Marathon introduces software to protect VM from physical failures

Ashlee Vance has a piece at The Register on a new product just leaving beta from Marathon that protects virtualized servers — possibly running business critical apps — from failures on the physical hardware.

…Marathon thinks it can encourage customers to move what could be considered as business critical software over to virtualized servers. At the beginning, such software might include SQL Server and Exchange applications, but eventually customers could put things like financial and retail software on virtual systems as well.

As you might expect, the Marathon everRun VM architecture requires a customer to run duplicate systems, so that one can pick up where the other left off when failures happen. We’re told that the failover magic can be applied to servers without any changes to existing software.

I’m not sure that virtualization will emerge as an acceptable solution for protecting multi-thousand processor jobs from the inevitable loss of processors on our ever larger machines, but it is useful to keep up with how other people are solving these problems.

Comments

  1. Cool story – thanks for posting this. It is interesting, as there are a lot of different sub-cultures of hardware that we can learn from.

    In fact, NEC markets the hardware version of what everRun does with software. They do this through lock-step processing capabilities in the Xeon and Itanium platform. I imagine that this could be applied to large scale solutions where single-image multi-core servers have an advantage over node-based compute clusters (if there is one – I’m not saying there is).

  2. Scott – thanks for the response. I’m often on the bubble in reporting stories not directly aligned with HPC. I want to look into the NEC thing though, that looks interesting.

Resource Links: