Interesting evaluation of AppEngine for transaction processing and a comparison to other cloud providers

9 views
Skip to first unread message

Ivan Zuzak

unread,
May 27, 2010, 7:27:02 AM5/27/10
to Google App Engine
Hi all,

Just thought I'd mention this here as it is probably interesting to
AppEngine developers -- there's a paper on "An Evaluation of
Alternative Architectures for Transaction Processing in the Cloud"
available at [1] and a short overview at [2].

Best,
Ivan

[1] http://systems.ethz.pubzone.org/servlet/Attachment?attachmentId=76&versionId=1363456

[2] http://highscalability.com/blog/2010/5/26/end-to-end-performance-study-of-cloud-services.html

bFlood

unread,
May 27, 2010, 8:36:51 AM5/27/10
to Google App Engine
hi ivan

when will the new GAE results using non-transactional writes be
available? also, were these tests run *after* the 500 rps limit was
removed?

at first glance, it looks like a pretty thorough test and GAE does not
look too good...

thanks
brian

On May 27, 7:27 am, Ivan Zuzak <izu...@gmail.com> wrote:
> Hi all,
>
> Just thought I'd mention this here as it is probably interesting to
> AppEngine developers -- there's a paper on "An Evaluation of
> Alternative Architectures for Transaction Processing in the Cloud"
> available at [1] and a short overview at [2].
>
> Best,
> Ivan
>

> [1]http://systems.ethz.pubzone.org/servlet/Attachment?attachmentId=76&ve...
>
> [2]http://highscalability.com/blog/2010/5/26/end-to-end-performance-stud...

Ivan Zuzak

unread,
May 27, 2010, 7:31:22 PM5/27/10
to Google App Engine
Brian -- I forgot to mention that I am not involved with the
referenced articles in any way, so I have no idea when new results
will be available. Contacting the authors of the paper and reporting
back here sounds like the thing to do. :)

Cheers,
Ivan

Sharp-Developer.Net

unread,
May 28, 2010, 2:40:57 AM5/28/10
to Google App Engine
As someone pointed to in comments to the post looks like they use same
entity group in GAE datastore - what is bogus and make the test
senseless as GAE designed to be scaled and this test trying to prove
it's not.

Their reply point that "we wanted maximum consistency" just does not
add any sense as well - there are ways to achieve consistence without
single entities group though you would need spend some time to design
it properly.

So not that interesting to GAE developers.

Alex
http://sharp-developer.net/

On May 27, 12:27 pm, Ivan Zuzak <izu...@gmail.com> wrote:
> Hi all,
>
> Just thought I'd mention this here as it is probably interesting to
> AppEngine developers -- there's a paper on "An Evaluation of
> Alternative Architectures for Transaction Processing in the Cloud"
> available at [1] and a short overview at [2].
>
> Best,
> Ivan
>
> [1]http://systems.ethz.pubzone.org/servlet/Attachment?attachmentId=76&ve...
>
> [2]http://highscalability.com/blog/2010/5/26/end-to-end-performance-stud...

Darien Caldwell

unread,
May 28, 2010, 12:07:02 PM5/28/10
to Google App Engine
Yes, its like testing 3 cars, two diesel powered, and one gasoline
powered, and putting diesel in all 3 for 'maximum consistency'. :)

N. Rosencrantz

unread,
May 28, 2010, 3:54:48 PM5/28/10
to Google App Engine


On May 27, 7:27 am, Ivan Zuzak <izu...@gmail.com> wrote:
> Hi all,
>
> Just thought I'd mention this here as it is probably interesting to
> AppEngine developers -- there's a paper on "An Evaluation of
> Alternative Architectures for Transaction Processing in the Cloud"
> available at [1] and a short overview at [2].
>
> Best,
> Ivan
>
> [1]http://systems.ethz.pubzone.org/servlet/Attachment?attachmentId=76&ve...
>
> [2]http://highscalability.com/blog/2010/5/26/end-to-end-performance-stud...
Alway think why you need it:Alla edits reversible. Same thing
Reply all
Reply to author
Forward
0 new messages