[Nsi-wg] Successful SuperComputing Demos

[Jerry Sobieski]

Hi all-

I want to apprise everyone of the successful NSI and AutomatedGOLE demos 
that took place last week at Supercomputing in Salt Lake City.

The demos consisted of deployments of NSI v2 into the Automated GOLE 
fabric.  We unfortunately did not have all the NSI implementations at v2 
yet, which posed a technical challenge to get v1 and v2 implementations 
to interoperate.  But as noted below, folks figured out how to address 
the issue and we were able to get v1 and v2 to co-exist!   Another 
bright spot: we had OSCARS active in the mix for the first time and 
running v2 (!) ... This is a great new capability for both the OSCARS 
user community and the NSI community.  The demos all went exceedingly 
well - not without challenges - but the problems were resolved and the 
demonstrations were showed to numerous folks.

I would like to thank especially Tomohiro Kudoh and the team at AIST for 
providing a v1-v2 translation capability in GLAMBDA-A that made it 
possible for the AutoGOLE to run in mixed mode.

And I would also like to recognize Jeroen van der Ham at University of 
Amsterdam for his efforts in translating and tweeking topologies in the 
new N3 format for the networks involved.   Jeroen was also responsible 
for adapting the automated connection scheduler to v2.

I also want everyone to recognize generally the effort the v2 
implementation required.  While there is much work still ahead to refine 
the v2 implementations, this was a mammoth effort with a very short 
timeline to get v2 implemented in time for this string of demos this 
fall.   So thank you all!   And congratulations for another successful 
milestone and demonstration!

Best regards to all!
Jerry