Techniques for quality of service guarantee in multiservice telecommunication networks
(MURST Research Program - Prot.9809321920)
The research program concerns the definition of techniques for quality of service guarantees in multiservice telecommunication networks.
Starting from a first valuation based on abstract models, subsequent step will be a simulation check.
The research program approaches are:
Both the approaches correspond to separate branches of the research program. The two research branches have a common first stage: the exact definition of various quality of service needs and the characterization of the correspondent traffic type.
As result, the program is articulated in three research lines:
The study will be focused on real time multicast QoS requirement, and on scalable stochastic models of voice and video sources.
QoS in multicast applications over IP networksThe problem of routing multicast packets over an internetwork and of providing a class of multicast transport services to the applications has been in recent years discussed in literature. The multicast routing may be provided through a variety of tree infrastructures: source-based tree, such as in DVMRP and MOSPF, shared trees, such as in the CBT, or a combined use of them, such as in PIM. All these proposals, however, still provide, for the multicast interaction scheme, the best effort service model of IP networks. On top of this service platform, quite a number of multicast transport protocols have been devised that address the general requirement of scalability, and of reliable and ordered packet delivery.
More recently, the availability of push technologies for the web and the increasing attention paid to emerging applications, e.g. desktop conferencing and webcasting, created the need for Quality of Service (QoS)-based multicast whose guarantees are beyond the best effort. To react to this new needs, the Internet world initiated the debate about the end-to-end QoS requirements by considering firstly the unicast traffic. On the contrary, the debate about QoS-based multicast routing and transport is at a very beginning stage.
The role of the University of Milan Research Unit (R.U.) in the project concerns the definition, the design, the evaluation and the experimentation of protocols for the transport and the delivery of multicast packets with real time constraints over best effort networks. The R.U. activity falls into two of the working packages defined by the project:
We will consider multimedia applications (mainly voice and data) with different end-to-end QoS requirements and we will specify the multicast service requirements by mainly addressing the guarantee of delivery to all the group members within the maximum specified delay. The introduction of new requirements should not modify the traditional ones of scalability and adaptation to dynamic membership and to multi-domain networks.
In the frame of this working package, our charge is studying and designing the multicast protocol in compliance with the specifications given at point 1 and the set of algorithms needed to modify the best effort service model accordingly.
The activities undertaken by the IETF WGs for the unicast traffic will be the base to initiate this activity. In particular, the algorithms of reservation, resource allocation and traffic control, will be designed to run over the routers that belong to the multicast shortest paths trees. At the same time, the admission control mechanisms to be performed at the upper layer must be defined to monitor the generated multicast flows.
STAGE 1: User and service requirement specification
1. classification of applications that need specified quality guarantees multicast services
2. definition of QoS regard to end-to-end delay, required bandwidth, limited jitter. QoS Control Model definition describing granularity, context, length of QoS required. Definition of metrics like loss rate, bandwidth, delay and jitter for specific QoS required.
3. definition of additional characteristics like reliability and security and their relation to defined QoS.
STAGE 2: Architecture specification
2. specification of support architecture and control policies to grant differentiated QoS
3. comparison of IETF solutions; documents and inherent publication collection
STAGE 3: Simulation model implementation and result evaluation
1. protocol for QoS multicast over best effort networks develop
2. additional security mechanisms for multicast communications
3. simulation model implementation
4. Performances comparison by simulation
Interesting linksHere you can find more information about QoS projects, applications, research and other people working for make QoS true!
IPMI is a forum where you can find information regarding:
- hardware and software vendors for multicast routing and applications
- available new technologies and solutions
IPMI home page is a good point to start exploring the multicast universe.
A list of all the official active IETF Working Groups. The list is ordered by contents; from there you can reach all the information about publications, people and addresses to contact them.
There you can find a list of IETF documents. I-D are working documents whose life is six months. The list is grouped by argument. You can find all the valid Internet draft publicized by all the IETF working groups. Note that you can reach some of the selected INTERNET DRAFT from our "Selected document" page.
A list of specific QoS Internet draft regarding integrated services (direct link). You can reach it from INTERNET DRAFT list too.
A list of specific QoS Internet draft regarding differentiated services (direct link). You can reach it from INTERNET DRAFT list too.
Selected documents" page.
(Project ended in 2002.)
|Last Updated on Wednesday, 12 November 2008 16:48|