Re: Effective compression ratio with minimal performance overhead

From: Mladen Gogala <gogala.mladen_at_gmail.com>
Date: Mon, 23 May 2022 07:29:18 -0400
Message-ID: <78de108c-3677-3573-8893-8a0cd1ccf8b5_at_gmail.com>



H Lok,
HCC is done on Exadata storage nodes, and is implemented in hardware. You don't see any overhead on the database side but that doesn't mean thee isn't any.
Regards

On 5/22/22 23:57, Lok P wrote:
> Thank you. Actually I read in many blogs(for e.g in below doc)
> stating there is "minimal performance overhead" with advanced
> compression, so this 100% degradation surprised me and thus wanted to
> check if this is expected behaviour or if I am checking it wrong. And
> at the same time , why is the HCC compression(compress for query low)
> experiencing zero performance overhead?
>
> https://www.oracle.com/technetwork/database/options/compression/advanced-compression-wp-12c-1896128.pdf
>
> On Mon, May 23, 2022 at 1:53 AM Mladen Gogala
> <gogala.mladen_at_gmail.com> wrote:
>
> On 5/22/22 15:31, Lok P wrote:
>> With Advanced compression, a table of size ~500GB is reduced to
>> ~100GB. But the execution time of a batch INSERT query (using a
>> batch size of 100) is increased from ~6minutes to ~13minutes for
>> 10million rows. And run time of a multi threaded 'UPDATE' using
>> pipelined function has been increased from ~11minutes to
>> ~29minutes. It was updating 16million rows.
>
> Well, there is no such thing as a free lunch. And thee is no such
> thing as cheap gasoline.
>
> --
> Mladen Gogala
> Database Consultant
> Tel: (347) 321-1217
> https://dbwhisperer.wordpress.com
>
> -- http://www.freelists.org/webpage/oracle-l
>

-- 
Mladen Gogala
Database Consultant
Tel: (347) 321-1217
https://dbwhisperer.wordpress.com

--
http://www.freelists.org/webpage/oracle-l
Received on Mon May 23 2022 - 13:29:18 CEST

Original text of this message