Stephen A. Uczekaj

Integrated quality of service (QoS) management in service-oriented enterprise architectures

One of the significant challenges for making service-oriented architectures (SOA) effective for enterprise systems is quality of service (QoS) management because of the dynamic, flexible, and compositional nature of SOA. QoS management must be integrated into service-oriented enterprise architectures. It must support a set of common QoS characteristics and provide comprehensive QoS services end to end, from application, to middleware, and to network and from source hosts to destination hosts across a network. We describe such an integrated QoS management architecture and its services. We classify QoS characteristics into four categories and each of which is decomposed into a set of measurable attributes. We integrate these characteristics into an XML-based language for applications and QoS providers to express QoS requirements and contracts. We model an integrated QoS management architecture based on standard specifications from organizations like ISO and OMG. We implement a comprehensive set of QoS management services with innovation resource management techniques and adaptation mechanisms. We provide test data to validate our architecture and solution first in a publish/subscribe style of enterprise SOA. In comparison with other work in QoS management, our architecture and solution provide innovative techniques, extensions, and generalizations beyond traditional task-oriented QoS management in object-oriented middleware and domain specific applications.

Service level management using QoS monitoring, diagnostics, and adaptation for networked enterprise systems

As enterprise services increasingly interconnect as networked services in a service-oriented architecture (SOA), service level management (SLM) is becoming a complex problem and can no longer be handled by traditional monitoring tools like Microsoft SMS. SLM is a process managing the quality of services demanded by clients and offered by providers. This paper presents two contributions to the research of SLM. First, instead of considering monitoring as an isolated service, it incorporates monitoring as an integral part of a comprehensive QoS management framework. This framework consists of QoS management concepts and services including service level contract management, admission control, resource management, monitoring, diagnostics, and adaptation. Using this framework, clients are able to negotiate quality of service contracts with providers and providers are able to optimize system resources to meet contract requirements. The second contribution is the incorporation of diagnostic service in the QoS management framework. Based on data feed from monitoring service, diagnostic service is able to detect any condition changes and to reason about the causes of any degradation conditions in the networked enterprise system. With condition detection and situation understanding, QoS management can then proactively activate adaptation mechanisms to maximize the systems ability to meet QoS contract requirements of concurrent clients. Our monitoring service uses both reporting approach and probing approach to acquire the information of the health status of elements of a networked system. The monitored data is then fed to our diagnostic service to reason about root causes of anomalies, using graphical models. Depending on the system health status and root causes, appropriate adaptations are triggered proactively to improve the system performance under the constraints of concurrent QoS contracts. We validate our SLM approach using QoS management services integrated in a publish/subscribe style of SOA. We then demonstrate via experiments some benefits of QoS monitoring, diagnostics, and adaptation services for responsiveness SLM.

A semantic infosphere

We describe a prototype implementation of a semantic filtering capability added to an existing XML-based publish and subscribe infrastructure. An ontology is used to provide vocabulary for expressing both 1) the semantic annotations that characterize the published documents and 2) the subscriptions specifying the class of documents to be routed to a given client. A description logic (DL) classifier is used to determine which subscribers an incoming document is routed to. We outline the key elements of the ontology for the battlefield domain and give some sample annotations and subscriptions. This is the basis for describing a number of scenarios showing how this filtering capability could be used practice. We critically analyze the suitability of a DL language and reasoner in general, and the particular implementation choices (DAML+OIL, FaCT and OilEd) for performing this task. A key result of the work is to demonstrate the importance of testing semantics-based technologies on practical problems. We discovered a number of new and interesting areas for future work, which in turn can direct the focus of the research community.

A policy-based approach for QoS specification and enforcement in distributed service-oriented architecture

A significant challenge of successful application of the service-oriented architecture (SOA) in large-scale distributed systems is the quality of service (QoS) management, which provides various QoS guarantee levels for concurrent clients through effective resource allocations and adaptations. In this paper, we propose a policy-based approach for specifying QoS management strategies and enforcing QoS guarantees. This approach enables easy adaptation of new business rules and adaptation to system resource changes. This approach is also effective for supporting QoS management, as demonstrated in our experiments in a publish/subscribe system.

Evolution of composition framework in a distributed system toolkit

Network Centric Information Infrastructure (NCII) is a toolkit of prefabricated software components together with a composition framework for integration and experimentation. Typical distributed system applications of a toolkit such as NCH can be applied to air traffic management, shop floor control and in-flight connection to Internet. The composition framework has gone though through three generations of changes in the past five years and it is a good illustration of a paradigm shift from object-oriented architecture, through component-based architecture to service-based architecture.

Dynamic Regeneration of Workflow Specification with Access Control Requirements in MANET

Distributed software systems are the basis for innovative applications. The key for achieving survivable and maintainable distributed systems is agility because the non-deterministic nature of distribution would otherwise leave the system uncontrollable, especially in emerging mobile ad-hoc networks. A mobile ad-hoc network (MANET) is based on a self-organizing and rapidly deployed network of mobile services to collaborate without using any pre-existing fixed network infrastructure. Survivability is defined as the capability of a service to fulfil its mission in a timely manner, even in the presence of attacks, failures, or accidents. There are four key survivability properties: resistance, recognition, recovery and adaptation. Recovery, a hallmark of survivability, is the capability to maintain critical components and resource during attack, limit the extent of damage, and restore full services following attack. Exception handling is a way to deal with the recovery aspect of survivability. Resistance can be viewed as the process of limiting access to critical and vulnerable resources only to authorized users, programs, processes, or other systems. This paper bridges the analysis of secure business process and its recovery aspect in terms of exception handling in the context of access control requirements. We propose an integrated approach to engineer a survivable distributed system through dynamic regeneration of workflow specifications in the context of Business Process Execution Language for Web Services (BPEL) and eXtensible Access Control Markup Language (XACML)

Dynamic Workflow Generation with Interoperable Security Alerts in MANET

Distributed software systems are the basis for innovative applications. Agility is the key for achieving survivable and maintainable distributed systems because the non-deterministic nature of distribution would otherwise leave the system uncontrollable, especially in emerging mobile ad-hoc networks. A mobile ad-hoc network (MANET) is based on a self-organizing and rapidly deployed network of mobile services that collaborate without using any preexisting fixed network infrastructure. Survivability is defined as the capability of a service to fulfill its mission in a timely manner, even in the presence of attacks, threats, or failures. Any failure of a single system may affect a business process; however, the monitoring of many mobile services simultaneously is not an easy task. This paper bridges the analyses of business processes and security threats with an alert mechanism. We propose an integrated approach to engineer a survivable distributed system through dynamic workflow generation in the context of Web Services Business Process Execution Language (WSBPEL), and Really Simple Syndication (RSS).