The documentation says: "...Memcache...is accessible by multiple instances of your application"
I don't know if this means that each server instance has its own Memcache or that all the instances access a single Memcache. In any case, when the Memcache needs to be invalidated for a particular key, then the Memcache needs to be invalidated for all instances of the application. So the Memcache cannot be entirely separate for each instance of the application. I have seen many developers implement all kinds of clever caches which have been like reinventing the wheel many times over, plus caching often leads to nasty bugs in the code and I suspect that the performance in many cases is in practice not significantly increased or not increased at all. A lot of time, effort and resources spent to create messy caching code that was never needed in the first place and only makes the application much more difficult, costly and time- consuming to develop, maintain and modify. On Nov 17, 8:13 am, Roberto Saccon <[EMAIL PROTECTED]> wrote: > Interesting questions. I don't know anything (beside the API) about > it. But the amount of available memcache is limited. I want to have as > much as possible for me and I am willing to put some coding effort in > to get my slice. And I hope others are lazy and put in less effort or > none at all and leave a bigger slice of memcache for me, so I hope > memcache won't ever be done automatically. > > Anything wrong with this way of thinking (beside of the moral > aspect) ? > > regards > Roberto > > On Nov 17, 4:56 am, Anders <[EMAIL PROTECTED]> wrote: > > > And how well does the Memcache scale? Is a sharded Memcache needed to > > prevent a performance bottleneck when the load increases? Is it > > similar to a sharded counter for datastore records: When the load is > > low then using a single counter for datastore records works fine, but > > when the load increases then the single counter becomes a bottleneck > > when it is being hit by many simultaneous requests. > > > Or is the Memcache already sharded behind the scene, or in some other > > way already made to scale well? > > --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Google App Engine" group. To post to this group, send email to google-appengine@googlegroups.com 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 -~----------~----~----~----~------~----~------~--~---
