Session Mobility and Quality of Service Specifications for Distributed Multimedia Applications

7
Session Mobility and Quality of Service Specifications for Distributed Multimedia Applications Edward Babulak, Member, IEEE, Member, ACM, Abstract - In this paper we present a detailed review on the Session Mobility and Quality of Service (QoS) Specifications for Distributed Multimedia (MM) Applications. Enhanced Trader lists the resources necessary to support the application’s QoS requirements and makes them available for the Session Manager, which co-ordinates the session switching. Our work demonstrates that all the components of distributed MM system should be involved in the QoS management process. The client’s requirements related to QoS must be adapted to various constraints supported by the distributed MM system components: client machines, server machines and transport system. Our studies in contrast to the existing ones present a comprehensive review on the Reflective Systems with relation to the session switching. Index Terms – Quality of Service (QoS), Common Request Broker Architecture (CORBA), Enhanced Trading Service, Session Mobility (SM) and Multimedia (MM). I. INTRODUCTION Increasing demands for interoperability and QoS on current MM communications encourages developers to look for new solutions and technologies. CORBA is well established industry standard and offers number of services including trading service, which constitutes founding mechanism for our approach to QoS provision within the distributed heterogeneous platforms. Emerging standards for distributed integrated environments, like CORBA from the Object Management Group (OMG) provide a software bus for a distributed MM applications [1]. Open system standards are important for the integration of distributed MM system support in order to achieve unified communications architecture. The QoS provides a unifying theme on which the functionality and facilities of new integrated standards can be constructed. For future applications, especially those, which are highly interactive and rely on MM information, it is essential that QoS be guaranteed for the whole system. Enhanced communications protocol support such as end- to-end QoS negotiation, renegotiations, degradation and co- ordination over multiple connections is required. Design of distributed MM applications, such as system for access to remote MM databases or teleconferencing, requires careful consideration of QoS issues, due to the networking bandwidth and processing power required by the presentation quality of live media, such as video. For applications running in shared environment, the allocation and management of these resources is very important. In general, the best-effort approaches offered by most existing systems are not suitable for distributed MM, because some users may be ready to pay some higher price for obtaining maximum quality, while others may prefer low-cost presentations with lower quality. We briefly review the trading service and define the QoS specifications essential for our research. This paper is structured in four sections. The first section presents introduction, motivation and the QoS problem statement. Second section presents proposed solution, the use of trading service, application scenario fundamentals, adaptive MM architecture, and the user interface SM. The section three and four presents enhanced trader’s specifications, locating the service and display, enhanced trader’s QoS specifications. Finally the last section presents conclusions. A. Motivation Today's mission-critical distributed systems in military, government or business applications have a specific functional and performance requirements being reflected by string QoS constrains [2], [3] and [4]. MM communications research explores human-computer interaction by making use of text, images, graphs, animation, sound, speech, and video [5] and [6]. When combined, developing distributed QoS-constrained MM applications (e.g., Audio/Video Conferencing) is a difficult task because of the strict requirements on system resources and the great diversity of MM standards and devices [7], [8] and [9]. QoS is becoming increasingly important in distributed MM systems [10] and [11]. The nature of distributed MM applications is such that they require peer-to-peer communication support mechanisms [12] and [13]. MM traffic needs to be delivered to end-systems, networks and end-users in a form that they can handle while satisfying the constraints imposed by the applications [14], [15] and [16]. The QoS mechanisms are required to ensure high quality media playout at high-performance workstations while at the same time providing appropriately filtered lower quality media playout at low-end systems, i.e. adaptive QoS [17] and [18]. Existing support mechanisms such as multicast protocols are performance deficient. They provide the service based on the least capable link or node involved in the multicast session [19], [20] and [21]. B. QoS Problem Statement. QoS is not a new concept; traditionally it is employed within the domain of computer networks to specify a set of parameters assigned to transport connections. As a rule, QoS is established through negotiation between service users and providers. The process of negotiation is simple if the resources are managed by a centralized entity (e.g. an operating system) or by a set of homogeneous entities. Unfortunately, in distributed MM applications the negotiation and management of resources is a difficult task PGNET 2000 1 ST ANNUAL SYMPOSIUM ON THE CONVERGENCE OF TELECOMMUNICATIONS , NETWORKING AND BROADCASTING.

