On Thu, 6 Jul 2017 14:25:21 -0300, Clark Morris wrote:
>>
>>In the case of z/OS Parallel Sysplex, it is "5 nines" 99.99999%
>>
Errr... I count seven nines there.
https://en.wikipedia.org/wiki/High_availability#.22Nines.22
>>This equates to about 3 seconds of non-availability per year.
>>
Five nines would be about 5 minutes of non-availability per year.
https://en.wikipedia.org/wiki/High_availability#Percentage_calculation
>>Tools/techniques exist in the *nix world to provide this availability (or at
>>least something much better that we currently receiving). It is a matter of
>>"IBM's will" to implement or not.
>>
Not "will", but economics. End-to-end reliability would be improved only
negligibly
by improving server reliability multiple orders of magnitude above the limiting
factor, the Internet. However, I agree with the consensus that KC isn't even
close.
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
