Marathon, which made a splash at the fall 2007 VMworld show with a preview of its everRun VM product for high availability and disaster recovery, will formally announce the product today; it's available in beta form and will ship in April, the company says. The product can be added to the XenServer Enterprise Edition license for an extra $2000, with no limits on the number of VMs running inside that server, Marathon says.

As Burton Group research analyst Chris Wolf noted recently, more CIOs are looking for high availability (HA) solutions now, since virtualization has moved beyond test and development work and into the realm of key business applications. If those apps are running in a VM, the last thing IT would want is a slowdown or outage for business users. The higher profile the app, the greater the need for a high availability solution for VMs. "VM high availability will be a significant concern in 2008 as virtualization technology improvements allow more high-end enterprise applications to run inside virtual machines," Wolf says.

Marathon is betting that its product will appeal in particular to midmarket CIOs, some of whom have had a hard time justifying the cost of traditional high availability/disaster recovery solutions, many of which require the added expense of a storage area network (SAN) and staffers with specialized SAN skills. Another competitive point: VMware's high availability solution also requires a SAN. Marathon's product can use direct connect attached storage, which is a plus for CIOs who don't have a SAN and donâ¬"t want to pay to install and maintain one.

The everRun solution may also appeal to IT leaders in larger enterprises who value simplicity of administration with regards to disaster recovery. Marathon estimates that an IT group can get everRun up and running to protect a server full of VMs in about 30 minutes, whereas traditional clustered high availability solutions can take days or weeks to set up and install. The product also automates fault and policy management, saving IT staffers' time on these chores.

And with what Marathon calls geographic fault tolerance capability, you can set up many VMs at one data center site (or several sites) to have automatic fault tolerance with another geographic site; this would keep production apps running in VMs up and running even if one location experienced a natural disaster such as a flood.