[MOBY-dev] Service quality and availability data

[Phillip Lord]

>>>>> "mwilkinson" == mwilkinson  <mwilkinson at mobile.rogers.com> writes:

  mwilkinson> I don't think that this is going to be part of the spec
  mwilkinson> itself. It's a bit too high-level to be part of the
  mwilkinson> discovery system, but could potentially be a layer over
  mwilkinson> MOBY central.  

  mwilkinson> M -----Original Message----- From: Matthew Wakefield
  mwilkinson> <matthew.wakefield at anu.edu.au> Date: Wed, 9 Apr 2003
  mwilkinson> 17:50:02 To:moby-dev at biomoby.org Subject: [MOBY-dev]
  mwilkinson> Service quality and availability data

  mwilkinson> In a discussion of BioMoby the issue of service
  mwilkinson> overloading and outages was discussed.  The situation
  mwilkinson> raised was in the context of two BLAST servers, one on a
  mwilkinson> 486 in mongolia, and the other NCBI.  Is there any
  mwilkinson> planned way to differentiate these services to the naive
  mwilkinson> BioMoby user?


The problem with quality of service data is that it will be different
for different people. So storing it in moby central may not make
sense. If you have NCBI, and a 486, generally, the 486, er, sorry NCBI
is going to be quicker. 

If one the other hand you have NCBI, and EBI, well I am going to
choose EBI every time, because EBI is going to be much quicker. Well
unless its 9am and everyone in American is still in bed. In which case
NCBI might yet get there first. 

I'm sure, that most of you will not agree with me, as most of you are
not sitting in front of a machine in Manchester, but in the US, or at
least one of the Americas.

Moby central, is, well, central. So the description of the service is
the same for everyone. The conclusion that I draw from this is that
MOBY central should store information about services which relates
only the service itself. Quality of service relates also to the user
of that service, and so should be stored elsewhere. 

Where else? Well I'm not sure....

Cheers

Phil