The newest version of Symantec’s application high availability product for virtual environments provides tight integration with VMware’s disaster recovery capabilities and a dashboard for monitoring and managing multiple applications through the VMware vCenter Server console. Application High Availability (HA), which came to market last fall, brings deeper application awareness to VMware so that failed apps can be quickly detected and restarted.
Ensuring application availability is the greatest challenge in managing virtualized servers, according to an IDC survey cited by Symantec, and is generally recognized, along with security, as an impediment to virtualizing business-critical applications.
"The biggest challenge within the context of a virtual infrastructure is to monitor and take corrective action, specifically from an HA perspective if applications fail," says Ashish Nadkarni, senior analyst and consultant for Taneja Group.
VMware HA provides high availability at the hypervisor and virtual machine level, but not for the applications running on the VM. There are several problematic scenarios that Application HA addresses, including infrastructure failure; the VM is running but the app is down; the VM recovers after a server failure, but the application doesn’t; the application is running but is not functional. Application HA monitors application health--that is to say, it does not measure application performance, but rather defines and monitors its service group dependencies in the VMware environment. It quickly detects failures and automatically restarts the app. It can trigger VMware HA for additional recovery if that is required. The goal is to keep critical business applications running and minimize downtime.
"Think about it this way: Why did people cluster applications and provide HA for applications in a physical environment?" says Nadkarni. "In a virtual infrastructure, there is no such protection. VMware HA does not have much application intelligence." So, VMware HA will restart the VM if it fails, the host server is restarted or if the VM is brought up in another location in a disaster recovery scenario. But non-functional applications have to be tracked and manually started.
Lacking a tool such as Application HA, enterprises might use OS-based clustering in the guest, Symantec says. However, this prevents them from taking advantage of key VMware capabilities, such as vMotion, dynamic resource scheduling and distributed power management with VMware Distributed Resource Scheduler. Scripts require development and maintenance and slower recovery than Application HA, Symantec says.
