Re: SQL Backtrack for Oracle vs. RMAN
Date: Fri, 26 Jul 2002 11:31:09 -0700
Message-ID: <ahs4nh$1pb$1_at_pith.uoregon.edu>
As we do not know your system. You will have to find the ROI on your own.
Usually take a MTTF and find the possible downtime per year. Do not count downtime that is already backed up by others systems: redundant fail over, Mirror; Raid5; etc.
First: Once you have a number of times that you require the recovery. Run through the cost of the rebuild. with the Backtrack vs. RMAN. If there is no savings at this point (which I doubt there is) have faith. Do not forget that downtime usually has more cost than just IT. You must calculate the cost of missing data. Assign $$$ value to each row of data and multiply that by the number of rows it is possible to loose. And of course do not forget the impact on production. If the DB goes down in the middle of the day and 10K workers can not do work the difference between a 2 hour and a 2.5 hour recovery can be huge (the hourly cost of personell is usualy 30-300% more than their hourly wage depending on the business).
Second calculate the cost of maintaining Backtrack vs RMAN. Once the proceedure is created it is probably just a matter of rolling tapes for both. So probably no savings here.
Third Determine the reliablility of the product and possible human error introduced in to the system by convouluted and poorly desinged software and system proceedures. (This may be where you make the ROI). If the newbie DBA 3 years from now is using the system and thinks he has good backups but the system crashes and we find that the backups were not running and there was no clue for the newbie DBA that there may be problems whole system and has not been for 5 months; Recalculate the First step with A total loss of data for those 5 months, the downtime for production, and the cost of reenteing the data in human hours plus overtime going forward. If this does not scare your mgmt then find a new job.
Finally I would say weigh all of these costs with your flimsy Backtrack software, RMAN, and the cost of a fully redundant hot failover system. The choice should be obvious.
I am sure you know all of this but:
Good Luck
"Wormie" <ordeith_at_coruscant.net> wrote in message
news:df31f670.0207260905.3e36be0e_at_posting.google.com...
> At my company, we are big users of SQL Backtrack for our disaster
> recovery of our Oracle databases (we also use if for Informix, MSSQL,
> Sybase and Oracle NT).
>
> However, we are being forced (from the top) to show hard dollar ROI
> for SQLBT in order to keep it as opposed to switching to RMAN (which
> none of us want to do).
>
> Does anyone have any information on how SQL Backtrack is a better tool
> vs. RMAN. Specifically, how we can realize a return on investment
> using SQLBT, as opposed to using RMAN. It is difficult to do, since
> disaster recovery is sort of like an insurance policy.
>
> Any input would be appreciated,
>
> Thanks!
>
> BTW, we are using Oracle 8.0.5 -> 8.1.7 on Sun and AIX.
Received on Fri Jul 26 2002 - 20:31:09 CEST