It not silly.

There is a chapter in Implementing Backup and Recovery that Dave Little and I 
just finished writing that illustrates how to plan your tape capacity.  It 
should be easy enough to translate those formulae into trans log only equations.

Unfortunately I am not at liberty to send you those pages right now as the book 
should be coming out this month...but the equations work...but as you mentioned 
there are a lot of variables to consider...so in the interest of simplicity we 
didn't introduce combined D-INC and C-INC backup equations but based on D-INC 
only or C-INC only...if you get the book you'll understand what I'm talking 
about.  Anyway, I can't take credit for the equations, I enlisted the help from 
my brother who's got an under grad in Mathematics, Masters in EE and a PhD in 
Imaging Science.  So I trust his work.

David
-------------------------------------< 
David A. Chapa                         
DataStaff, inc. 
312 683 1144                 
mailto:consult AT datastaff DOT com 
http://www.datastaff.com     
                            
Co-Author: 
Implementing Backup and Recovery:       
The Readiness Guide for the Enterprise 
published by John Wiley and Sons 
A Veritas Solution Series Publication 
ISBN:0471227145 
-------------------------------------<

Quoting "Yosifovski, Tammy" <Tammy.Yosifovski AT usfc DOT com>:

> Okay, Okay so my question is silly but doesn't anyone have to predict tape
> usage to charge against a capital expenditure, before actually using the
> tapes not after?  I can easily do this for disk space but transaction logs
> are a whole ballpark of its own.   
> 
> 
> 
> -----Original Message-----
> From: Joost Mulders [mailto:mail AT j-mulders.demon DOT nl] 
> Sent: Wednesday, June 04, 2003 5:25 PM
> To: Veritas-bu AT mailman.eng.auburn DOT edu; Yosifovski, Tammy
> Subject: Re: [Veritas-bu] Tape Predictions for Transaction Logs
> 
> >Does anyone have a mathematical formula for predicting tape usage for
> >transaction logs?  Any idea on how to do this for something like this that
> >has so many variables?  
> 
> Hmm, let's see. 
> 
> ((X / (1024 ^3)) x Y) / Z) x W
> 
> Whereas:
>  X = The number of bytes your db's generate for transactionlogs
> 
>      A nice estimation of X would be, the number of users connected 
>      to your db x the number of keystrokes they do per day x their 
>      intelligence factor divided by the coffee cocient, which is an index
>      for the quality of the coffee your company serves. 
>      0.1 is bad coffee. 1 is very good coffee. So 0 >= X <= 1
>      
>  Y = is your retention, expressed in days. 
>  
>  Z = is the capacity of one tape, expressed in GB's. 
>  
>  W = a factor expressing the quality of the backup admin. 
>      W = 100 = Bad admin
>      W = 1   = Good admin
>      So, 1 >= W <= 100
>      
> Sorry, couldn't help it. Nothing better to do :-)
>  
>  
> -- 
> You're to have, not to hold.
> _______________________________________________
> Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
> http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
>