GAE memcache is limited. 

 

I'm about to crash so here is how  I'd do it really fast.

 

User A posts update. 

All Subscribed users get marker to check for updates



If UserB has Marker to check for updates he does.

If no marker last "report" is repeated.

 

Marker is a single byte you can keep lots of them in memcache.

 

GAE is going to be expensive because you basically have to plan on 1 read
and write per update from the user and 1 read for each updated online
friend, and 1 write for the final version

 

If average user is online 1 hour a day, and during their hour they see 20
updates. That is  40 reads and 20 writes

 

And you can't very well miss stuff, though Fb and twitter do. but they have
as much memcache as they want. We don't.

 

 

 


Brandon Wirtz 
BlackWaterOps: President / Lead Mercenary 

Description: http://www.linkedin.com/img/signature/bg_slate_385x42.jpg



Work: 510-992-6548 
Toll Free: 866-400-4536 

IM: [email protected] (Google Talk) 
Skype: drakegreene 
YouTube:  <http://www.youtube.com/blackwateropsdotcom> BlackWaterOpsDotCom 



 <http://www.blackwaterops.com/> BlackWater Ops 

 <http://www.cloudonastring.com/> Cloud On A String Mastermind Group



                

 

 

From: [email protected]
[mailto:[email protected]] On Behalf Of Kaan Soral
Sent: Wednesday, February 08, 2012 11:14 AM
To: [email protected]
Subject: [google-appengine] How to achieve Twtter of FB-like dynamism on
Appengine

 

I have searched "twitter" in this group and went back 1 year in posts but
couldn't find anything but I remember a discussion like this before

Problem:
On both Twtter and FB, when you visit their Homepage, you see a summary of
what your friends or interests are doing.
You might have up to 1000 of these on average.
How to combine this information?

Some thinking:
a) A naive approach would be to query for everyone, every time =
impossible-ish
b) To make this naive approach work we might determine popular friends of a
user and achieve a query once, construct a callback for near time updates
approach (to achieve real time info)
c) Another approach is to not query at all but update everyone who are
subscribed to a user, when that user posts an action
This would work for FB I guess, but for Twtter, considering some people have
1Ms of followers, this may be a problem
But the complexity is the sum of all followers over the network, and this
number may be low, considering not everyone has that many followers, so this
might be worth it


I am leaning over the (c) method, what do you guys think, how would you
implement these Social Networks?
(I combined these 2 networks in the problem definition, but you may provide
separate solutions)

-- 
You received this message because you are subscribed to the Google Groups
"Google App Engine" group.
To view this discussion on the web visit
https://groups.google.com/d/msg/google-appengine/-/1WEnOApSj4sJ.
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.

<<image003.jpg>>

Reply via email to