Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.
Comment: Migrated to Confluence 4.0

...

New pinger+perfSonar schema is here

Talking to Yee, he points out a few things about the database:

1. You do not want a table to exceed 1 Million rows. For PingER we have 48 measurements / day, roughly 1000 pairs, 365 days in a year or ~ 1.75Million rows/year.   Thus perfSONAR  shards the data into 1 table per month. Then there is a method to access data spread across multiple tables. PerfSONAR has classes to facilitate this, see for example http://anonsvn.internet2.edu/viewvc/viewvc.py/perfSONAR-PS/trunk/perfSONAR_PS-PingER/lib/perfSONAR_PS/DB/SQL/PingER.pm?view=log and these could be used.

2. The extra columns in data_data (e.g. throughput, MOS, alpha etc.) could simply be appended to the data table (thus not requiring a join).

3. perfSONAR has many methods for calculating IPDV, IQRLPD etc. They are written in Perl. We could use these. Familiarize yourself with http://anonsvn.internet2.edu/viewvc/viewvc.py/perfSONAR-PS/trunk/perfSONAR_PS-PingER/lib/perfSONAR_PS/Utils/PingStats.pm?view=log