QoSME: QoS Management Environment

Florissi, PatrĂ­cia Gomes Soares

Distributed multimedia applications are sensitive to the Quality of Service (QoS) delivered by underlying communication networks. For example, a video conference exchange can be very sensitive to the effective network throughput. Network jitter can greatly disrupt a speech stream. The main question this thesis addresses is how to adapt multimedia applications to the QoS delivered by the network and vice versa. Such adaptation is especially important because current networks are unable to assure the QoS required by applications and the latter is usually unprepared for periods of QoS degradation. This work introduces the QoS Management Environment (QoSME) that provides mechanisms for such adaptation. The main contributions of this thesis are: Language level abstractions for QoS management. The Quality Assurance Language (QuAL) in QoSME enables the specification of how to allocate, monitor, analyze, and adapt to delivered QoS. Applications can express in QuAL their QoS needs and how to handle potential violations. Automatic QoS monitoring. QoSME automatically generates the instrumentation to monitor QoS when applications use QuAL constructs. The QoSME runtime scrutinizes interactions among applications, transport protocols, and Operating Systems (OS) and collects in QoS Management Information Bases (MIBs) statistics on the QoS delivered. Integration of QoS and standard network management. A Simple Network Management Protocol (SNMP) agent embedded in QoSME provides QoS MIB access to SNMP managers. The latter can use this feature to monitor end-to-end QoS delivery and adapt network resource allocation and operations accordingly. A partial prototype of QoSME has been released for public access. It runs on SunOS 4.3 and Solaris 2.3 and supports communication on ATM adaptation layer, ST-II, UDP/IP, TCP/IP, and Unix internal protocols.



More About This Work

Academic Units
Computer Science
Department of Computer Science, Columbia University
Columbia University Computer Science Technical Reports, CUCS-036-95
Published Here
February 10, 2012