[254] in Athena_Backup_System
Status of Jeff's review of the ABS requirements.
daemon@ATHENA.MIT.EDU (Bill Cattey)
Mon Jun 10 19:31:56 1996
Date: Mon, 10 Jun 1996 19:31:42 -0400 (EDT)
From: Bill Cattey <wdc@MIT.EDU>
To: athena-backup@MIT.EDU
Cc: rar@MIT.EDU, jis@MIT.EDU, azary@MIT.EDU, mbarker@MIT.EDU
This is just a note to make sure that all Athena Backup team members are
current on the situation with Jeff's review of the Athena Backup system
requirements.
To review:
I showed the requirements to Jeff to get a sanity check.
Also participating in the discussion were Tom Coppeto, Paul Hill, and
Matt Braun.
The outcome:
1. A few minor concerns which I've incorporated in the requirements doc
which I will more fully discuss in a subsequent note.
2. As I suspected at the ABS meeting of 30 May, a read-only crippled
mode was not acceptable. Although being able to perform regularly
scheduled dumps with the database offline was deemed ideal, being able
to dump the contents of a server onto a blank tape was considered
acceptable.
3. Jeff expressed the requirement that an abstraction layer be coded to
hide the relational database back end.
A quick discussion at the Athena Backup meeting of 6 June confirmed that
we could do the read/write crippled mode without much pain, but that
coding an abstraction layer would be very painful, and that the team did
not see much benefit to so great a cost.
Much discussion offline has taken place regarding what an appropriate
abstraction might be for the database coding, or what other action might
be appropriate.
Current state of affairs:
A discussion will be held with Roger Roach, Susan Minai-Azari, Jeff
Schiller, and myself to decide how to proceed.
In preparation for that discussion I am tuning up the requirements
document, and consulting with various folks with greater expertise than
I in database operation and programming, and in analysis of risks, costs
and benefits to different ways of proceeding.
I will keep you posted.
-wdc