Transcript of Session Mobility and Quality of Service Specifications for Distributed Multimedia Applications

Session Mobility and Quality of Service Specificationsfor Distributed Multimedia Applications

Edward Babulak, Member, IEEE, Member, ACM,

Abstract - In this paper we present a detailed review on the SessionMobility and Quality of Service (QoS) Specifications for DistributedMultimedia (MM) Applications. Enhanced Trader lists the resourcesnecessary to support the application’s QoS requirements and makesthem available for the Session Manager, which co-ordinates thesession switching. Our work demonstrates that all the components ofdistributed MM system should be involved in the QoS managementprocess. The client’s requirements related to QoS must be adapted tovarious constraints supported by the distributed MM systemcomponents: client machines, server machines and transport system.Our studies in contrast to the existing ones present a comprehensivereview on the Reflective Systems with relation to the sessionswitching.

Index Terms – Quality of Service (QoS), Common Request BrokerArchitecture (CORBA), Enhanced Trading Service, Session Mobility(SM) and Multimedia (MM).

I. INTRODUCTION

Increasing demands for interoperability and QoS oncurrent MM communications encourages developers to lookfor new solutions and technologies. CORBA is wellestablished industry standard and offers number of servicesincluding trading service, which constitutes foundingmechanism for our approach to QoS provision within thedistributed heterogeneous platforms. Emerging standards fordistributed integrated environments, like CORBA from theObject Management Group (OMG) provide a software busfor a distributed MM applications [1]. Open systemstandards are important for the integration of distributedMM system support in order to achieve unifiedcommunications architecture. The QoS provides a unifyingtheme on which the functionality and facilities of newintegrated standards can be constructed. For futureapplications, especially those, which are highly interactiveand rely on MM information, it is essential that QoS beguaranteed for the whole system.

Enhanced communications protocol support such as end-to-end QoS negotiation, renegotiations, degradation and co-ordination over multiple connections is required. Design ofdistributed MM applications, such as system for access toremote MM databases or teleconferencing, requires carefulconsideration of QoS issues, due to the networkingbandwidth and processing power required by thepresentation quality of live media, such as video.

For applications running in shared environment, theallocation and management of these resources is veryimportant. In general, the best-effort approaches offered bymost existing systems are not suitable for distributed MM,because some users may be ready to pay some higher pricefor obtaining maximum quality, while others may preferlow-cost presentations with lower quality. We briefly review

the trading service and define the QoS specificationsessential for our research.

This paper is structured in four sections. The first sectionpresents introduction, motivation and the QoS problemstatement. Second section presents proposed solution, theuse of trading service, application scenario fundamentals,adaptive MM architecture, and the user interface SM. Thesection three and four presents enhanced trader’sspecifications, locating the service and display, enhancedtrader’s QoS specifications. Finally the last section presentsconclusions.

A. Motivation

Today's mission-critical distributed systems in military,government or business applications have a specificfunctional and performance requirements being reflected bystring QoS constrains [2], [3] and [4]. MM communicationsresearch explores human-computer interaction by makinguse of text, images, graphs, animation, sound, speech, andvideo [5] and [6]. When combined, developing distributedQoS-constrained MM applications (e.g., Audio/VideoConferencing) is a difficult task because of the strictrequirements on system resources and the great diversity ofMM standards and devices [7], [8] and [9]. QoS is becomingincreasingly important in distributed MM systems [10] and[11]. The nature of distributed MM applications is such thatthey require peer-to-peer communication supportmechanisms [12] and [13]. MM traffic needs to be deliveredto end-systems, networks and end-users in a form that theycan handle while satisfying the constraints imposed by theapplications [14], [15] and [16].

The QoS mechanisms are required to ensure high qualitymedia playout at high-performance workstations while atthe same time providing appropriately filtered lower qualitymedia playout at low-end systems, i.e. adaptive QoS [17]and [18]. Existing support mechanisms such as multicastprotocols are performance deficient. They provide theservice based on the least capable link or node involved inthe multicast session [19], [20] and [21].

