Abstract: One of the most exciting technologies in today's Internet is the MBone. MBone stands for Multicast Backbone; it provides the infrastructure for efficient multipoint packet delivery in the Internet. The most popular scenario is worldwide audio-video conferencing. However, so far there are no satisfactory solutions available on how to archive multimedia data streams of multicast videoconferences, and how to make them accessible for remote sites (i.e. how to remotely record and remotely play back MBone conferences). In this paper we present architectural considerations, design issues and describe a prototypical implementation of a Video Conference Recording on Demand (VCRoD) service for the MBone. Our system is called MBone VCRoD Service (MVoD) and is a clientserver based architecture for interactive remote recording and playback of MBone sessions. MVoD is implemented based on open standards (e.g. CORBA), making it possible for other applications to interface it. Since the MVoD client application is implemented using JAVA, it will be possible to access the service from almost any platform.

Abstract: We present a soft real-time CPU server for Continuous Media processing in the UNIX environment The server is a daemon process from which applications can request and acquire soft real-time QoS (Quality of Service) Guarantees Our server is an extension of the URsched scheduler It provides (1) protection among real-time(RT) processes (2) fairness among RT and non-RT processes, (3) rate monotonic scheduling, (4) a fix to the UNIX security problem We have implemented our protocol in the SUN Solaris 25 Operating System, and we have shown through experiments that our soft RT server provides predictable QoS for continuous media applications We also discuss how we will fit the real-time server into our general Resource Broker Architecture in our future work

Abstract: This paper presents a system, called mTunnel, for application level tunneling of multicast traffic in a lightweight manner, where the end-user is responsible for deciding which MBone-sessions and multicast groups to tunnel. mTunnel is primarily designed for easy deployment and easy-to-manage tunneling. Therefore it runs as an user application and does not need access to restricted system resources.

Abstract: Their small size and ease of use makes Personal Digital Assistants (PDAs) very attractive for inclusion as clients in mobile multimedia applications. Alas, the cost for integration of PDA in distributed multimedia systems is high, as the resource limitations on PDA require client adaptation, and as software development on PDA is comparatively expensive. We present PocketWeb, a tool facilitating PDA integration in multimedia applications based on the web protocols while freeing application developers from PDA specific software development. We have applied PocketWeb for integration of PDA as multimedia clients in MILLION, a distributed multimedia system developed in an ESPRIT project.

Abstract: Multimedia presentations are applicable in various domains such as advertising, commercial presentations or education. If multimedia documents which describe multimedia presentations can be accessed on-line via different network types and be presented on various types of terminals different amounts of resources may be available at presentation time. Hence, it can happen that there are not enough resources to render a multimedia document according to the specification. For usual multimedia documents resource scarcity implies an arbitrary reduced presentation quality. To handle resource scarcity in a better way, multimedia documents can be made adaptable to different resource situations. The presented temporal model provides abstractions to specify multimedia documents with alternative presentation parts. Further on, the presentation behavior of media objects can vary within specified limits. Hence, the presented temporal model allows to compose presentations which have a defined behavior when resource restrictions occur and a resource scarcity need not result in an arbitrary reduced presentation quality.

Abstract: Synchronous collaborative work environments, which are mainly based on video-conferencing systems, suffer a lack of human communication channels and social awareness because mostly only audio, video, and joint editing of documents are supported. Collaborative services such as floor control, session control, and telepointers provide additional communication mechanisms to support persons co-working through computers. We propose a collaborative services model (CSM) for distributed learning environments. Our object-oriented approach includes floor and session control mechanisms and policies. Due to the realization of collaborative services as operations on the model, no specific network collaboration protocol is required. We present an optimistic synchronization scheme which provides consistency for the distributed model.

Abstract: Multimedia applications require continuous delivery of time-dependent data. In best-effort systems where no resource reservation is possible, adaptation strategies are necessary in case of high system load. We propose an adaptation framework for a client-pull architecture where continuous data are requested and buffered by the presenting client. The current client buffer utilization is used as an indicator for performance bottlenecks and determines which adaptation strategy the client invokes. The aim of the invoked adaptation is to reduce the data volume which has to be transferred to the client in a controlled manner leading to a graceful degradation of presentation quality. Frequent switches between different qualities are avoided by using a hysteresis and by specifying smooth transitions between different degrees of data dropping. As we consider adaptation control only at the client side we believe that our approach can easily be used for achieving intra-media synchronized presentations in most open and heterogeneous environments.

Abstract: Because of stateless characteristics of http, it is not possible for http servers to adapt responses on the basis of previous interactions with the same user. Although efficient, this is a limiting aspect of the protocol, particularly for Web documents that are directed toward a broad audience, including experts and novices and an adaptive response would be helpful in providing information with the right style.

