wiki:GeniNetworkStitchingTestStatus/Measurements

Version 14 (modified by lnevers@bbn.com, 11 years ago) (diff)

--

Experiment Measurements

This page captures status and execution details for the Experiment Measurements. For overall status see the GENI Network Stitching Test Status page and for test details see the GENI Network Stitching Test Plan page.

Last update: 06/11/13

Measurements State Ticket Comments
GNS-T1 - Utah InstaGENI to GPO InstaGENI
GNS-T2 - InstaGENI Utah to GPO ExoGENI
GNS-T3 - InstaGENI Utah to GPO ExoGENI to LONI and MAX
GNS-T4 - InstaGENI Utah to GPO ExoGENI to CENIC
GNS-T5 - InstaGENI Utah to GPO ExoGENI to NYSERNet
GNS-T6 - Utah PG to GPO InstaGENI
PG Utah to IG Utah (not originally planned) Color(#63B8FF,In Progress)?


State Legend Description
Color(green,Pass)? Test completed and met all criteria
Color(#98FB98,Pass: most criteria)? Test completed and met most criteria. Exceptions documented
Color(red,Fail)? Test completed and failed to meet criteria.
Color(yellow,Complete)? Test completed but will require re-execution due to expected changes
Color(orange,Blocked)? Blocked by ticketed issue(s).
Color(#63B8FF,In Progress)? Currently under test.


Test Execution

Assumptions:

  • GPO PG (pgeni.gpolab.bbn.com) credentials are used for tests in this page.
  • Some end-points exist in multiple test cases, so their measurements will be not be duplicated.

Notes:

  • Stitcher currently does not support AM API v3, is delivered to run AM API V2.
  • Capacity must be specified for scenarios where Utah IG is end-point.

Measurements Goals

The set of GENI Network Stitching (GNS) test topologies selected to validate stitching experiment scenarios are also used to benchmark measurements to characterize various areas including setup timing, tear down timing, timing for an experiment to change data paths from one stitched VLAN to another, end-host to end-host data flow set up timing and delay. Throughput will also be measured with iperf between endpoint hosts using the following methodology:

  • All results are captured for default bandwidth allocation, unless otherwise stated.
  • All results are for a 60 second test runs.
  • TCP iperf measurements are collected for:
    • '1 client' scenario using command: 'iperf -c dest_host -t 60'
    • '5 clients' scenario using command: 'iperf -c dest_host -t 60 -P 5'
    • '10 clients' scenario command: 'iperf -c dest_host -t 60 -P 10'
  • UDP iperf measurements are run for 10 seconds increments to determine bandwidth supported

Measurements Collected

This sections captures all measurements collected for the Network Stitching test effort.

GNS-T1 - Topology 1 - Utah InstaGENI to GPO InstaGENI

Description Results Collected
VLAN stitching setup timing
VLAN stitching tear down timing
Data path change from stitched VLAN to another
End-host to end-host data flow set up timing
End-host to end-host delay
Throughput between end-hosts

GNS-T2 - Topology 2 - InstaGENI Utah to GPO ExoGENI

Description Results Collected
VLAN stitching setup timing
VLAN stitching tear down timing
Data path change from stitched VLAN to another
End-host to end-host data flow set up timing
End-host to end-host delay
Throughput between end-hosts

GNS-T3 - Topology 3 - InstaGENI Utah to GPO ExoGENI to LONI and MAX

Description Results Collected
VLAN stitching setup timing
VLAN stitching tear down timing
Data path change from stitched VLAN to another
End-host to end-host data flow set up timing
End-host to end-host delay
Throughput between end-hosts

GNS-T4 - Topology 4 - InstaGENI Utah to GPO ExoGENI to CENIC

Description Results Collected
VLAN stitching setup timing
VLAN stitching tear down timing
Data path change from stitched VLAN to another
End-host to end-host data flow set up timing
End-host to end-host delay
Throughput between end-hosts

GNS-T5 - Topology 5 - InstaGENI Utah to GPO ExoGENI to NYSERNet

Description Results Collected
VLAN stitching setup timing
VLAN stitching tear down timing
Data path change from stitched VLAN to another
End-host to end-host data flow set up timing
End-host to end-host delay
Throughput between end-hosts

GNS-T6 - Topology 6 - Utah PG to GPO InstaGENI

Description Results Collected
VLAN stitching setup timing
VLAN stitching tear down timing
Data path change from stitched VLAN to another
End-host to end-host data flow set up timing
End-host to end-host delay
Throughput between end-hosts

Utah PG to Utah InstaGENI (not originally planned)

Collected: 2013-06-11

Description Results Collected
VLAN stitching setup timing 1 minute 50 seconds
VLAN stitching tear down timing 2 minutes 9 seconds
Data path change from stitched VLAN to another ?
End-host to end-host data flow set up timing less than 1 second
End-host to end-host delay rtt min/avg/max/mdev = 0.180/0.256/0.319/0.036 ms
TCP Throughput between end-hosts (1 client) 95.8 Mbits/sec
TCP Throughput between end-hosts (5 clients) 96.0 Mbits/sec
TCP Throughput between end-hosts (10 clients) 96.0 Mbits/sec

Sliver set-up and tear-down

Collected: 2013-06-18 and 2013-06-19

End-point Aggregates Create SliverDelete sliver
GPO IG <-2links-> Utah IG 2 min 36 sec 8 min 14 sec
GPO IG <-> Utah IG 2 min 45 sec 2 min 32 sec
GPO IG <-> Kentucky PG 2 min 11 sec 1 min 52 sec
GPO IG <-> Utah PG 2 min 26 sec 2 min 11 sec
MAX MyPLC <-> GPO IG 4 min 00 sec 1 min 02 sec
MAX <-> Utah PG 4 min 59 sec 1 min 49 sec
MAX <-> PG KY 40 min 14 sec (*) 1 min 07 sec
MAX <-> PG KY 05 min 52 sec 1 min 40 sec
Utah IG <-> Utah PG <-> GPO IG 2 min 38 sec 3 min 10 sec
Kentucky PG to Utah IG
Utah PG <-> Kentucky PG <-> GPO IG <-> Utah PG 2 min 38 sec 3 min 10 sec
Utah IG <-> Utah PG <-> Kentucky PG <-> GPO IG <-> Utah IG

(*) No VLAN available due to SCS doing a poor job at randomizing VLAN tags out of a range with sparse small blocks. Fix was applied and test was re-ran test to collect new timing.