed demonstrate how MetaSockets respond to dynamic wireless channel conditions in order to improve the quality of interactive audio streams delivered to iPAQ handheld computers.

Keywords: adaptive middleware, reflection, aspect-oriented programming, forward error correction

This paper presents an aspect-oriented approach to dynamic adaptation. A systematic process for defining where, when, and how an adaptation is to be incorporated into an application is presented. A formal model for describing characteristics of dynamically adaptive applications is introduced, which enables precise descriptions of different approaches to dynamic adaptation.

Keywords: aspect-oriented programming, adaptation, security, group communication

Many middleware platforms use computational reflection to support adaptive functionality. Most approaches intertwine the activity of observing behavior (introspection) with the activity of changing behavior (intercession). This paper explores the use of language constructs to separate these parts of reflective functionality. This separation and ``packaging'' of reflective primitives is intended to facilitate the design of correct and consistent adaptive middleware. A prototype implementation is described in which this functionality is realized through extensions to the Java programming language. A case study is described in which metamorphic socket components are created from regular socket classes and used to realize adaptive behavior on wireless network connections.

Keywords: Adaptive middleware, reflection, component design, mobile computing, wireless networks, forward error correction.