CSTO_PROJECT_INFORMATION
PROJECT_SUMMARY
AO_NUMBER
D079
CONTRACTOR_OR_ORGANIZATION_NAME
University College London
PRINCIPAL_INVESTIGATOR
Peter Kirstein
and
Jon Crowcroft
Department of Computer Science
Gower St
London, WC1E 6BT
UK
tel +44-171-380-7286
fax +44-171-387-1397
e-mail kirstein@cs.ucl.ac.uk
Supporting Internet Multicast Multimedia.
See also
Home Page
OBJECTIVE
Multicast has emerged as one of the key enabling technologies for
supporting next generation networking applications such as video
conferencing. We believe that the following mechanisms have to be
developed before multicast can be used by wider community at a more
regular basis.
Mechanisms for supporting multicast routing protocols such as
coreplacement strategies for CBT and PIM, Quality of Service path
selection in multicast tree formation, and unicast support for reverse
path calculation.
Mechanisms for utilising multicast capability in applications such as
feedback control in a multicast environment, management of
applications in a multicast group, and reliable multicasttransport.
Mechanisms for integrating resource management with multicast
forwarding, such as interface between forwarding engine and caching
issues.
The purpose of the proposed research project is to develop novel
mechanisms for supporting and utilising Internet multicast. In
particular, we will carry out research and development in the three areas
identified above.
We intend to carry out in the following tasks
-
1. Design of a core placement algorithm for CBT and PIM.
-
2. Investigate Quality of Service (QoS) support in multicast routing.
-
3. Study security issues and protection mechanisms for multicast
networks.
-
4. Investigate unicast routing support for multicast tree formation with
reverse path forwarding.
-
5. Experiment on feedback traffic control algorithms for multicast-based
applications.
-
6. Investigate reliable multicast transport protocols based on the concept of
Application Layer Framing.
-
7. Design and implement management and control protocols for multicast-
based applications.
8. Integrate resource management and multicast forwarding.
-
9. Investigate the impacts of Mobility, Multimedia and Multicast traffic
RECENT_ACCOMPLISHMENT
While we have had recent accomplishments in this area, they have not been
carried out under the current task which has not yet started..
FY-96_PLAN
-
1. Core placement and management
CBT and PIM require a core-based multicast tree structure. At the
present, cores are placed "by hand" based on the topology and group
membership distribution. Automatic core placement and management is an
important research issue that has not been solved. The problem is known to
be complex. We will develop efficient heuristic algorithms for common
case multicast styles and from the knowledge of the traffic distribution.
We will also look into the issues such as core selection, placement, set-up
and error recovery.
-
2. Quality of Service (QoS) support in multicast routing.
Traditional routing metrics such as delay and hop-count are inadequate for
the new applications which often have stringent quality of service (QoS)
requirements. For a network to support diverse QoS requirements, it is
necessary for routing to have a more complex model of the network, for
example, characterised with multiple metrics such as bandwidth, delay
and loss probability, and to be able to find paths that satisfy specific
application QoS requirements. We will investigate QoS support in
multicast routing based on our early work on the complexity of QoS
routing.
-
3. Security issues for multicast networks.
Security issues in multicast networks are of extremely importance as
multicast communication is at a substantial increased risk from specific
security threats, compared with the same threats in unicast. Since
multicast involves one-to-many communications, a single node may affect
the whole network. A few incidents on the experimental MBone has shown
that an innocent mistake may cause the flooding of the whole MBone. We
will first try to identify the specific threat either malicious and non-
malicious, and investigate protection mechanisms based on both
preventative and detective approaches.
-
4. Unicast routing support for multicast tree formation.
Multicast algorithm such as DVMRP and PIM use the reverse path of
unicast routing to build the multicast tree. However, current unicast
routing does not support reverse path calculation. Ad-hoc mechanisms
have been used to bypass this problem (e.g., a built-in unicast routing
protocol in current implementation of DVMRP, and the PATH message
in RSVP). In the future, the problem must be addressed in unicast routing.
We will investigate mechanisms for integrating and supporting both
forward-path and reverse-path forwarding in a unified path calculation.
TECHNOLOGY_TRANSITION
In fact several router vendors have already started using some of our CBQ
code in their routers, and others plan to do so in the very near future.
However, since this project has not yet started, it seems premature to fill in
this section
Anticipated start 15 Jan 1995
Expected_end_date 30 June 1998
CURRENT_CONTRACT_FUNDING_PROFILE_BY_YEAR
FY94 Base 0
FY95 Base $77281
FY96 Base $
FY97 Base $
TOTAL Base $479625
OPTION_LIST
FY96 Option1 210 000
FY97 Option2 181 984
TOTAL Option 391 984
EXPENDITURES_BY_FISCAL_QUARTER
Base Effort Required In-hand Needed
Funding for 7/95-9/95 0 0 0
Funding for 10/95-12/95 0 0 0
CY95 Base 0 0 0
Funding for 1/96 - 3/96 43771 0 43771
Funding for 4/96 - 6/96 43771 0 43771
Funding for 7/96 - 9/96 52525 0 52525
Funding for 10/96 - 12/96 52525 0 52525
CY96 Base 192592 0 192592
DATE_OF_ADMINISTRATIVE_DATA
2/1/96
ANYTHING_ELSE_YOU_NEED
The project requires the current 768 Kbps link between the UK and the US
which is co-funded by ARPA and the UK MoD and UKERNA; It would be
desirable to increase the bandwidth of this link to at least T1 (1.544 Mbps),
and possibly higher to allow a full participation in DARTNET-II. It would
be ideal if this link could be upgraded in some way to nearer DS-3 to allow
proper use of ATM.