Comments on: Newsletter – Trends http://akfpartners.com/techblog/2009/10/27/newsletter_trends/ Technical and Leadership Thoughts Thu, 29 Jul 2010 20:19:57 +0000 hourly 1 http://wordpress.org/?v=3.0 By: Karen Ellers http://akfpartners.com/techblog/2009/10/27/newsletter_trends/comment-page-1/#comment-1507 Karen Ellers Sun, 06 Jun 2010 19:31:25 +0000 http://akfpartners.com/techblog/?p=509#comment-1507 Nice articles, I'll bookmark this page for my references Nice articles, I’ll bookmark this page for my references

]]> By: Ori Lahav http://akfpartners.com/techblog/2009/10/27/newsletter_trends/comment-page-1/#comment-468 Ori Lahav Sun, 01 Nov 2009 10:07:18 +0000 http://akfpartners.com/techblog/?p=509#comment-468 Hey AKF fellows Thanks for the News Letter update. The continuous deployment part was educating and I must admit it might be wise to learn more about it. About Key-Value Stores, the list of systems you have mentioned is a very partial list and include those who are very limited in terms of scaling. meaning: memcached and redis are memory based and will always take advantage of the RAM you can give them. outbrain uses memcached heavily and vary happy with it but as cache solution and not really as a (persistent) data store. redis is very much the same but in new versions also support dumping to disk. CouchDB is kind of the opposite, it is not a memory only DB and uses earlang map-reduce capabilities very nicely. however couchDB is distributed only on terms of data replication and NOT on data sharding and distribution. so basically you run into the same walls that you get using MySql replication. I'm currently doing some research about NoSql, distributed, highly scalable solutions and I would love sharing my notes with you once done. About cloud computing - my view of it is here :-) http://olahav.typepad.com/weblog_files/2009/05/my-take-on-cloud-services.html again, thanks for the post Ori Hey AKF fellows
Thanks for the News Letter update.

The continuous deployment part was educating and I must admit it might be wise to learn more about it.

About Key-Value Stores, the list of systems you have mentioned is a very partial list and include those who are very limited in terms of scaling. meaning:
memcached and redis are memory based and will always take advantage of the RAM you can give them. outbrain uses memcached heavily and vary happy with it but as cache solution and not really as a (persistent) data store. redis is very much the same but in new versions also support dumping to disk.

CouchDB is kind of the opposite, it is not a memory only DB and uses earlang map-reduce capabilities very nicely. however couchDB is distributed only on terms of data replication and NOT on data sharding and distribution. so basically you run into the same walls that you get using MySql replication.

I’m currently doing some research about NoSql, distributed, highly scalable solutions and I would love sharing my notes with you once done.

About cloud computing – my view of it is here :-)
http://olahav.typepad.com/weblog_files/2009/05/my-take-on-cloud-services.html

again, thanks for the post

Ori

]]>