B. QoS Problem Statement.

QoS is not a new concept; traditionally it is employedwithin the domain of computer networks to specify a set ofparameters assigned to transport connections. As a rule, QoSis established through negotiation between service users andproviders. The process of negotiation is simple if theresources are managed by a centralized entity (e.g. anoperating system) or by a set of homogeneous entities.Unfortunately, in distributed MM applications thenegotiation and management of resources is a difficult task

PGNET 2000 1ST ANNUAL SYMPOSIUM ON THE CONVERGENCE OF TELECOMMUNICATIONS , NETWORKING AND BROADCASTING.

since resources are very diverse, dispersed and maintainedby heterogeneous entities. Much attention has been paid tothe use of CORBA Trader at the initial stage of our work,but less to the use of more general trading mechanism thatwould relate to a SM function, QoS re-negotiation and QoSadaptability.

II. PROPOSED SOLUTION

Our work is motivated by the challenge of providing adynamic adaptation of QoS parameters during the MMapplication set-up. In order to satisfy user’s QoS needs wemay apply trading services1 to allocate resources necessary[22]. The trading service would provide resources necessaryfor optimum QoS connection2. The customers subscribethrough their account for specific QoS parameters.According to the account privileges, the client will beallocated devices and network resources via proposed QoSProvision and SM mechanisms. An enhanced TradingService will provide reference of access points to networkresources and end-devices necessary for a QoS provision.

A. The Use of Trading Service

The growth of computer networks and transmissioncapacity facilitates an increased amount of service offers indistributed systems [23], [24] and [25]. One of the CORBAservices is the trading service, which supports clients insearching for suitable services. To support a client insearching for a special service, a trader can be used. Thistrader has a service directory, which contains availableservices specified by service type and service propertyvalues. If a client requests a service, the trader procures aconvenient one [26]. In case of MM application, CORBAcould provide access to the remote objects requested by theapplication [27]. The selection of objects andcommunication resource allocation must be optimized forthe desired QoS. Trading problem in particular could beapproached from three different angles:• clients searching for the suitable service,• integrating different local trader functions in order to

extend service access support beyond local networkboundaries, or

• The service that would take into the consideration theQoS aspects dynamically. Because of the growingquality needs of many communication services, a tradershould be able to allow service request with QoSaspects.

Trading could be useful in underlying infrastructures. Inother words, the QoS will reflect on what we want ascompared to what we get and how that relates to the totalresource capacity. A trader is a third party object thatenables clients to find information about suitable services

1 Trading service represents the process of arbitration of services.2 Providing we have “high bandwidth” ATM connection it makes senseto make use of high quality end-devices, such as high-end terminals,workstations, video and audio devices. In case of having a very poorconnection available, the use of high quality end-device may notimprove the quality of service expected.

and servers. Services provided by service providers can beadvertised, or exported, in a trader. Such advertisements,known as service offers, are stored in the trader’s database.The advertising object (the service provider or object actingon behalf of the service provider) is called an exporter.Potential service users (importers) can import the trader whomeans obtaining information on available service and theiraccessibility. The trader tries to match the importer’srequest against the service offers in its database. After asuccessful match the importer can interact with the serviceprovider. The relationships between trader, importer,exporter, and administrator are illustrated in Figure 1.

Figure 1. Trader Function

B. Application Scenarios Fundamentals

This section presents possible application scenariosillustrating the use of trading service, to allow for the SM.We introduce the Server and Client side applicationscenarios where sessions are switched either due to theserver failure or due to the client initiated request. In ourapproach we extend work on QoS Negotiation andAdaptation [5]. We define a framework, which allows forthe resource allocation necessary to support the applicationQoS requirements across distributed heterogeneousplatforms. While current work in MM SM has usually beenconcerned with maintaining the network connection of amobile clients, it is believed that session hand-off betweenterminal devices or service providers is essential as well.

