JSON will not beat python's marshal. It wont even come close.
See my other blogpost:
http://devblog.miumeet.com/2012/06/storing-json-efficiently-in-python-on.html
It also discusses zip options on marshalled data.

Cheers,
-Andrin

On Wed, Sep 5, 2012 at 7:53 PM, Drake <[email protected]> wrote:

> Proto-buffers are mostly fast, but not always. And they are typically
> larger, I think more info about the objects are stored.
>
> > -----Original Message-----
> > From: [email protected] [mailto:google-
> > [email protected]] On Behalf Of Jeff Schnitzer
> > Sent: Wednesday, September 05, 2012 10:27 AM
> > To: [email protected]
> > Subject: Re: [google-appengine] Speeding up memcache by 3x for datastore
> > entities
> >
> > Have you compared this against the performance of ndb's integrated
> > caching?  I believe it uses protocol buffers.
> >
> > Also, getting entity cache behavior right - so that contention doesn't
> throw
> > the cache out of sync - is very hard.
> >
> > Jeff
> >
> > On Wed, Sep 5, 2012 at 9:32 AM, Drake <[email protected]> wrote:
> > > Yes, this has been discussed before. We use marshal for a lot of our
> > > stuff, but it isn't always faster. It depends on how large the objects
> > > are and the types.
> > >
> > >
> > >
> > > The size difference is interesting, as we rarely see that.
> > >
> > >
> > >
> > > You should also benchmark against JSON which is often faster than
> > > Marshall, and often smaller.
> > >
> > >
> > >
> > > Lastly you should bench each with Zip compression. The size of the
> > > object does matter, and at different compression levels you get
> > > different performance curves.
> > >
> > >
> > >
> > >
> > >
> > > From: [email protected]
> > > [mailto:[email protected]] On Behalf Of Andrin von
> > > Rechenberg
> > > Sent: Wednesday, September 05, 2012 2:15 AM
> > > To: [email protected]
> > > Subject: [google-appengine] Speeding up memcache by 3x for datastore
> > > entities
> > >
> > >
> > >
> > > There is an amazingly simple trick how to speed up memcache by 3x for
> > > datastore models
> > >
> > > and entities use only half the size in memcache, see my blogpost:
> > >
> > >
> > >
> > > http://devblog.miumeet.com/2012/09/speeding-up-memcache-by-3x-
> > for.html
> > >
> > >
> > >
> > > Cheers,
> > >
> > > -Andrin
> > >
> > > --
> > > You received this message because you are subscribed to the Google
> > > Groups "Google App Engine" group.
> > > To post to this group, send email to [email protected]
> .
> > > To unsubscribe from this group, send email to
> > > [email protected].
> > > For more options, visit this group at
> > > http://groups.google.com/group/google-appengine?hl=en.
> > >
> > > --
> > > You received this message because you are subscribed to the Google
> > > Groups "Google App Engine" group.
> > > To post to this group, send email to [email protected]
> .
> > > To unsubscribe from this group, send email to
> > > [email protected].
> > > For more options, visit this group at
> > > http://groups.google.com/group/google-appengine?hl=en.
> >
> > --
> > You received this message because you are subscribed to the Google Groups
> > "Google App Engine" group.
> > To post to this group, send email to [email protected].
> > To unsubscribe from this group, send email to google-
> > [email protected].
> > For more options, visit this group at
> > http://groups.google.com/group/google-appengine?hl=en.
>
>
>
> --
> You received this message because you are subscribed to the Google Groups
> "Google App Engine" group.
> To post to this group, send email to [email protected].
> To unsubscribe from this group, send email to
> [email protected].
> For more options, visit this group at
> http://groups.google.com/group/google-appengine?hl=en.
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"Google App Engine" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/google-appengine?hl=en.

Reply via email to