Skip to main content

Liaison statement
LS on harmonization of IP Capacity and Latency Parameters: Revision of Draft Rec. Y.1540 on IP packet transfer performance parameters and New Annex A with Lab Evaluation Plan

State Posted
Submitted Date 2018-12-11
From Group ITU-T-SG-15
From Contact Judit Katona Kiss
To Group ippm
To Contacts Brian Trammell <ietf@trammell.ch>
Bill Cerveny <ietf@wjcerveny.com>
Tommy Pauly <tpauly@apple.com>
Cc Scott Mansfield <Scott.Mansfield@Ericsson.com>
Bill Cerveny <ietf@wjcerveny.com>
Brian Trammell <ietf@trammell.ch>
IP Performance Measurement Discussion List <ippm@ietf.org>
John Drake <jdrake@juniper.net>
Mirja Kühlewind <ietf@kuehlewind.net>
Spencer Dawkins <spencerdawkins.ietf@gmail.com>
itu-t-liaison@iab.org
Tommy Pauly <tpauly@apple.com>
Response Contact acmorton@att.com
Purpose For comment
Deadline 2019-03-01 Action Taken
Attachments oLS70att-td619R1
Body
As an update to our communications from the Q17/12 Interim meeting (Darmstadt,
16-17 October 2018) we now share further results from our Geneva, 27 November –
6 December 2018 meeting with the full SG 12. We have made operational further
progress with our test bed that supports our Y.1540/Annex A Evaluation Plan for
IP Capacity and Latency methods of measurement.

The first test bed was demonstrated for the Q17 meeting. The T-rex measurement
tool and iPerf 2 test tool are now running reliably, and this allows us to test
with both TCP connections and UDP streams with iPerf 2.

Calibrated rates of 100, 200, 300, and 500 Mbps were tested. The UDP benchmark
accurately assessed the case with a 500 Mbps traffic shaper when accounting for
the fact that iPerf2 reports UDP payload in the results. The headers of
Ethernet, IP, and UDP (18+20+8 octets) add another 3.0% on 1470 byte packet
payloads, yielding 500.5 Mbps. The tests with 3 TCP connections (iPerf2)
produced measured rates that were nominally 5% lower than the 200/300/500 Mbps
traffic shaper settings and the UDP benchmark; they have not been corrected for
the Ethernet, IP and TCP header overhead, but this is the typical report
displayed to users. Even with correction for the header overhead, the combined
TCP connections appear to underestimate the calibrated rate targets.

Q17/12 has successfully recruited new participants to develop our evaluation
plan and Lab work, and the invitation remains open in the near-term.

Attachment:
SG12-TD619R1: Revisions to Y.1540 and text of new Annex A with Initial
Evaluation Plan