Trace selection, schedules and duration

New Message Reply About this list Date view Thread view Subject view Author view

From: Joerg Micheel (joerg@cs.waikato.ac.nz)
Date: Tue Jan 09 2001 - 21:24:42 PST


>From our Web logs we gather that about 25 to 30 trace files get downloaded
by users every day. We currently have 17 systems alive, each capturing a
90 second snapshot 8 times a day, which makes for over 100 traces per day.
Assuming that each file gets downloaded only once, we could estimate about
20-25% efficiency.

Our current plan for 2001 is to roughly double the number of active monitors
to about 30-35. Included are a number of high-bandwidth connections, such
as 622 MBit/sec PoS links. The volume of data captured from those links
may be easily an order of magnitude higher than on the traditional monitors.
At the moment, we face between 1.5 and 3.2 GBytes of compressed data per
day. We need a fresh DDS-4 tape for archival roughly every week. Post-
processing time on some busy monitors becomes scarce. It is very clear
that we will not be able to stick with the current trace schedule and
trace durations for much longer.

=========================================================================
We need your input in order to rearrange the trace taking and provide a
service that suits your needs best.
=========================================================================

Please let us know:

o Whether you think 90 second snapshots are sufficient (do you need
  longer traces) ?
o Whether 8 samples per day are appropriate (would you rather have
  less samples per day, only a few days per week, only one week out
  of a month, ...) ?
o How do you select trace files (by trace size, preference for particular
  sites, alternating, historical, ...) ?
o Whether you are interested in historic data (shall we delete most
  of the trace data online, rather than tape archiving everything) ?
o Would you like to see correlated traces from multiple monitors
  (including time syncronization) ?
o How do you find the WWW interface to traces ? (we see a lot of
  aborted downloads, would you prefer a different medium, would
  you prefer ftp with reget option)
o What kind of documentation/graphs would you like to see along the
  raw traces ?

Thank you for your time.

For the PMA team
        Joerg

-- 
Joerg B. Micheel			Email: <joerg@cs.waikato.ac.nz>
WAND and NLANR MOAT			Email: <joerg@nlanr.net>
The University of Waikato, CompScience	Phone: +64 7 8384794
Private Bag 3105			Fax:   +64 7 8585095
Hamilton, New Zealand			Plan:  PMA, TINE and the DAG's


New Message Reply About this list Date view Thread view Subject view Author view

This archive was generated by hypermail 2b30 : Thu Sep 27 2001 - 16:24:40 PDT