Comments on: How Much of Twitter’s $100M is for Servers? http://www.datacenterknowledge.com/archives/2009/09/28/how-much-of-twitters-100m-is-for-servers/ News and analysis about data centers, cloud computing, managed hosting and disaster recovery Thu, 18 Mar 2010 14:34:45 +0000 http://wordpress.org/?v=abc hourly 1 By: The Cloud, Hadoop Marched Toward the Mainstream in Q3 http://www.datacenterknowledge.com/archives/2009/09/28/how-much-of-twitters-100m-is-for-servers/comment-page-1/#comment-7353 The Cloud, Hadoop Marched Toward the Mainstream in Q3 Tue, 20 Oct 2009 22:55:28 +0000 http://www.datacenterknowledge.com/?p=16066#comment-7353 [...] Facebook announced in September that it is seriously upping its infrastructure investment, and industry experts believe Twitter will have to spend a substantial portion of its latest $100 million on infrastructure, as [...] [...] Facebook announced in September that it is seriously upping its infrastructure investment, and industry experts believe Twitter will have to spend a substantial portion of its latest $100 million on infrastructure, as [...]

]]> By: Keith Hanna http://www.datacenterknowledge.com/archives/2009/09/28/how-much-of-twitters-100m-is-for-servers/comment-page-1/#comment-6672 Keith Hanna Tue, 29 Sep 2009 10:27:14 +0000 http://www.datacenterknowledge.com/?p=16066#comment-6672 Twitter also and possibly more importantly needs to invest in best practices like ITL. Specifically, Capacity Management methodologies and tools must be a part of their data center operations. Just blowing cold hard cash on servers will not guarantee service reliability and quality. Twitter and NTT America need to stay ahead of the wave of demand with predictive modeling technologies that are based on queuing theory which model non-linear behavior. Otherwise they're just guessing how much capacity they need and when. Twitter also and possibly more importantly needs to invest in best practices like ITL. Specifically, Capacity Management methodologies and tools must be a part of their data center operations. Just blowing cold hard cash on servers will not guarantee service reliability and quality. Twitter and NTT America need to stay ahead of the wave of demand with predictive modeling technologies that are based on queuing theory which model non-linear behavior. Otherwise they’re just guessing how much capacity they need and when.

]]>