The myth and reality of software-defined data centersDate: Jul 22, 2014
As software-defined approaches are applied to information technology, organizations can have the choice of outsourcing IT functions to service providers.
"The concept -- simply in a nutshell -- is that by virtualizing our servers, by placing them into failover clusters, we've eliminated the need for disaster recovery," Toigo told his Storage Decisions audience.
"Now we've got continuous operations courtesy of these nonstop, never-fail networks, never-fail storage and never-fail servers that are basically going to change the rules of the game," said Toigo, the CEO and managing principal of analyst firm Toigo Partners International, as well as chairman and co-founder of the Data Management Institute.
Toigo called this outlook a fundamental misunderstanding of DR and the ability of technology to be self-sustaining. He pointed to the availability of networks as the "Achilles heel" of software-defined data centers.
"Regardless of how good your cloud service provider is, regardless of the high SLAs (service-level agreements) he's able to deliver, regardless of the availability of the systems that he builds and offers to you, the fact is his SLAs are subject to the gating factor that is implemented by the network," said Toigo. "The network is not usually under the service provider's control."
He said that even if a service provider is able to keep his operation up and running 100 percent of the time, access to that resource is limited directly by the speeds and availability of the network that connects a customer to that provider.
"The whole SLA concept becomes problematic," said Toigo.