Following reception of:

>>> Title: LS - Reply to ITU-T SG11 relevant activities on “Internet 

>>> Speed Measurements” (reply to SG11-LS91) Submission Date: 2015-11-24 

>>> URL of the IETF Web page: https://datatracker.ietf.org/liaison/1444/



(Proposed) Liaison Reply to ITU-T SG 12, SG 11, and TSAG



The IP Performance Metrics (IPPM) working group of the IETF

appreciates the opportunity to comment on methods intending

to measure Internet Speed, such as those described in the 

attachments to your liaison (originating in SG 11).  This is

one of the topics which IPPM was chartered to address, and it

has proven more difficult than originally thought.



The cited proposals for International Recommendations 

depend on the assumption that TCP-based measurements can meet

requirements for standardized metrics, in terms of applicability

to subpaths as well as end-to-end paths, and repeatability

across independent implementations or measurement systems. 

IPPM's current work in progress (see reference below) 

describes why a TCP-based approach would be unsuccessful,

due to the properties of TCP's control loop (see sections 4.0

and 4.1 of the reference below for many more details).

IPPM is currently pursuing methods of measurement which avoid

TCP's control loop, and retain the ability to measure subpaths

and complete paths in a repeatable way. 



IPPM urges all members of the measurement and testing community

to carefully consider our findings and experience with 

capacity measurement, and to join our work in IETF to specify 

methods of measurement that avoid the pitfalls of TCP-based

testing. 



Referenced Work-in-progress:

https://tools.ietf.org/html/draft-ietf-ippm-model-based-metrics-07

Specific sections cited above:

https://tools.ietf.org/html/draft-ietf-ippm-model-based-metrics-07#section-4