SM in general is the ability to simultaneously utilizemultiple display devices allowing for a new applicationscenarios that will ease the restrictions of session playbackon a portable MM device. On the other hand, the ability totransparently re-bind a MM service provider, allows thesystem to scale well with respect to the number of mobileclients, and will result in better utilization of the board-bandnetwork infrastructure of the future. MM SM can be broadlyclassified into:• Mobility caused by the actual movement of a device,

and• Mobility caused by session transfer from one device to

another.

Most current work in MM falls into the first category, andis usually restricted to maintaining the network connectivityof a mobile device. However, we believe that the user ratherthan the mobile device is central to MM SM, which can evenbe achieved without the physical movement of any device.SM between display devices may be necessary because ofthe user preference or physical limitations of portablemobile devices, while SM between remote service providersmay be needed to balance the utilization of the networkinfrastructure and improve scalability. The limitations ofportable MM device arise primarily because of its smallform factor. For instance, the physical dimensions of aportable display device will never be the same as those of alarge screen. In the scenario of MM presentation or video-on-demand where the slide show has been prepared on aportable device, it is beneficial to move the session to anavailable large screen in order for it to be better visible tolarge audience (see Figure 2). Clearly, it is often desirable toutilize alternate display devices during a mobile MMsession.

Figure 2. SM Scenario for display switching.

We may expect some performance gains in a system withportable MM devices, where selecting a media service basedon proximity to the device will result in reduced latency, andpossibly cost. From the system point of view, utilizingnearby media services will also better utilize the network,resulting in a system which scales well with increasingnumber of mobile users.

Furthermore, such media service rebinding may benecessary because the changes of a display device requiresdifferent service parameters. It is clearly an advantage, andoften necessary, to change media services during a long-running mobile MM session. Both, MM content providers,and publicly accessible display devices are essential in orderto fully realize the potential of portable mobile MM (seeFigure 3).

C. Adaptive MM Architecture

The increasing diversity of available hardware and istallednetworks makes it imperative that MM applications adapt toavailable resources. Furthermore, application adaptivenessmust be dynamic in order to adjust to an increasinglyvariable availability of resources.

The principal pattern of interaction between componentsin adaptive MM architecture for MM application software

is a service contract between a service provider and a serviceconsumer.

Figure 3. Reflection on MM SM.

This pattern provides a powerful framework todecompose a MM application into a hierarchy of services.Components higher up in the hierarchy are associated withmore abstract notions, resulting in the hiding of details andtherefore, a simpler application. The same service may beprovided by more that one server, and service providers canbe changed in order to maintain a service contract. In thisarchitecture, vertical interaction betweeen components atdifferent levels of the hierarchy can be considered controlrelated, while interaction between components at the samelevel, typically the lowest, is efficient data movement.

Consider a video-on-demand application shown in Figure4. At the highest layer, the application is composed of a userinterface components (UIC) and media service component(MSC). Both the UIC and MSC play the role of a serviceprovider to the top layer of the application, which maintainscontracts with each of them. The user in turn utilizes adisplay and an audio playback device. The contract betweenthe user interface and the application is decomposed intocontracts with the audio device and the video display.

Similarly, the media service component decomposes itscontract with the application into contracts utilizing networkresources and video servers. Although the primarymotivation for this architecture was to solve the problem ofdecomposing an abstract notion of QoS into low-level QoSparameters, the architecture can also form the basis fordistributed components infrastructure for a MM application.

Details of specific service providers can be effectivelyhidden from higher layers of the application. For instance,the top layer of the video-on-demand application in Figure4, need not to know the existence of video serveralternatives or which alternative is being currently used.This knowledge is hidden by the intermediate media service,which can transparently choose a different video serverwhen conditions make this necessary. In the figure the redarrow indicated devices for display, audio and video serverare represent those currently being used.

The need for dynamically changing service provider canbe seen in the following scenario. A mobile user, who wantsto redirect a media session from the Liquid Crystal Display

(LCD) display of his palm-top computer to a locallyavailable large screen television.

