(Continuation of Question 13/13 studied during 1997-2000)

Background and justification

As critical communications services increase their reliance on IP networking, the user-perceived performance of IP networks is becoming increasingly important. When several IP network providers work together to provide end-to-end communications, each provider needs to understand its fair allocation of the end-to-end performance objectives. Those allocations must be both adequate for the service being offered and feasible based on the available networking technologies.

A framework is needed to guide the development of performance Recommendations for new network capabilities, transmission facilities, and services/applications, including those supported by the emerging global information infrastructure, and to relate performance Recommendations focused on different protocol and service layers.

Items for study

IP network performance

1. What QoS will be required by the services intended for IP networks (e.g., VoIP)?

2. What QoS is feasible on IP networks, given their "best effort" heritage, and given the IP QoS mechanisms being developed in the IETF and elsewhere? What QoS objectives can be specified for mobile/wireless IP networks? For MPLS paths?

3. How will the end-to-end QoS objectives for IP-based services be fairly allocated when more than one IP network participates in the provision of communications?

4. To what extent will QoS commitments depend on the existence of IP traffic contracts that completely specify the characteristics of the offered IP traffic?

5. How will users of IP-based services communicate their need for an IP QoS commitment?

6. How will QoS commitments of IP networks be verified?

7. How can small IPLR, IPER and SPR values be estimated from small traffic sample sizes?

8. What additional parameters and statistics should be added to Recommendation Y.1540 so that users can fully understand important performance attributes of IP networks?

9. What Recommendations are required to describe the throughput performance of IP "connections" and the relationship of throughput to availability?

General performance studies

10. How should the generic measurement points, reference events, communication functions, performance outcomes, and performance parameters defined in Recommendations I.350 be supplemented to address new network capabilities (e.g. multipoint connections, multiconnection calls, and modification of connection attributes), new access arrangements (e.g. wireless, satellites (including LEOs), HFC, xDSL), and new services/applications (e.g. interactive multimedia communications, personal and terminal mobility, flexible routing and charging, security, Internet access, and virtual private networks)?

11. How should the general structure of performance Recommendations summarized in Recommendation I.351/Y.801/Y.1540 be revised to reflect these innovations?

12. How should performance Recommendations address communications built on heterogeneous networking technologies?

Availability performance relationships

13. What Recommendations are required to show how availability performance is mapped among multiple, dynamically reconfigurable layers of networking (e.g., what is the relationship between the availability of the physical layer, the ATM layer and the IP layer)? How can availability specifications for different technologies be related?

14. How should performance Recommendations be enhanced to assure the success of critical communications during emergency situations (e.g. assuring the success of any communications that are critical to disaster response and relief)?

Specific tasks with deadlines

1. New Recommendation "Y.1541" defining QoS classes for IP networks based on the performance parameters of Y.1540 by year end 2001.

• end-to-end objectives suitable for IP applications
• per network allocations of the end-to-end objectives based on available QoS mechanisms information about how user’s QoS needs are communicated and interpreted.

2. Additions to Recommendation Y.1540, especially a delay variation parameter and (at least) a discussion of flow performance, decided by early 2002.

3. New Recommendation Y.800 similar to I.350 but for the global information infrastructure by year end 2002. This Recommendation is envisioned to have up to four aspects:

• a new conceptual model for understanding performance issues in the global information infrastructure with its non-homogeneous structure and its vertical applications,
• pointers to current Recommendations that are relevant to global information infrastructure performance,
• information on specific performance issues that result from network interworking and
• an availability performance model for heterogeneous networking.

4. Updates to Recommendation I.350 by 2004

5. Eliminating all references to I.353 and then deleting Recommendation I.353 by year-end 2001.

6. Updates, as appropriate, to Recommendation I.351/Y.801/Y.1501.

7. Maintenance, as necessary, for Recommendation I.355.

Relationships

• ITU-T Study Groups on services and network operations
• ITU-T Study Groups on measurements and maintenance
• ITU-T Study Groups on public data networks
• ITU-T Study Groups on performance aspects of TV and sound transmission
• ITU-T Study Groups on voice and multimedia performance
• ITU-R Study Groups on satellite systems
• ITU-R Study Groups on mobile-satellite services
• ITU-R Study Groups on radio systems
• ATM Forum on performance issues
• IETF on performance issues