Workshop on IP/Optical |
Chitose, Japan, 9-11 July 2002
Abstract: Session 11 - Network Performance (IP/OPTICAL)
Workshop Session
NETWORK PERFORMANCE (IP/OPTICAL)
Thursday, 11 July 2002
In conjunction with the
Workshop on IP/Optical
Convened by ITU-T at the request of Study Groups 13 and 15
Chitose, Japan, 9-11 July 2002
SESSION DESCRIPTION
This workshop session will focus on network performance and
related Quality of Service (QoS) requirements for IP over Optical networks, and
on how those requirements are being addressed in international standards and
emerging IP/Optical network infrastructures. The session will be a panel
discussion featuring experts participating in key international standards and
network planning groups including ITU-T Study Groups 13, 15, 12, and 4, the IETF,
and the TMForum. The panelists will each give a 20-minute presentation
summarizing their perspectives (and the work of industry groups in which they
participate) on a particular aspect of IP/Optical network performance. The
panelists and workshop attendees will then exchange views in a question-driven
discussion period. Before concluding the session, the panel will consolidate the
views expressed and identify insights and action items contributing to a more
coherent and better-coordinated plan for achieving IP/Optical network
performance requirements. The presentations and a session summary will be made
available to attendees in soft copy form.
Motivation
Many network planners believe the long-awaited convergence of
IP and telephony services will be based on a two-layer network architecture, in
which wavelength-switched optical "core" networks (comprised of DWDM
links, OXCs, and OADMs) interconnect conventional IP router "client"
networks that in turn serve IP end users – and all types of application data,
including real-time voice and video signals, are transported in IP packets. Such
"IP/Optical" networks will support dynamic resource reservation,
multi-layer traffic engineering, automatic protection switching and path
restoration, and many other advanced features. They will likely employ MPLS
technologies for network management and control. IP/Optical networks are
expected to reduce the costs of network deployment and operation by reducing the
number of network elements (and network layers) that must be installed and
managed, by providing bit rate transparency, and by automating the
time-consuming process of service provisioning. Such networks are expected to
promote new applications and enhance service creation by expanding network
capacity and providing a common platform for technology and service integration.
IP/Optical networks should also provide substantial
performance benefits. In addition to increasing available capacity, switched
optical core networks should offer attached IP client networks reduced packet
transfer delay and delay variation, lower packet loss and error rates, and
shorter delays in path provisioning and restoral. End users of concatenated
IP/Optical networks should experience all of these improvements, plus lower
delays and blocking probabilities in setting up individual managed IP flows or
connections. IP end users should also experience improved service availability
as a result of the path protection and restoration features implemented in the
optical cores. However, realizing these performance benefits in practical
IP/Optical networks will not be straightforward. A new performance specification
and measurement framework (and associated performance parameters and objectives)
will need to be defined for switched optical networks – and reconciled with
the frameworks used in characterizing conventional IP (and MPLS) router
networks. Unified signaling procedures will need to be developed to relate IP
end user QoS requirements and traffic specifications with routing, multiplexing,
traffic engineering, wavelength assignment, and other QoS control and resource
allocation mechanisms used in the IP router networks and switched optical cores.
A unified network management strategy will need to be identified to allow
independent IP/Optical network providers to monitor performance, detect and
isolate failures, and coordinate protection, restoration, and provisioning
activities.
Prospective Questions
How will the performance of IP-based multi-service networks
be specified on an end-to-end basis? What numerical performance requirements will VoIP and other anticipated real-time IP
applications impose? What new applications will drive the deployment of
integrated IP/Optical infrastructures, and what new performance requirements
will they entail? How will such requirements be allocated among IP network
providers? How will IP network performance be monitored, in and out of service?
How will IP network failures be localized? What mechanisms will IP-based networks
implement to provide assured-quality services?
How will the interactions between switched optical networks and client IP networks
be modeled from a performance point of view? What kinds of parameters will be
used in specifying the performance levels switched optical networks provide to
their IP clients? Will standard service classes or SLAs be used to document such
specifications? How will performance requirements be allocated among switched
optical networks operated by different providers? How will the performance of
switched optical networks be monitored? How will failures in such networks be
localized? What mechanisms will be used to control resource allocation and
network performance in switched optical networks?
How will performance requirements (and associated traffic contracts) for IP
flows supporting particular applications be communicated from IP users to the
serving IP router networks? What signaling mechanisms will be used to
communicate optical path performance and capacity requirements from IP clients
to serving optical core networks, and among concatenated optical networks? How
will IP QoS signaling be coordinated with optical network performance signaling
in IP/Optical networks? How will the requirements of IP end users be reflected
in optical network requirements? How will performance monitoring, failure
localization, and other network management functions (e.g., protection,
restoration, OAM&P) be coordinated between the IP and optical layers in
IP/Optical networks? Additional questions are encouraged.*
PANELISTS
(To be confirmed by May 17, 2002)
Al Morton, AT&T -- IP Layer Performance
Peter Huckett, Acterna -- Optical Layer Performance
Tobey Trygar, Telcordia -- IP/Optical Performance Management
CONTACTS
For further information about this session, please contact:
Neal Seitz - Session Chair
Vice Chair, ITU-T SG 13
Chair, Working Party 4/13 (Network Performance and Resource Management)
NTIA/ITS
Tel: +1 303 497 3106
E-mail: neal@its.bldrdoc.gov
BACKGROUND
In October of 2000, the WTSA confirmed Study Group 13
(Multi-protocol and IP-based Networks) as ITU-T Lead Study Group on IP related
matters, with responsibilities including the coordination of ITU-T efforts to
integrate heterogeneous telecommunication and IP-based networks and services.
The WTSA also confirmed Study Group 15 as ITU-T Lead Study Group on optical
technology, with responsibility for coordinating ITU-T activities involving the
transport, multiplexing, routing, management, supervision, and survivability of
optical channels carrying client signals. The IP/Optical Workshop will be an
opportunity to further develop these roles and activities, to coordinate them
with important responsibilities of other Study Groups (e.g., SGs 4 and 12) and
other organizations (e.g., IETF, TMForum), and to identify key issues and
strategies for the advancing the work.
OTHER WORKSHOP TOPICS IN CHITOSE
- Tuesday Business and Market aspects, Switched Optical Networks.
- Wednesday Optics in Access and Metropolitan Networks, Optical Interfaces, Network OAM & Protection
and Restoration, WDM and DWDM, Signaling and Routing, Optical Fibers,
Cables and Components, Optical Network Management.
- Thursday Optical Switching and Equipment, Optical Network Clients and Services, Service and Network
Evolution.
* The discussion will not address, tariff, regulatory, or policy issues.
|
|
|