Abstract: Many new distributed multimedia applications involve dynamic multiple participants, have stringent end-to-end delay requirement and consume large amount of network resources. In this paper, we propose a distributed delay-constrained dynamic multicast routing algorithm (DCDMR) to support these applications. DCDMR scales well because the source of the multicast tree needs only limited computation or may even not be involved in the route computation. When group membership changes, the existing multicast tree is perturbed as little as possible and the average resulting tree cost is very satisfactory.

Abstract: The growing number of digital audio and video repositories has resulted in a desperate need for effective techniques to deliver data to users in a timely manner. Due to geographical distribution of users, it is not cost effective to have a centralized media server. In this paper, we investigate issues involved in the design of a distributed video server (DVS) to support movie-on-demand (MOD) application. We propose a redundant hierarchical (RedHi) architecture for DVS where the nodes are continuous media servers and the edges are dedicated network lines. With RedHi, each node has two or more parents. We show that the redundant links in RedHi yield a more reliable and efficient system. Our simulation results demonstrate that RedHi can tolerate a single link failure with no degradation in performance while with pure hierarchy almost 2.5% of requests are rejected due to the failure. In normal mode of operation, RedHi outperforms pure hierarchy significantly (160% improvement on the average when counting the number of rejections). In the context of RedHi, we also propose and evaluate alternative object management policies, and load balancing heuristics.

Abstract: Recent advances on computing and communication technologies have made it possible to provide digital video services through network. One of the economic and practical applications is near video-on-demand systems. In this paper, a novel data placement scheme on optical discs is proposed for providing near video-on-demand services. By applying the scheme, the seek time of an optical disc drive is almost eliminated, and the bandwidth utilization can be optimized. Then, the maximal number of supported streams from an optical disc drive with the minimal buffer requirement can be easily achieved. The scheme can be applied not only to CDROMs storing MPEG-1, but also to Digital Versatile Disks (DVDs) storing MPEG-2 in the near future.

Abstract: We describe three disk scheduling algorithms that can be used in a multimedia server for sustaining a number of heterogeneous variable-rate data streams. A data stream is supported by repeatedly fetching a block of data from the storage device and storing it in a corresponding buffer. For each of the disk scheduling algorithms we give necessary and sufficient conditions for avoiding under- and overflow of the buffers. In addition, the algorithms are compared with respect to buffer requirements as well as average response times.

Abstract: This paper analyses a tree hierarchical network architecture employing pooled cache memories and multicasting network capability to deliver Interactive Video on Demand (I-VoD) services. The cache memories are located close to the customer and are able to store segments of the most popular programs in a dynamic way. They can reduce both transmission bandwidth and storage capacity required in the network, thus reinforcing the substantial network resource savings achievable through the multicasting feature allowed in the ATM switches. In the present paper, we analyse this type of VoD network with the aim to provide optimal dimensioning of the cache memories and to evaluate the load reduction produced in the servers and in the links as a function of the main system parameters (cache memory dimensions, number and popularity of programs, number of user clusters involved, etc.).

Abstract: Today, Forward Error-Correcting (FEC) codes are mainly implemented in hardware, and many believe that their complexity prohibit their software implementation. This paper presents in detail how the performances of a software implementantion can be significantly improved. Different levels of optimization which are independent of the working environment are presented and discussed. The coding throughput of 100 Mbps on an UltraSparc 1 shows that FEC codes can be easily added to multimedia applications without requiring dedicated hardware support. As a case study, we use FEC codes to protect AAL5-PDUs from cell losses in ATM networks.

Abstract: With the evolving availability of wireless communication services and of affordable mobile devices such as notebooks or Personal Digital Assistants, mobile computing is becoming widely accepted and applied. It is one further step towards the vision of information access for anyone, anytime, anywhere. Now, mobile multimedia applications are needed to make the vision more real. However, because of the narrow bandwidth of wireless wide-area networks and the limited resources of mobile devices in comparison to stationary systems, the handling of multimedia data faces severe problems. This leads to a need for effective solutions that enable the interactive handling with multimedia services even over a wireless link.

Abstract: There is a convergence between the idea of design patterns and the goal to achieve reuse of component design in the Telecommunications Information Networking Architecture (TINA). Patterns can be specified using graphical and syntactical notations proposed by the TINA Consortium (TINA-C) for the information and computational viewpoint languages of the Open Distributed Processing Reference Model (ODP-RM). In this article four design patterns are proposed that efficiently support Computer Supported Collaborative Work (CSCW) applications in TINA: voting, floor control, role management and group integrity criteria. The design is done with the goal of high configurability in mind. It is planned to implement the patterns to achieve an extension of a prototype of a TINA multimedia communication service. This service is thus to evolve to a collaboration service. Eventually, the implementation of the patterns should become part of an object-oriented framework for TINA applications.

