Showing papers by "Jon Crowcroft published in 1998"

Recommendations on Queue Management and Congestion Avoidance in the Internet

Abstract: This memo presents two recommendations to the Internet community concerning measures to improve and preserve Internet performance. It presents a strong recommendation for testing, standardization, and widespread deployment of active queue management in routers, to improve the performance of today's Internet. It also urges a concerted effort of research, measurement, and ultimate deployment of router mechanisms to protect the Internet from flows that are not sufficiently responsive to congestion notification.

TCP-like congestion control for layered multicast data transfer

Abstract: We present a novel congestion control algorithm suitable for use with cumulative, layered data streams in the MBone. Our algorithm behaves similarly to TCP congestion control algorithms, and shares bandwidth fairly with other instances of the protocol and with TCP flows. It is entirely receiver driven and requires no per-receiver status at the sender, in order to scale to large numbers of receivers. It relies on standard functionalities of multicast routers, and is suitable for continuous stream and reliable bulk data transfer. In the paper we illustrate the algorithm, characterize its response to losses both analytically and by simulations, and analyse its behaviour using simulations and experiments in real networks. We also show how error recovery can be dealt with independently from congestion control by using FEC techniques, so as to provide reliable bulk data transfer.

Differentiated end-to-end Internet services using a weighted proportional fair sharing TCP

Abstract: In this document we study the application of weighted proportional fairness to data flows in the Internet. We let the users set the weights of their connections in order to maximise the utility they get from the network. When combined with a pricing scheme where connections are billed by weight and time, such a system is known to maximise the total utility of the network. Our study case is a national Web cache server connected to long distance links. We propose two ways of weighting TCP connections by manipulating some parameters of the protocol and present results from simulations and prototypes. We finally discuss how proportional fairness could be used to implement an Internet with differentiated services.

A case for dynamic sender-based reservations in the Internet

Abstract: In this paper we discuss the need for resource reservation in the Internet and examine some of the strengths and weaknesses of RSVP, which is currently the most popular of Internet reservation protocols that have been developed. We also discuss some alternative reservation protocols for packet networks, in particular the ATM Block Transfer (ABT) reservation protocol that has been designed for use in Asynchronous Transfer Mode (ATM) networks and which uses in-line control packets to modify reservations on the fly to achieve very efficient bandwidth utilisation. Finally, we present a proposal for a new reservation protocol, known as DRP (Dynamic Reservation Protocol) which combines many of the strengths of RSVP and ABT with few of the weaknesses to achieve a highly bandwidth-efficient reservation mechanism with excellent scalability with regards to round trip time, data rate and number of hosts.

A Dynamic Sender-Initiated Reservation Protocol for the Internet

Abstract: In this paper we discuss the need for resource reservation in the Internet and examine some of the strengths and weaknesses of RSVP, which is currently the most popular of Internet reservation protocols that have been developed. The deficiencies of RSVP motivate our design of a new resource reservation protocol which uses dynamic sender-initiated reservations to achieve a highly bandwidth-efficient reservation mechanism with excellent scalability with regards to round trip time, data rate and number of hosts.

CRAM: cell re-labeling at merge-points for ATM multicast

Abstract: Many distributed multimedia applications involve data delivery from a source to multiple destinations, the participating nodes forming a multicast group. In the naive solution, separate connections can be established from each source to other group members. However a tree can be established for each source with the participants as the leaf nodes or just have one tree spanning all the participants. In this paper, we introduce a data forwarding model to support such shared multicast trees over the ATM networks called CRAM (cell re-labeling at merge-points for ATM multicast). Our work allows VC merging in the MPLS architecture and supports recently proposed wide area multicast protocols (like CBT and PIM) in ATM networks.

End to end reliable multicast transport protocol requirements for collaborative multimedia systems

Abstract: Multi-party collaborative multimedia applications require data to be transmitted reliably and efficiently in order to provide a guaranteed quality of service (QoS). The multimedia applications can vary from distributed games and shared whiteboards to interactive video conferencing. These applications often involve a large number of participants and are interactive in nature, with participants dynamically joining and leaving the applications. In order to provide many-to-many interaction when the number of participants is large, IP multicasting is a very good option for communication. IP multicasting provides scalability and efficient routing but does not provide the reliability that these multimedia applications may require. Though a lot of research has been done on reliable multicast transport protocols, it really seems that the only way of doing a reliable multicast is to build it for a given purpose like conference control in multimedia conferencing. This paper compares some of the available multicast transport protocols and analyses the most suitable features and functionalities provided by these protocols for a facet of conference control: floor control. The goal is to find or design a reliable multicast transport protocol which would scale to tens or hundreds of participants scattered across the Internet and which would deliver the control messages reliably.

General conference control protocol

Abstract: This paper describes a protocol by which people communicating using different conference control mechanisms can be seamlessly integrated into a single conference control mechanism. A generic approach to conference control, General Conference Control Protocol (GCCP) is proposed to support and bind formal/tightly coupled multimedia conference (ITU standardised) with the informal/loosely coupled conference (IETF based). These two types of conferences are used by two different types of community and the innate differences in the underlying architecturc of IETF and ITU based conferencing makes it difficult to interoperate. The differences and the similarities of these two architectures are explored and a protocol is proposed in this paper which will interoperate between two different types of conferences. A single generic conference control system covers a range of services provided by different conference controls. Participants in different kind of conferencing with different kind of underlying architecture can use a particular type of service(s) from generic conference control which will integrate them into a single mechanism. The appropriate and relevant features of existing conference control mechanisms have been used to design a tool for conference management.