Figure 4. Architecture of Video-on-Demand Application

In contrast to most other work in session redirection,redirection as a re-binding to a different display service is anovelty. Typically, the capability of the new display isdetermined, and is used to adjust existing contractsthroughout the system. For instance, the new display mayhave the capability of displaying better quality video track(eg. HDTV) from the video server. The current server maynot support the desired format, and the media service mustchoose a different server which can provide the video streamin the desired format. Changing the display provider in thisapplication has included changes throughout the system,requiring a change in even the video service provider.Described architecture of hierarchical contracts is apowerful basis for building adaptiveness into a MMapplication. The architecture is generic, unifying theadaptiveness (mobility) necessary for QoS, resourcevariability, and changes in service providers.

Figure 5. MM Capabilities.

D. User Interface SM.

It is a common practice that session transfer from onedisplay device to another is be user initiated for manyreasons including security and privacy. It is important toconsider user-interface design issues when building a

system that supports MM session migration. The importantfactors to be considered are:• The herogeneity of alternative display devices;• Making the user aware of devices in his proximity;• Initiating session transfer;• Preserving the ability to control the session.

The user-interface model, represents a MM sessionconsisting of a set of application objects, each with its owncapabilities for sending and receiving media data. A displaydevice is also associated with the capabilities for presentingmedia data. For example, a movie object is capable of bothaudio and viceo output, while a pair of headphones iscapable only of presenting the audio output. A capability canbe represented as a plane (see Figure 5), with a region on theplane representing the range of quality which can besupported. A device is considerd capable of presenting thedata associated with a media object if the device and objectcapabilities intersects. Furthermore, the presentation of amedia object on a particular device is constrained to therange of quality determined by the intersection of theirrespective capabilities.

When a session containing a movie object, which has theinherent capabilities of video and audio output, istransferred to a loud-speaker, obviously only the audio trackof that movie will be heard.

IV. ENHANCED TRADER’S SPECIFICATIONS

Trading Service advertises the state of network objects.Let us consider network elements QoS, the applicationselects the appropriate resources for execution. QoS couldbe considered as a usage of resources whereby applying theappropriate control mechanism, trading service selects theappropriate resource that would satisfy the userrequirements (e.g., user should express the QoSrequirements). Let us assume the scenario where we wish toprovide a fast printing without taking care of the QoS. If weadvertise the printer through the trading service, we willselect one of the advertised printers as fast. Or similarly wemay register two services, if we have two mechanisms totransport data, one over TCP and one over UDP (user datapacket). In case that we start the application that requires thefast delivery then the trader would be able to select betweenthe data transport resources available (i.e., TCP or UDP). Inorder to access the information about the network resourceswe could store the data in a local MIB (managementinformation base).

In case we may want to reserve bandwidth and at the sametime we want to ensure the QoS in this way, we mayconsider to provide the gateway to the RSVP, which shouldbe activated in case of providing the appropriate QoSparameters. The selection of the RSVP would be enablethrough the use of trading service. The mechanism thatwould provide a control of trader is in fact a program, whichis involved with the trader. This program, running as aCORBA object or having a CORBA gateway, is advertisedto the trader. Trader must provide following answers:• Why? Service user must find service provider

at run-time.

• What? Trading service provides match makingbetween objects. Service Users consult the TradingService to obtain information about suitable service.Client & Server are the Application Programs. Serverhas objects with specific methods that are registeredwith trader through a Service Offer. Client doesn’tknow what Server has to offer, so Trader has a list ofObject Services Contained in Server

In case we want to chose between ATM and Fast Ethernetor Ethernet, we should have the control program into theswitch. If we would have for example a backbone and thenwe can use FDDI, Ethernet, and ATM Networks, we wouldhave the server object into the router (similar to RSVP,which is installed in on the router and the client runs on thehost). Then we can use as selection criteria not only the typeof the link but also the state.

The principal operations involve the offer export, import,and final process of binding. In such architectures, eachservice provider first exports its availability to a tradingservice. A client imports a provider based on thespecifications of a required service, which is indicated in thestandard specification format. URN's3 can be considered anattempt towards such a universal format for locationindependent names. A trading service usually also providesa directory service to determine the available service of aparticualr type.