Abstract: ATM, the asynchronous transfer mode, was developed to meet two main goals. Firstly, it provides a flexible transmission technology. Secondly, it offers a set of new services to the user [Part94]. In order to take advantage of ATM's strengths, one might want to use its services directly. To be able to do so, in particular to be able to integrate access to ATM services into own applications, one is in need of a powerful ATM application programming interface (API). In this paper, we first present the design and implementation of such an ATM API for Digital Alpha workstations running Digital UNIX. Moreover, we describe how the basic ATM service is enhanced by setting an implementation of SandiaXTP (Xpress Transport Protocol) on top of it. Finally, we give an overview over a monitor for this ATM environment designed to compute quality of service (QoS) parameters from protocol information, detect QoS violations by comparing them to desired values, and display the results graphically.

Abstract: Multimedia presentations are applicable in various domains such as advertising, commercial presentations or education. If multimedia documents which describe multimedia presentations can be accessed on-line via different network types and be presented on various types of terminals different amounts of resources may be available at presentation time. Hence, it can happen that there are not enough resources to render a multimedia document according to the specification. For usual multimedia documents resource scarcity implies an arbitrary reduced presentation quality. To handle resource scarcity in a better way, multimedia documents can be made adaptable to different resource situations. The presented temporal model provides abstractions to specify multimedia documents with alternative presentation parts. Further on, the presentation behavior of media objects can vary within specified limits. Hence, the presented temporal model allows to compose presentations which have a defined behavior when resource restrictions occur and a resource scarcity need not result in an arbitrary reduced presentation quality.

Abstract: It has been widely recognized that QoS is important in designing distributed, interactive multimedia systems. However, although much has been said about QoS in the networking domain, and to a lesser degree in processor scheduling, there has hardly been any research activity concerning QoS in the user interface community. We show how the use of 3D graphics techniques in the user interface can lead to a natural way of implicitly specifying and presenting QoS to the end-user. Further, we introduce the concept of generic QoS control tools that allow users to explicitly control and monitor quality of service across different media types. In order to support such a novel user interface, a mechanism is needed to communicate these quality requirements to lower level system components. Although there exists a plethora of QoS architectures that define the semantics and interface of every component in many different ways, none offers a definitive way of structuring QoS aware systems. We instead propose a generalized, abstract concept of QoS for all layers of a software architecture. Each layer in a software system deals with QoS at its appropriate level of abstraction using a generic API for communicating QoS parameters and values to layers above and below. We call the aggregation of these parameters and values a “service contract”. This abstract concept can be applied recursively to build hierarchies of services. This paper describes our framework for building QoS aware software systems and explains in detail the user interface for a multimedia browser as an example application.

Abstract: Distributed multimedia systems and advanced network and Internet technologies enable novel application areas such as teleteaching. This paper first presents an experimental distributed infrastructure for such teleteaching applications and discusses recent experiences with various distributed lectures and exercises. In addition to key technological issues such as multimedia information management and distribution, aspects of user interaction, content provision, and acceptance of given solutions are also discussed.

Abstract: One of the most important standardization efforts in the broad field of audiovisual coding is MPEG-4 that introduces objects as smallest accessible units inside frames. This creates new possibilities in scalable scene compositions where more interesting foreground objects (e. g. a news speaker) can be lossy encoded with a better quality than a less interesting background object. This leads to new degrees of freedom in achieving high compression rates. Furthermore, this new audiovisual coding paradigm provides new functionalities in interactive audiovisual presentations. In this paper, we present a prototype realization of a video-on-demand application that enables distributed event handling of object-related events created by a user. Our system consists of a client/server architecture, where the client side handles events like zooming or moving single video objects at presentation time, while the server side considers user control concerning the quality of a lossy encoding of the single video objects. The term object-oriented stands for the access of objects rather than frames in our coding scheme as well as the object-oriented design of our software system by means of aggregation, association or inheritance. This paper describes the architecture of our system by explaining how the desired distributed functionality can be achieved. The software is entirely written in the Java language. Therefore, platform independence is achieved by using the Java Virtual Machine. Java's multithreading capabilities are used to design a highly extendible and flexible system that can be dynamically configured at runtime. The overall functionality of the system is explained with OMT class diagrams, block diagrams and state diagrams describing a protocol exchange by client and server, including the influence of the user interaction on the quality of the transmitted video objects.

