Vinod Muthusamy

A distributed service-oriented architecture for business process execution

The Business Process Execution Language (BPEL) standardizes the development of composite enterprise applications that make use of software components exposed as Web services. BPEL processes are currently executed by a centralized orchestration engine, in which issues such as scalability, platform heterogeneity, and division across administrative domains can be difficult to manage. We propose a distributed agent-based orchestration engine in which several lightweight agents execute a portion of the original business process and collaborate in order to execute the complete process. The complete set of standard BPEL activities are supported, and the transformations of several BPEL activities to the agent-based architecture are described. Evaluations of an implementation of this architecture demonstrate that agent-based execution scales better than a non-distributed approach, with at least 70% and 120% improvements in process execution time, and throughput, respectively, even with a large number of concurrent process instances. In addition, the distributed architecture successfully executes large processes that are shown to be infeasible to execute with a nondistributed engine.

Distributed automatic service composition in large-scale systems

Automatic service composition is an active research area in the field of service computing. This paper presents a distributed approach to automatically discover a composition of services based on the desired input to and output from the process. The algorithm makes use of the content-based publish/subscribe model, with service inputs modeled as subscriptions, and outputs as advertisements. Service interfaces are mapped to publish/subscribe messages in such a way that publish/subscribe matching is used to evaluate service compatibility. In this way, large-scale distributed service composition and process discovery is achieved with a distributed publish/subscribe network. Evaluations in a distributed environment of a real implementation of the system demonstrate the scalability of the distributed approach, especially with respect to the number of services, the complexity of the discovered processes, and the number of concurrent searches.

Adaptive content-based routing in general overlay topologies

This paper develops content-based publish/subscribe algorithms to support general overlay topologies, as opposed to traditional acyclic or tree-based topologies. Among other benefits, publication routes can adapt to dynamic conditions by choosing among alternate routing paths, and composite events can be detected at optimal points in the network. The algorithms are implemented in the PADRES publish/ subscribe system and evaluated in a controlled local environment and a wide-area PlanetLab deployment. Atomic subscription notification delivery time improves by 20% in a well connected network, and composite subscriptions can be processed with 80% less network traffic and notifications delivered with about half the end to end delay.

Content-based routing in mobile ad hoc networks

The publish/subscribe model of communication provides sender/receiver decoupling and selective information dissemination that is appropriate for mobile environments characterized by scarce resources and a lack of fixed infrastructure. We propose and evaluate three content-based routing protocols: CBR is an adaptation of existing distributed publish/subscribe protocols for wired networks, FT-CBR extends CBR to provide fault-tolerance, and RAFT-CBR provides both fault-tolerance and reliability. Using network simulations we analyze the applicability and test the tradeoffs of these algorithms. We show that RAFT-CBR can guarantee 100% delivery to small groups, at the expense of transmission delay. CBR, with a low message overhead and low delay, is more suitable for larger groups at the expense of reliability. FT-CBR provides comparable delivery rates to RAFT-CBR, as well as low delay, at the expense of increased message cost.

Effects of routing computations in content-based routing networks with mobile data sources

This paper presents the first quantitative evaluation of the role of routing computations on performance when mobility is introduced to a content-based routing network. Additionally, the paper identifies the factors that affect the performance of a distributed publish/subscribe architecture supporting mobile publishers, formalizes publisher mobility protocols for distributed publish/subscribe systems, and develops and evaluates protocols that reduce the costs associated with supporting mobile publishers in publish/subscribe systems. Our results show that ignoring route computation time paints a false picture of the scalability of content-based routing networks, but that with appropriate protocols the adverse effects can be mitigated.

Historic data access in publish/subscribe

We develop a content-based publish/subscribe platform, called PADRES, which is a distributed middleware platform with features inspired by the requirements of workflow management and business process execution. These features constitute original additions to publish/subscribe systems and include an expressive subscription language, historic, query-based data access, composite subscription processing, a rule-based matching and routing mechanism, and the support for the decentralized execution of service-oriented applications.

SLA-driven business process management in SOA

The management of non-functional goals, or Service Level Agreements (SLA), in the development of business processes in a Service Oriented Architecture (SOA) often requires much manual and error-prone effort by all parties throughout the entire lifecycle of the processes. The formal specification of SLAs into development tools can simplify some of this effort. In particular, the runtime provisioning and monitoring of processes can be achieved by an autonomic system that adapts to changing conditions to maintain the SLAs goals. SOA supports partitioning a system into services that are running in a distributed execution environment. When coupled with an associated cost model, a process can be both executed and monitored in an optimal manner, based on a declarative, user-specified optimality function.

Efficient event-based resource discovery

The ability to find services or resources that satisfy some criteria is an important aspect of distributed systems. This paper presents an event-based architecture to support more dynamic discovery scenarios, including efficient discovery of resources whose attributes can change, and continuous monitoring for resources that satisfy a set of constraints. Furthermore, algorithms are developed to optimize the discovery cost by reusing results among similar concurrent discovery requests. Detailed evaluations under various workload distributions demonstrate the feasibility of the architecture and show significant benefits of the optimizations in terms of network traffic and discovery processing time.

Publisher mobility in distributed publish/subscribe systems

The decoupling of producers and consumers in the publish/subscribe paradigm lends itself well to the support of mobile users who roam about the environment with intermittent network connectivity. This paper presents the first quantitative evaluation of publisher mobility in a distributed publish/subscribe system. Our results indicate that publisher mobility breaks a fundamental assumption of publish/subscribe systems and has a significant performance impact. We formalize publisher mobility algorithms for a distributed publish/subscribe system, and develop and evaluate optimizations to the mobile publisher algorithms.

Automating SLA modeling

Service Level Agreements (SLAs) define the level of service that a service provider must deliver. An SLA is a contract between service provider and consumer, and includes appropriate actions to be taken upon violation of the contractual obligations. However, implementing an SLA using existing IT infrastructure is difficult, requiring a lot of manual effort to translate an SLA into code, model it with the given programming language, and ensure the required monitoring support is available for efficient monitoring and tracking of the SLAs. In this paper, we present a solution for modeling an SLA contract. It is designed to be configurable, reusable, extensible and inheritable, thus providing great flexibility to construct complex SLAs. We also introduce an algorithmic generation pattern to create the necessary artifacts to implement an SLA presented in this paper. The resulting artifacts automatically monitor a business process and evaluate whether the SLA is violated during runtime execution. The proposed approach is designed to require minimal human intervention.