Version 6.0 replaces polling--in which the server frequently checks with the application to make sure it has the computing resources it needs--with what Symantec calls an Intelligent Network Framework (IMF). It is already a feature of Symantec's UNIX line of management software, said V. Don Angspatt, VP of product management for Symantec's storage and availability management group.
"The problem is that polling takes time but it also consumes CPU cycles and so as more CPU cycles are consumed, then that aggravates the failure detection time even more. So now, you've all of a sudden just made things a lot worse," Angspatt said. With the IMF approach, there's an agent framework within the network that registers all the resources and then the framework is immediately notified when there's a change of state, he said.
Veritas SFHA 6.0 also reduces failover time by avoiding the time-consuming tasks associated with connecting a new server to the storage array that was connected to the failed server, such as scanning the storage disks and importing disk groups. Instead, it uses what Symantec calls "multi-node disk group access," which means that all the server nodes in a cluster can see the storage all the time, Angspatt explained.
Other features of the new version include automated end-to-end disaster recovery for Windows Hyper-V virtual environments and Hyper-V Storage Live Migration, which moves virtual storage assets as needed in a data center. Angspatt compared it to VMware's VMotion for moving virtual servers to different physical servers.
Sensitive customer and business data is scattered in hidden corners of your infrastructure. Find and protect it before it winds up in the wrong hands. Also in the new issue of Dark Reading: The practical side of data defense. Download the issue now. (Free registration required.)