The existence of multiple service providers or devicesresult in a heterogeneous environment in which it becomesnecessary to determine a service or device capability. It isbelieved that while a user may desire to retain control overthe session transfer between terminal devices, he would berelatively uniterested in controlling session transfer betweenservers. Server session transfer should therefore be hiddenfrom the user, and the triggered by changes to globalparameters such as relative server load and network load, orimplicitly by the users mobility. System support is necessaryto trigger such server transfer.

A. Locating a Service and Display

Locating a nearby terminal device is substly different fromlocating a nearby server in that the physical proximity of aterminal device may be an insufficient criterion. Forinstance, a display screen that is few meters away is uselessin a situation where a user, and the group he wants to sharethe information with, are located behind the screen. On theother hand, the same screen may be useful if the user intendsdisplaying information to communicate with large audienceseated in front of the large screen. Clearly the orientation ofthe display screen with respect to the user is an importantcriterion.

There are two approaches to locating the nearby devicesand services: using the physical characteristics of a wirelesstechnology to determine proximity, and using a directoryservice for which the physical location of the client is aparameter. It is believed that neither of these approaches issufficiently general for locating display devices andservices. For instance, a directory service is unsuitable to

3 R. Moats. RFC 2141: URN syntax, May 1997.

locate an appropriate display device because it may beimpossible to determine if the device is visible from themobile users location, or whether it had been occluded by atemporary partition.

Infra-red technology, which may be suitable to advertisethe location and orientation of such a display device, wouldbe unsuitable to advertise the presence of a nearby printerwhich may be in a printer-room next door.

B. Enhanced Trader’s QoS Specifications

In order to design a system where multiple applicationscoexist within a QoS management framework, it isnecessary for them to either have a common understandingof how QoS should be specified, or to be able to map theirindividual specifications into a common one. To support theintegrated resource management that will be required for theefficient and dependable operation of future distributedsystems, research is needed in the areas of application-levelQoS, application-specific QoS trade-offs and resourcemanagement algorithms.

A considerable amount of work is in progress in the areaof providing QoS guarantees mover high-speed networks[28], [29] and [30]. The ongoing work mainly concentrateson the problem of bandwidth management and switch-basedscheduling to provide deterministic or statistical guaranteeson end-to-end delay, throughput, and packet losses. Thesolutions proposed in this area are valuable in ensuringcertain QoS for network traffic. However, another aspect ofQoS management, namely, the problem of end-system QoSmanagement is also important. It is not sufficient to ensurethat the network traffic is delivered with desired QoS acrossa path through the network, but it is also essential tosupplement the network QoS with mechanisms that ensurethat data can be delivered (and processed) in a timelyfashion across the data path inside the end-system. The data(and control) path inside the end-system connects thenetwork interface with the source (or sink) of networkcommunication such as MM application running in the userspace. To guarantee end-system QoS support one mustconsider contention for resources such as network interface,CPU (time), memory, and bus bandwidth.

Compare to the research in the area of network-level QoSguarantees, relatively little work has been done in the area ofend-system architectures for QoS management [31] and[32]. Currently, general-purpose operating system (OS) suchas Unix do not include support for QoS specifications orinclude policies and mechanisms to provide predictableservice and guaranteed access to end-system resources. Thetraditional OS model of ensuring fair, on-demand sharing ofresources is inadequate to provide such predictive QoSguarantees. In our work we are currently investigating a newarchitecture for managing resources within the distributedheterogeneous system via trader. The architecture provides acommon framework for managing resources, such asVideoServers, Client’s Display Monitors, and CPU, networkinterface, bandwidth taking into account the adaptive anddynamic nature of MM applications.

The principal task of trader is to provide the resourceallocation. For the purpose of QoS provision we must define

clearly the service parameters at both the application and thenetwork level. The Application-level QoS parameters aredependent on Network-level QoS or Server-level QoS.

In our studies assume that:

• The Session Manager is machine independent (i.e.independent of Operating System, Video Player, etc.);

• Network is bottleneck free;• if not the we would have consider additional

network delay,• Jitter (i.e. delay variation) is introduced by network and

physical distance between the Client and Server.• We must consider both Network-level and Application-

level QoS parameters.• The distributed heterogeneous environment may

introduce the requirement of registering more that oneVideo Server, as well as creating the federation oftraders depending on the location and the availability ofthe resources necessary for the application specifiedQoS.

