Re: a cume install time estimation formula

["Alexei Pytel" <pytel@xxxxxxxxxx>]

I see a couple of problems with your formula:
TimeForEachPtfOnCpw1Sys - PTF sizes are so different that PTF application
time is wildly different. Did you mean average time across all known PTFs ?
Besides you make an assumption that PTF load/temp apply time and PTF perm
apply/supercede time are the same. Actually it's much faster to perm
apply/supercede PTF already on system than to load and temp apply a new
one. This is why the more PTFs already on system, the faster cume apply.

I think this formula needs enhancement to differentiate between different
types of processing PTFs may require.

    Alexei Pytel




                                                                                
                   
                    "Steve Richter"                                             
                   
                    <srichter@AutoCoder       To:     <MIDRANGE-L@midrange.com> 
                   
                    .com>                     cc:                               
                   
                    Sent by:                  Subject:     a cume install time 
estimation formula  
                    owner-midrange-l@mi                                         
                   
                    drange.com                                                  
                   
                                                                                
                   
                                                                                
                   
                    05/16/2001 09:50 AM                                         
                   
                    Please respond to                                           
                   
                    MIDRANGE-L                                                  
                   
                                                                                
                   
                                                                                
                   



people,

here is my suggestion for a formula ibm could use to estimate cume ptf
install time:

( NbrPtf_In_Cume * CdLoadTimePerPtf ) +
(( ( NbrPtfToBeLoaded + NbrPtfOnTheSys ) * TimeForEachPtfOnCpw1Sys )  /
CpwOfTargetSys )

All times in milli seconds.

note: My formula assumes that a cpw 100 system will apply a ptf 10 times as
fast as a cpw 10 system. Even if this is wrong, there must still be a way,
using the cpw rating of the system, to calc the relative rate at which the
system can apply an average ptf.

Ibm would provide an option 51 = Calc cume ptf load time estimate on the Go
LicPgm menu.

NbrPtf_In_Cume is count of all ptf on the cume. Those that apply to pgm
products on the system and those that do not.

NbrPtfToBeLoaded is count of all ptf on the cume that apply to pgm products
on the system.

NbrPtfOnTheSys is count of all ptf currently on the target system. The more
ptf currently on the system, the more ptf that will have to be perm applied
and superceded during the cume load.

TimeForEachPtfOnCpw1Sys. On a system with cpw speed of 1, the time required
to perm apply a new ptf or to perm apply an existing ptf.

The first part of the formula calcs the time needed to copy the ptf's from
the cd to the system. I assume the speed of the cd is the same on all
system.

The 2nd part of the formula multiplies the time needed on a cpw1 speed
system to apply a single ptf by the total nbr of ptfs that will have to be
applied.

The 3rd part of the formula tries to factor how long it will take the
actual target system to apply the ptf's.

Steve Richter




+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---