"Bob Jones" <email_at_me.not> wrote in
news:wEy2g.13707$4L1.802_at_newssvr11.news.prodigy.com:
>
> "ianal Vista" <ianal_vista_at_hotmail.com> wrote in message
> news:Xns97AD9992DC1BFianalvistahotmailcom_at_70.169.32.36...
>> "Bob Jones" <email_at_me.not> wrote in
>>>
>>> The first statement is a little vague. Low means how low? I would
>>> change "does not" to "sometimes does not".
>>> Again, in the second statement, good means how good? Close to 100%?
>>> Do you still consider cache size being inadequate then?
>>>
>>
>> Bob (Mr. BCHR is golden) Jones,
>>
>> http://oracledba.co.uk/tips/choose.htm
>>
>> "In these days where people are still mistakenly worrying about hit
>> ratios, here is a simple routine to bump up your hit ratio to any
>> value desired. Thanks to Jonathan Lewis for some of the strategy
>> used. "
>>
>> Assume an OLTP system that consistentantly run with 50% BCHR & steady
>> load.
>>
>> From 10:00 until 10:15 The code above is run on the OLTP system
>> selecting the BCHR to be 20%.
>>
>> From 10:15 until 10:30 The code above is run on the OLTP system
>> selecting the BCHR to be 80%.
>>
>> Which of the following TWO statements are more correct & why?
>> (comparing against response times from 09:30 - 10:00)
>>
>> a) From 10:00 - 10:15 the OLTP users will experience faster response
>> b) From 10:00 - 10:15 the OLTP users will experience slower response
>> c) From 10:00 - 10:15 the OLTP users will experience same response
>> d) From 10:15 - 10:30 the OLTP users will experience faster response
>> e) From 10:15 - 10:30 the OLTP users will experience slower response
>> f) From 10:15 - 10:30 the OLTP users will experience same response
>>
>> Inquirying minds await your response
>>
>
> My goodness. Obviously you have not read my previous postings. What
> did I said about artificially manipulating BCHR?
>
>
>
What is "artificial" in your eyes may be a manifestation of a separate
part of the application s/w.
BCHR knows nothing about the software that produces the (meaningless)
value. There is NOTHING artificial about the ratio. It is what it is.
Either the BCHR accurately reflects reality and the performace of the
database & is meaningful or it is a meaningless indicator of
performance.
I contend that a RELIABLE performance metric, can not be manipulated.
You don't like to answer the hard question, so you try to dismiss it.
HAND!
Received on Sat Apr 22 2006 - 18:40:55 CDT