IV. CONCLUSIONS

The objective of this research was to address the problemof QoS Provision and the SM for the CORBA-basedDistributed MM across Heterogeneous Platforms viaenhanced Trader.

Major contribution of our research is in providing detailedstudies on the concept of the Session Mobility and the QoSSpecification for Distributed Multimedia. We allow fordifferent servers and clients to run on different platformsconnected through inter-connected networks. Result of ourresearch leads to further analytical studies on systemperformance measures and efforts to solve the optimum QoSprovision, based on the network resource availability and theapplications-defined QoS requirements.

ACKNOWLEDGMENTS

Author is grateful for guidance given to him during hisvisit at the University of Ottawa. All including, Dr. N.D.Georganas the Thesis Director, Dr. Ionescu the Thesis Co-Director and Dr. Groza have contributed to this research.Author is currently a full-time Faculty at the Department ofElectrical & Computer Engineering at PennState Erie at theBehrend College. Author wishes to express his sinceregratitude to Dr. R. Simoneau, the Director of School ofEngineering & Engineering Technology for his guidanceand motivation, to Dr. R. Ford, In Charge of Department ofElectrical & Computer Engineering for his support andsupervision. Finally, my gratitude goes to my colleaguesincluding, Dr. R. Weissbach for his encouragement and Dr.Chris Coulston for his critical review.

REFERENCES

[1] Object Management Group “CORBA 2.0Specification”, OMG Technical DocumentPTC/96-03-04, USA, March 1996.

[2] C. M. Adam, M.C. Chan, A. A. Lazar, J.-F. Huard andK. S. Lim, “The Binding Interface Base SpecificationRevision 2.0”, OPENSIG Workshop on Open Signalingfor ATM, Internet and Mobile Networks, The MollerCentre, Cambridge, UK, April 17-18, 1997.

[3] C. Aurrecoechea, A. Campbell, and L. Hauw. “Surveyof QoS Architectures.” MM Systems Journal,Special Issue on QoS Architectures, 1997.

[4] N.D.Georganas, "MM ApplicationsDevelopment: Experiences", J.MM Tools andApplications, Vol.44, No. 3, May 1997, pp.313-332.

[5] G.V. Bochmann, B. Kerherve, A. Hafid, P. Dini andA. Pons, Functional Design of Adaptive Systems. InProceedings of the IEEE Workshop on MMSoftware Development, Berlin, Germany 1996.

[6] G.v.Bochmann, R.Dssouli, J.Gecsei and B.KerhervePosters on: ” QoS Negotiation andAdaptation: Project Overview “ The Third Annual ITRhttp://www.iro.umontreal.ca/labs/citr/posters.html,29 Sep 1995.

[7] Masaki, S., et al. “ Personal MM-multipointteleconferencing system for broadband ISDN”, 3rd IFIPWG6.4 Conf. On High Speed Networking, BerlinGermany (March 18-22, 1991) pp.263-278).

[8] D. Anderson, R. Herwitch and C. Schaefer, SRP: Aresource Reservation Protocol for GuaranteedPerformance Communications in the Internet, TheInternational Computer Science Institute, Berkley, 991.

[9] Ahuja, S., Ensor, J. and Horn, D.N. “The RapportMM conferencing system” Conf. On OfficeInfo. Syst. Palo Alto, CA., March, 1988, pp. 1-8.

[10] R. Steinmetz and K. Nahrstedt, MM:Computing, Communications and Applications,Prentice Hall, 1995.

[11] Towards an Architecture for Distributed MMApplications Support; Proceedings of theInternational Conference on MM Computingand Systems, Boston, Massachusetts, May 1994.

[12] P. J. Kuhn, “ QoS in ATM Networks”,Slides, 1995.

[13] ISO/IEC JTC1/SC21/N9309: Open SystemInterconnection,Data management and OpenDistributed Processing, QoS, BasicFramework,Working Draft, January 1995.

[14] J. T. Schuilenga, “ QoS Monitoring and Modelling ofContinuos Streams in a Distributed Environment”,M.Sc. Thesis KPN Research/ University of Twente,April 1995.

