IEPM Group Task List Current

Action Items

Terapaths Netflow (see http://iepmbw.bnl.gov/netflow/index.html) - Yee

1. Get security requirements from SLAC/John H (Yee will talk to Gary 11/14/06) - Yee
2. Presentation/Front-End - Yee
   1. Reimplement spider and pie charts for SVG::TT - Shahryar, Yee
      1. Implement pie chart class in SVG::TT - Shahryar, Yee
      2. Implement spider class in SVG/Template format - Shahryar, Yee
      3. Tidy up legends for pie and spider charts - Shahryar, Yee

   2. Fix labels on time series plots - Yee [DONE 20061019]

3. Processing Back-End - Yee, Akbar
   1. Factorise out TopN code in JKFlow.pm (to do 11/14/06) - Akbar
   2. Refactorise JKFlow code for QoS analysis (restructure file structure and possibly QoS analysis) - Akbar
   3. Technical Documentation - Akbar
4. Testing and Installation - Yee, Akbar
   1. Finalise installation script - Akbar, Yee
      1. Add patching - need distribution mechanism for tarball storage - Yee, Akbar

      2. Test perl prefix installations - Akbar, Yee [done 20060905]

   2. Add UI/cgi code to installation - Akbar, Yee
   3. Try out installation on fresh machine at SLAC

Transport services evaluation - Yee

1. Work with Microsoft - Yee, Les
   1. Get latest privates from Microsoft
   2. Decide what is needed for stage 2
   3. Install Vista/Longhorn (Windows Vista does not install 11/14/06) - Yee
2. NDT server at SLAC (see http://nettest5:7123) - Yee

   1. Installed on nettest5 NDT Install - Yee [DONE 20060831]

   2. Talk security into allowing public access - Yee.

PingER

Redesign and Implement Guthrie to cover both IEPM and PingER

IEPM-BW

1. Work with DESY to get new monitoring host  (Jerrod will read Connie' documentation and see if he can accomplish this, adding to the documentation as he goes along) - Jerrod
2. Make RAL a remote node
   1. Have account but cannot ssh to it (sent email to tasker 9/8/06) - Les
3. Make FZK an IEPM Monitoring node - Connie - DONE
   1. Get contact for Connie (sent email 8/22/06, now awaits Connie) - DOME
4. Update metrics used
   1. ID and add more targets for pathload - Connie - NOTE: pathload is problematical.
5. Get distribution kit for iepm monitoring nodes to install & configure (Done adding owamp to the installkit as requested 10/3/06) - Jerrod
   1. Update pre-reqs document - Jerrod

6. Add LHC Atlas hosts to IEPM-BW (list sent to Jerrod 9/6/06, added needs testing 9/6/06) [Done 9/8/06] - Jerrod, Les, Connie

7. Add group for US-ATLAS [Done 9/20/06] -  Connie

8. Diagnose orphan sockets problem - Connie - IN PROGRESS 11/15/06
9. Get architecture of remote nodes and create a web page (wil get back to in Sept 06) - Jerrod
10. Write script to use ssh to get the configurations of IEPM monitor and remote hosts (in progress 4/26/06, will revisit Sep '06) - Jerrod
11. Do we want to get reverse traceroutes (at least where we have reverse traceroute servers, awaits time) - Connie
12. Fix up TCP receive buffer sizes, add sanity checks - needs root on target hosts, which we do not have -
13. Bugs:

Traceanal Modularisation - Yee, Asif

1. Integrate new topology into web server - awaits wan-mon appropoval - Yee
2. Rendering of topology much slower on www.slac.stanford.edu - likely to be related to the web server, not code, as it runs quickly elsewhere (awaits WANMON web server, see below) - Yee
3. Traceroute_analysis - Asif, Yee
   1. Prepare distributable version of traceanal - Asif, Yee
   2. Tidy up presentation - Asif
   3. Profile code for speed improvements - Asif, Yee
   4. Provide example of how to allow integration into Non-IEPM-BW data sets - Asif, Yee
   5. Update all Test_cases -Asif, Yee

IEPM-BW Ping Visualization - Les, Asif

1. Update Traceanal Modular code to gather and visualize IEPM ping data - Asif [BASE Code DONE 20061116: Rest will be assigned to student at NIIT]

Alerts and Diagnosis - Les, Yee

1. Look at multivariate event detection (collect data for SLAC, BNL, Caltech pathchirp, thrulay,ping) - Adnan
2. Look at improvements to plateau (Yee plans a rewrite)
   1. Ability to find step ups - Yee
   2. Extend to allow up & down then compare down with original - Adnan
   3. Allow for small number of samples (e.g. at start) -
3. Look at other detection algorithms and compare - Akbar
   1. Holt-Winters - Les, Mahesh, Felipe
      1. Go back 7 weeks -
      2. Check unusual results -
      3. Consider other ways to optimize parameters -
   2. Neural networks, Bayesian ...
4. Prepare table of canonical events and how various algorithms react - Adnan
   1. Build case studies of email event (see https://confluence.slac.stanford.edu/display/IEPM/Anomaly+Case+Studies)  - Adnan
5. Look at detecting outages for ping - Connie
   1. Analyze what constitutes a significant outage  - Connie
6. Understand cause of delayed alerts and see if can improve - Connie - fixed - done Connie

Install WANMON as IEPM web server - Yee

1. Install machine (awaits SCCS to install Dell host, Yee will track down 11/14/06)
2. Get traceroute.pl and pingtable.pl working and in production (will be via John B. as they will be maintaining it) - Yee

PerfSONAR - Yee

1. IEPM-BW Web Services - Yee, Asif

   1. -Measurement Archive Service for IEPM-BW data - Yee [DONE 20061110]

      1. Install SQL-MA* info*- Yee [STALLED due to complexity of Java instalaltion and code base compared to Perl]

         1. Write Ibatis configs for IEPM-BW data - Asif
2. NMWG requirements? - Yee

Presentations/Talks/Visits/Papers/Documentation

IPv6