Traffic Jams on the Internet

Abstract: We find the basic premise of the report “Social dilemmas and Internet congestion” by Bernardo Huberman and Rajan Lukose (25 July, [p. 535][1]) quite intriguing. Social forces such as those at work in the well-known “tragedy of the commons” ([1][2]) would indeed influence patterns of some

Differentiated End-to-End Internet Services using a Weighted Proportional Fair Sharing TCP

Abstract: In this document we study the application of weighted proportional fairness to data flows in the Internet. We let the users set the weights of their connections in order to maximise the utility they get from the network. When combined with a pricing scheme where connections are billed by weight and time, such a system is known to maximise the total utility of the network. Our study case is a national Web cache server connected to long distance links. We propose two ways of weighting TCP connections by manipulating some parameters of the protocol and present results from simulations and prototypes. We finally discuss how proportional fairness could be used to implement an Internet with differentiated services.

Performance comparison of CRAM, SEAM and SPAM multipoint VC schemes for ATM networks

Abstract: Multicast service is an important part of any modern routing architecture. Motivations for many-to-many multicast include general unpredictability of membership in many real applications; simplicity of the rendezvous for the application programmer; and low cost to end systems and switches in terms of state to maintain for the delivery tree. Shared trees have an even greater advantage over source based trees in the latter respect. Shared trees (as in the CBT model for Internet) are supported in the form of a single logical VC per multicast group (i.e., multipoint-to-multipoint VC or mp-mp VC) in the ATM networks. CRAM, SEAM and SPAM have been previously proposed for supporting mp-mp VC. The work in this paper does a performance comparison of these three schemes, with respect to buffer requirements, end-to-end delay, packet jitter and traffic overhead. Our evaluation is carried out through extensive simulations with different topologies and sender traffic types.

Static Internet multicast

Abstract: IP multicasting (Cheriton and Deering, 1985) appears to achieve a level of simplicity by extending the IP unicast addressing model (historically the class A, B, and C net numbers) from the mask and longest match schemes of CIDR (Braun et al. 1993), with a new classful address space, class D. The routing systems have been also built in a deceptively simple way in one of three manners-either broadcast and prune, destination list based tree computation or single centered trees. The multicast service creates the illusion of a spectrum that one can tune in to, as an application writer. Thus, many have seen the multicast pilot service, Mbone, as a worldwide Ethernet, where simple distributed algorithms can be used to allocate "wavelengths" and advertise them through "broadcasting" on a channel associated with a spectrum. These three pieces of the picture have tempted people to construct a distributed architecture for a number of next level services that cannot work at more than a modest scale, since they ignore the basic spirit of location independence for senders and receivers of IP packets. The problem is that many of these services are attempting to group activities at source, when it is only at join time that user grouping becomes apparent. This memo proposes approaches to solve some current multicast problems rather statically with DNS (Postel 1994) and the URL based approach, and avoid the pitfalls of trying to use address allocation to implement traffic aggregation for different sources or aggregation of multicast route policy control through control of such aggregated sources.

HYbrid Monitoring and Networks Simulation - HYMNS

Abstract: This note is about some of the technical work that UCL CS people have done, partly Jon at HP's Internet Research Institute during the first quarter of 1998, but also previous work using NS under several projects, notably MERCI, HIPAPRCH and DARPA work. The work described is a series of separate efforts to add protocols to NS, and the design effort to enhance NS to provide interworking between real and simulated (actual and virtual) systems.

Java performance for multimedia applications

Abstract: Java is a relatively new, cross-platform, object-oriented language, which has rapidly evolved into an entire operating environment, encompassing most real-world programming requirements. Traditionally, its interpreted nature has been regarded as a severe restriction against real-time application implementation. However, constant performance improvements with subsequent releases of the language, as well as its technical advantages in networking and multithreading support, have recently allowed its use for multimedia delivery by a variety of sources. In this text, presented are experiences on relevant work carried out during the last three years, almost immediately after Java’s public release, at the Department of Computer Science, University College London (UCL-CS). Java’s suitability for (soft) real-time media is verified, and current issues that need to be addressed, including workarounds for common implementation problems, are also examined. In addition, a performance-oriented case study, based on audio transmission over packet networks, an area where UCL-CS enjoys worldwide recognition, is included. The advent of higher speed access (e.g., 33.6-56 Kbps modems, but also xDSL, cable modems, etc.), combined with wide backbone connectivity and progress in software (applications and protocols), proved in practice the suitability of packet networks to carry multimedia (audio, graphics, video, etc.) information. During recent years, there has been considerable work on this field, with a significant number of relevant tools seeing publicity ([5] , [15] , [34] ). Most of these applications are implemented in C and/or C++, for performance and functionality reasons. Recently, however, a number of efforts to develop multimedia applications using a relative newcomer, Java, have been recorded ([1] , [8] , [14] , [17] , [30] ). The Department of Computer Science at University College London (UCL-CS) has long-term research and development experience in Networked Multimedia, and in particular audio and video conferencing, with several previous and current projects in the area ([35] ). With the advent of the Java programming language and its rapid evolution as the preferred platform for many, significant effort has been put into investigating its applicability for multimedia applications, with emphasis being given to audio ([1] , [8] ).

End-to-End Reliable Multicast Transport Protocol Adaptation for Floor Control and Other Conference Control Functions Requirements

Abstract: In order to provide guaranteed QoS multiparty collaborative multimedia applications require reliable transmission of data. The multimedia applications can vary from distributed games, shared whiteboard to interactive video conferencing. These applications often involve a large number of participants and are interactive in nature with participants dynamically joining and leaving the applications[Sudan95]. In order to provide many-to-many interaction when the number of participants is large IP multicast is a very good option for communication. IP multicast provides scalability and efficient routing but does not provide the reliability these multimedia applications may require. Though a lot of research has been done on reliable multicast transport protocol, it really seems that the only way of doing a reliable multicast is to build it for a given purpose like conference control in multimedia conferencing.