Multimedia data are characterized by large, complex values withtemporal dependencies. Such data are used by distributed multimediaapplications offering long-lived multimedia presentations indistributed environments. An end-client may choose to specify qualityrequirements for a multimedia presentation, using quality of service(QoS) parameters known from traditional communication technology. Theclient's demands and requirements should be satisfied throughout theentire presentation, thus QoS management is required. Thesemanagement tasks may be performed by software components residing ateach functional component in a distributed multimedia system, forexample the network or database system. We refer to such a softwarecomponent as a QoS manager, who, among other things, is responsiblefor adjusting the level of QoS prior to and during a multimediapresentation. Whenever the user's requirements regarding QoS are notsatisfied, the adaptation mechanism is responsible for bringing theQoS level within the accepted boundaries.
We focus on adaptation mechanisms for QoS in a multimedia databasemanagement system (MMDBMS) offering services to applications fordistance learning called Learning-on-Demand (LoD). LoD offersinteractive services and a high degree of user interactions.
In the OMODIS (Object-Oriented Modeling and Database Support forDistributed Multimedia Systems) project at the Department ofInformatics at the University of Oslo, a framework for QoS managementhas been developed. The framework defines messages for interactionbetween QoS managers in a dynamic, reconfigurable QoS managementhierarchy. Each manager consists of 7 internal agents, one being aQoS adaptation agent. In our work we propose messages for interactioninside a QoS manager connected to a MMDBMS and a QoS managementhierarchy. To decide the tasks of a manager for a MMDBMS we identify aset of events that may occur in a MMDBMS supporting LoD applications.These events are used to design protocols for internal communicationbetween the agents of a QoS manager for a MMDBMS.
The internal messages are realized through formal specifications inPromela (protocol meta language). The formal specifications are usedto simulate and validate the behavior of the internal message passing.Simulations and conformance testing are used to show that theprotocols as specified, are correct with respect to a set ofcorrectness criteria, and that the protocols support the agents inperforming their QoS management tasks.
Temporal claims and correctness requirements are used duringvalidation. The validation results show that the protocols forinternal communication meet all correctness criteria. In particular,the protocols are free of deadlocks and livelocks. All protocolexecutions end in legal end states, and they all follow our definitionof legal sessions in respect to the session states defined in theOMODIS QoS framework.
In addition, we propose an adaptation model for the QoS adaptationagent of the QoS manager. The model defines behavior for local QoSadaptation prior to and during a presentation. Our adaptation model isbased on QoS monitoring results of the local component. Thus, it iscapable of discovering QoS violations caused by the MMDBMS at anearlier stage than adaptation models depending exclusively uponmonitoring results from the client side. Problems at the client sidewill not arise until an end-to-end QoS violation has occurred.