Throughput Meeting --- August 17, 2016

======================================

 

Attending:  Shawn, Marian, Xinran, Ilija, Andy, Jason

Excused:  Horst

 

AGENDA:

 

0) Agenda review --  No changes.  Must be August since not many showed up for today's meeting :)

 

1) perfSONAR status

     i) New network problems  - RAL issue (see slides today),  MIT issue: MANLAN->MIT mbit/sec. Jason: LHCONE complicating the problem. See https://www.dropbox.com/s/qkcswsil3x8lhs0/perfsonar-traffic-flow.jpg?dl=0

           and https://www.dropbox.com/sh/9hfgfbh56ut1ue0/AABs0Dt3v7vovsoZae2Dqf4ra?dl=0 from Michael Tiernan.  Marian: Try tracepath from MANLAN to MIT to check MTU?

 

    ii) Updates on perfSONAR, MaDDash, mesh-config, etc.   Andy: Prepping for RC for perfSONAR in a few weeks.    Also test version of MaDDash available.

              Mesh-configuration challenges:  who own this?  Need long-term home.

 

2) Presentation on Network Analytics  (Xinran Wang & Ilija Vukotic / UC) See the network analytics presentation at :

    https://docs.google.com/presentation/d/1cvmR7nfCE8RzlHQ2LsVOERwgEPMQrg8EgWBu3tm45cc/edit?usp=sharing

 

        Question about whether ML is taking into account the topology? No not yet.    Alerting system is chatty?  Ilija: not too much so...tests showed

         a few emails a day.  Test it yourself.   Marian: question about NTP issues and bad time?  Lots of discussion.  Marian also noted new student

         will be at CERN and available to work on network analytics for ~2 months to see if there is a suitable bachelor's level topic there.  Will

         open email thread with Ilija on possibilities.

 

 

3) Round-table: Items to discuss?   No time this meeting.

 

4) AOB and next meeting:  We will be having the next meeting sometime in September (8th) and will be the European variant. 

     We will send email notification as we get closer.