Performance benchmark of PG

[Manish Lad]

>
> Dear all,
>
We are planning to migrate Oracle exadata database to postgresql and db
size ranges from 1 tb to 60 TB.

Will the PG support this with the performance matching to that of exadata
applince?
If anyone could point me in the right direction where i xan get the
benchmarking done for these two databases either on prime or any cloud
would be great.

Thanks all in advance.

Manish

>

Re: Performance benchmark of PG

[Laurenz Albe]

On Mon, 2021-07-19 at 15:39 +0530, Manish Lad wrote:
> We are planning to migrate Oracle exadata database to postgresql and db size 
> ranges from 1 tb to 60 TB. 
> 
> Will the PG support this with the performance matching to that of exadata 
> applince? 
> If anyone could point me in the right direction where i xan get the 
> benchmarking done
>  for these two databases either on prime or any cloud would be great. 

You won't find any trustworthy benchmarks anywhere, because Oracle expressedly
forbids publishing of benchmark results in its license, unless Oracle has given
its permission.

The question cannot be answered, because performance depends on your workload,
configuration, software and hardware.  Perhaps PostgreSQL will be faster, 
perhaps not.

Test and see.

Yours,
Laurenz Albe
-- 
Cybertec | https://www.cybertec-postgresql.com

Re: Performance benchmark of PG

[Manish Lad]

Yes you are right. I also experienced same in one such migration from db2
to PG which had read faster but the write was not meeting the need.

We then noticed the differences in disk types.

Once changed it matched the source.

Thanks and Regards

Manish

On Mon, 19 Jul 2021, 16:34 Laurenz Albe,  wrote:

> On Mon, 2021-07-19 at 15:39 +0530, Manish Lad wrote:
> > We are planning to migrate Oracle exadata database to postgresql and db
> size ranges from 1 tb to 60 TB.
> >
> > Will the PG support this with the performance matching to that of
> exadata applince?
> > If anyone could point me in the right direction where i xan get the
> benchmarking done
> >  for these two databases either on prime or any cloud would be great.
>
> You won't find any trustworthy benchmarks anywhere, because Oracle
> expressedly
> forbids publishing of benchmark results in its license, unless Oracle has
> given
> its permission.
>
> The question cannot be answered, because performance depends on your
> workload,
> configuration, software and hardware.  Perhaps PostgreSQL will be faster,
> perhaps not.
>
> Test and see.
>
> Yours,
> Laurenz Albe
> --
> Cybertec | https://www.cybertec-postgresql.com
>
>

Re: Performance benchmark of PG

[Hüseyin Demir]

Hi,

The question can not be answered in a proper way. Because, in PostgreSQL,
performance(response time in query execution events) depends on

1. Your disk/storage hardware. The performance can vary between SSD and HDD
for example.
2. Your PostgreSQL configurations. In other words, configuration parameters
can change your performance metrics. But you have to define your
queries,data size that a query can SELECT each time and queries that
INSERTS/UPDATES to database.
3. Your CPU and MEMORY hardwares can also change your performance metrics.
You have to compare your hardware infrastructure with Exadata appliances.
4. You also have to consider the connection pooling part in your
application part. PostgreSQL can suffer from performance problems because
of lack of connection pooling.

Regards.


Manish Lad , 19 Tem 2021 Pzt, 14:09 tarihinde şunu
yazdı:

> Yes you are right. I also experienced same in one such migration from db2
> to PG which had read faster but the write was not meeting the need.
>
> We then noticed the differences in disk types.
>
> Once changed it matched the source.
>
> Thanks and Regards
>
> Manish
>
> On Mon, 19 Jul 2021, 16:34 Laurenz Albe,  wrote:
>
>> On Mon, 2021-07-19 at 15:39 +0530, Manish Lad wrote:
>> > We are planning to migrate Oracle exadata database to postgresql and db
>> size ranges from 1 tb to 60 TB.
>> >
>> > Will the PG support this with the performance matching to that of
>> exadata applince?
>> > If anyone could point me in the right direction where i xan get the
>> benchmarking done
>> >  for these two databases either on prime or any cloud would be great.
>>
>> You won't find any trustworthy benchmarks anywhere, because Oracle
>> expressedly
>> forbids publishing of benchmark results in its license, unless Oracle has
>> given
>> its permission.
>>
>> The question cannot be answered, because performance depends on your
>> workload,
>> configuration, software and hardware.  Perhaps PostgreSQL will be faster,
>> perhaps not.
>>
>> Test and see.
>>
>> Yours,
>> Laurenz Albe
>> --
>> Cybertec | https://www.cybertec-postgresql.com
>>
>>

-- 
Hüseyin Demir

Senior Database Platform Engineer

Twitter:  https://twitter.com/d3rh5n
Linkedin: hseyindemir

Github: https://github.com/hseyindemir
Gitlab: https://gitlab.com/demirhuseyinn.94
Medium: https://demirhuseyinn-94.medium.com/

Re: Performance benchmark of PG

[Thomas Kellerer]

Manish Lad schrieb am 19.07.2021 um 12:09:
> We are planning to migrate Oracle exadata database to postgresql and
> db size ranges from 1 tb to 60 TB.
>
> Will the PG support this with the performance matching to that of
> exadata applince? If anyone could point me in the right direction
> where i xan get the benchmarking done for these two databases either
> on prime or any cloud would be great.


As already pointed out, you won't find such a benchmark.

You will have to run such a benchmark yourself. Ideally with a workload
that represents your use case. Or maybe with something like HammerDB.

But Exadata isn't only software, it's also hardware especially designed
to work together with Oracle's enterprise edition.

So if you want to get any reasonable results, you will at least have to
buy hardware that matches the Exadata HW specifications.

So if you run your own tests, make sure you buy comparable HW for
Postgres as well (lots of RAM and many fast server grade NVMes)

Re: Performance benchmark of PG

[Manish Lad]

Thank you all for your swift response.

Thank you again.

Manish

On Mon, 19 Jul 2021, 15:39 Manish Lad,  wrote:

> Dear all,
>>
> We are planning to migrate Oracle exadata database to postgresql and db
> size ranges from 1 tb to 60 TB.
>
> Will the PG support this with the performance matching to that of exadata
> applince?
> If anyone could point me in the right direction where i xan get the
> benchmarking done for these two databases either on prime or any cloud
> would be great.
>
> Thanks all in advance.
>
> Manish
>
>>

Re: Performance benchmark of PG

[Ninad Shah]

As Thomas rightly pointed about the feasibility of benchmarking. You may
still compare performance of queries on both Exadata as well as PostgreSQL.
IMO, it may not be on par, but it must be acceptable.

In the contemporary world, 60TB isn't really a huge database. So, I hardly
think you should find any performance issues on PostgreSQL.

All the best.


Regards,
Ninad Shah


On Mon, 19 Jul 2021 at 16:54, Manish Lad  wrote:

> Thank you all for your swift response.
>
> Thank you again.
>
> Manish
>
> On Mon, 19 Jul 2021, 15:39 Manish Lad,  wrote:
>
>> Dear all,
>>>
>> We are planning to migrate Oracle exadata database to postgresql and db
>> size ranges from 1 tb to 60 TB.
>>
>> Will the PG support this with the performance matching to that of exadata
>> applince?
>> If anyone could point me in the right direction where i xan get the
>> benchmarking done for these two databases either on prime or any cloud
>> would be great.
>>
>> Thanks all in advance.
>>
>> Manish
>>
>>>