Abstract: A main feature of the next generation's computing software is supporting convenient user interactions. Interactive multimedia presentations that are executed in distributed environments raise some new computer communication behaviors, which we call “interactive multimedia communications”. In this paper, we (1) classify user interaction types, (2) point out the main problems in interactive multimedia communications, (3) distinguish two approaches for solving these main problems, (4) identify the possible control architectures for processing user interactions, and (5) propose some methods and control schemes accordingly.

Abstract: This paper presents the approach of the ESPRIT project HyNoDe (EP 22160 - Hypermedia News on Demand) towards a novel Personalised News-On-Demand (PNoD) Service. The HyNoDe Service Personalised feature refers to customisation of the service to suit individual needs and preferences of the clients based on the information filtering concept. The service addresses all the involved phases in the news publishing industry, from News Capturing up to end-user News Presentation. The paper compares the HyNoDe service with existing ones, and presents the architecture of the service, and the HyNoDe modules and tools related with each of the news publishing phases.

Abstract: Application Output Recording (called AOR from now on) is the task of recording the output of an application so that the reproduction of the recording is sufficiently similar to the original output performance. The paper proposes a taxonomy to structure the task and describes a generic architecture covering the complete range of possible complexities in a recorder. Variables determining the usability potential, technical aspects and design decisions are defined. The usage of AOR is described in the context of the European Union funded ACTS-project DIANE (Design, Implementation and Operation of a Distributed Annotation Environment), which implements a generic distributed multimedia annotation environment running on broadband systems. A usage scenario clarifies the utilisation of AOR in DIANE. DIANE is evaluated in a telemedicine and a tutorial environment. Special requirements can be derived from these and have implications to the process of recording generic applications. A short overview on the first implementation of AOR in DIANE is given. The taxonomy is applied to solutions from related work as well in research as in the commercial sector.

Abstract: An efficient conference service facilitates the implementation and the run-time control of conference applications. Key features of a conference service are conference management, multicast communication support, application state synchronization, and user data marshalling. The paper compares the features offered by the T.120 standards and the CORBA distribution platform. Since they complement each other, it is discussed how a conference service based on a combination of T. 120 standards and CORBA can be realized.

Abstract: Distributed Multimedia Applications have high Quality-of-Service (QoS) requirements regarding throughput, delay, and delay jitter. Current communication systems do not provide QoS to the applications even in the cases where the network technology may provide QoS guarantees. This paper describes a Internet audio/video conferencing application on top of a native ATM communication system. The UDP/IP communication system has been substituted by AAL5/ATM. The paper describes how the applications can now benefit from the QoS capabilities of ATM and the required modifications of the audio/video conferencing applications in order to use the native ATM service.

Abstract: An electronic product catalogue (EPC) is a computer controlled information system with multimedia product presentations and navigation facilities. The paper presents the results of the EPK-fix project for the systematic construction of EPCs. These include a software engineering process model, a high level specification language for EPCs, and an integrated set of tools supporting the entire EPC development process. The EPK-fix process model supports the classical development phases: requirements analysis, specification and design, implementation, and test. In each of the phases emphasis is put on the particular multimedia aspects, human machine interaction, production of prototypes, and the quality of the produced documents.

Abstract: The speed of Internet access over GSM can be improved by optimizing the performance of TCP/IP connections over the air link, and by using smart proxies for HTTP transfers or other high-level protocols The goal is to improve the performance as perceived by the user, which implies the minimization of the end-to-end latency This document describes the setup of GSM data connections and points to a number of possibilities to improve its basic performance It then takes a more detailed look into WWW access and discusses how a double HTTP proxy can improve the web access over a (very) slow link This double proxy is made of two parts: a small and simple HTTP proxy on the mobile station and a more complex proxy serving several users from a specific Internet Cellular Access Gateway The users on the mobile station can configure their usual WWW browser to use the proxy on their computer — the mobile proxy — for accessing Web pages and everything will work as before, except that the overall performance should be considerably better

Abstract: A stable clock indication is needed for the processing and playout of MPEG-2 streams. ATM networks can support the transfer of such streams but contribute significantly (as well as other system operations) to delay and delay variation. The MPEG-2 standard specifies strong temporal constraints for processing and presentation. MPEG-2 is also based on the assumption of a common time base shared among sender and receiver. Clock synchronization covers this requirement and is traditionally performed using PLLs. Unfortunately, their linear nature cannot simultaneously guarantee rapidity and good clock reconstruction. The non-linear technique we propose in this paper is able to offer such guarantees and also can handle higher amounts of delay variations. Such technique is not restricted to the scope of MPEG-2 clock synchronization and can also succeed in the reconstruction of time bases for generic distributed applications. It does not require any specific device, like very precise oscillators or filters.