[15] Steinmetz, R., and Meyer, T., “Modeling distributedMM applications”, Int. Workshop on AdvancedCommunications and Applications for High SpeedNetworks, Munich, Germany (March 16-19, 1992).

[16] A. Campbell, G. Coulson and D. Hutchison, A Qualityof Service Architecture, ACM SIGCOM ComputerCommunication Review, Vol. 24, No. 2, pp. 6-27, April1994.

[17] K. Raymond, “Stream and Qos: a White Paper”,Contribution to OMG Telecommunications SpecialInterest Group (TELSIG), December 1995.

[18] Zhonghua Yang and Keith Duddy, "Distributed Object

Computing with CORBA", DSTC TechnicalReport 23, June, 1995.

[19] S. Deering. Multicast Routing in a DatagramInternetwork. Ph.D. thesis, Stanford University, 1991

[20] N. Shacham. Multipoint communication behierarchically encoded data. In IEEE Infocom’92,pages 2107-2114, 1992.

[21] J.M.S. Doar “Multicast in the Asynchronous TransferMode Environment”, PhD Dissertation, St. John’sCollege, University of Cambridge, January 1993.

[22] QoS – Framework, ISO/IEC CD13236.2, October 1995.

[23] Schill, A., Zitterbart, M.: A System Framework forOpen Distributed Processing; Journal of Network andSystems Management, Vol. 1, No.1, pp. 71-93, 1993.

[24] Spaniol, O., Popien, C., Meyer, B.: Services andService Trading in Client/Server Systems (in German).Thomson Publishing GmBH, 1994.

[25] P.G. Bosco (CSELT), E. Dahle (Telenor), M. Gien(Chorus), Grace (BritishTelecom), N. Mercouroff, N.Perdigues (Alcatel Alshtom Recherche), J.B.Stefani(France Telecom) “The ReTINA Project: AnOverview”, ReTINA Technical Report 96-15, 1996.

[26] D. Thißen, C. Linnhoff-Popien, “Finding OptimalService Within a CORBA Trader,” to appear.

[27] Lazar, A. A. "Programming TelecommunicationNetworks", IEEE Network, September/October 1997,pp. 2-12. This paper is a Revised Version of thekeynote address given by the author at theInternational Workshop on QoS,Columbia University, New York, May 21-23, 1997.Published in the workshop proceedings (pp. 3-23).

[28] Tawbi, W., and Horlait, E., “Video Coding and Qualityof Service”, MASI Internal Report, University of ParisVI, France (1992).

[29] "The xbind Project at Columbia University", Lazar,A.A., Web Site, at http://comet.ctr.columbia.edu/xbind

[30] "The Binding Interface Base", Lazar, A.A., Lim, K.S.and Marconcini, F., CTR Technical Report# 412-95-18, Center for TelecommunicationsResearch, Columbia University, New York,June 1995.

[31] Popien, C., Schürman, G., Weiß, K.-H.: DistributedProcessing in Open Systems (in German). Teubner, 1996.

[32] Andrea V. et al. “Distributed MM and QoS: ASurvey,” IEEE MM, summer 1995.

Edward Babulakreceived the HighNational Diploma inElectronics in 1976 andthe B.Sc. Degree inElectrical Engineeringin 1982, both from theSlovak TechnicalUniversity. Edward

worked 10 years in industry as a project manager, consultantand a system engineer. He received a High NationalCertificate in Computer Aided Engineering from BrightonCollege of Technology, U.K. in 1990 and a Master ofScience Degree in Computer Science from the University ofEast London, U.K. in 1991. Edward is currently a part-timegraduate student at the University of Ottawa, working on thecompletion of his doctoral thesis in Electrical Engineering.He has been a faculty member in Electrical and ComputerEngineering at Penn State Erie since August 1999 and holdsthe rank of a lecturer. His research interests are in the areasof QoS Provision for a CORBA-based MM applications,distributed heterogeneous systems and parallel computing.He is a member of the IEEE, the ACM, the AmericanMathematical Society and Mathematical Association ofAmerica.