The lightning strike-induced outage of an Amazon EC2 data center underscores the importance of eventually getting to a state of hybrid clouds. Having applications running in multiple locations simultaneously, applications that can run in-house or in the cloud (see Appistry), or having automatic live migration to another cloud or in-house servers can help to eliminate this kind of downtime (although I don’t know that the latter would help in the case of lightning). This strike also should drive use of Amazon’s Availability Zones, which are based on the notion that a lightning strike in the U.S. won’t affect a data center on another continent.
Infrastructure Links for this Week
-
With a Big Push, IBM Gives Cloud Computing Its Blessing

This is a smart move by IBM. Early signs pointed toward a Tivoli-driven management focus (software that does exist), but given the glut of internal cloud management software, IBM seems content to sell enterprise solutions, not technology.
Submitted by Derrick Harris
-
What Intel Can Teach Google About the Cloud

Vendor bias aside, WAN optimization will be important to cloud providers. Performance improvements alone can take a service so far before delivery needs to catch up.
Submitted by Derrick Harris
-
Software Licensing in a ‘Cloudy and Virtualized” World

A good question that appears to have been answered: Instances running certain software will have premium pricing.
Submitted by Derrick Harris
-
The coolest part to me: 500,000 names registered in 15 minutes with no reported technical problems.
Submitted by Derrick Harris
-
MMO Games are Still Screwed Up in Their Database Technology

I get the impression some gaming companies prefer form over function. I was surprised a few years ago to learn that the Second Life Grid isn't (or wasn't) a grid at all.
Submitted by Derrick Harris
