QoS-Aware Service Composition

Florian Rosenberg (CSIRO ICT Centre)

CSIRO ICT

DATE: 2010-05-06
TIME: 14:00:00 - 15:00:00
LOCATION: Bld. 108, S206 (CSIRO Seminar Room)
CONTACT: JavaScript must be enabled to display this email address.

ABSTRACT:
Business processes in enterprise applications are often implemented by composing a set of services into so-called composite services. This process is commonly referred to as service composition. An efficient execution and a continuous provision of these services is crucial to meet the desired non-functional requirements. Quality of Service (QoS) attributes are commonly used as a means to describe such non-functional properties of a service. This includes both domain- agnostic attributes such as performance- and dependability and domain-specific QoS attributes, such as voice or traffic data rate in the telecommunication domain. In enterprise systems, such composite services frequently need to be adapted to reflect changing business and QoS requirements. This requires a high degree of adaptability that needs to be supported by the underlying composition infrastructure. This talks presents some of the models and methods to adapt such composite services at runtime. First, we propose efficient optimization algorithms for QoS-aware compositions that can then dynamically select the ``best'' service to the fulfill composite service request. Second, we propose an approach for domain- specific service selection that can be adapted at runtime without any disruption of the infrastructure. Such an approach is especially suited for scenarios where a global optimization is not required because the number of available services small enough.
BIO:
Florian Rosenberg is a Research Scientist in the CSIRO ICT Centre, Australia since November 2009. He received his PhD in Computer Science from the Technical University Vienna in June 2009. Previously, he was a university assistant at the Distributed Systems Group, Technical University Vienna from 12/2005 to 10/2009. He was actively involved in different research projects in the area of service- oriented computing. He also worked as a technical co-op at the IBM T.J. Watson Research Center in New York, USA.