Thu 15 Jun 2006
Marathon Technologies offers an interesting fault tolerance solution for Windows using virtualization technology. Server virtualization usually means that you have multiple virtual servers running on one physical server. EverRun FT is software that does the opposite. It unites two physical servers to one virtual server.
So its purpose is not server consolidation but fault tolerance. This sounds like server clustering. However, there are some fundamental differences.
The biggest advantage of EverRun FT compared to clustering solutions is that the applications don’t have to be cluster-aware. Furthermore, there will be no failover if one of the servers breaks down, i.e. the virtual server will continue running without any downtime.
The disadvantage is that both physical servers have to be absolutely identical. You even have to patch them simultaneously. During the patching process you will have a short downtime.
This technology seems to be very promising to me. The price of $14,000 is not so promising though.
Newsletter
Loading ...
Hello,
I have heard a product called Virtual Iron. I thought that did the same thing. I had been scouring the net looking for a Free and possibly open-sourced solution. I thought a while back that XEN would have been in a way the answer (funny its not hardly mainstream news anymore) but that is not quite following some of the hype I had read at the time.
The real big possibility people on the low end side look for is getting something like 5 or 10 $100 - $150 computers together in a ungodly seamless setup like what EverRun Proclaims. Now I will keep up the hunt and see what this is all about. Good day.
Yemi Bedu
Hello,
I just finished reading a good number of the product features, requirements, and FAQs. The setup seems to be a one on one, active-passive server setup. one virtual server on one physical server that has its data hot-updated with the passive physical server. When needed, the application in the stand by takes over the execution state.
That wasn’t what I was looking for entirely in my previous comment. I was looking for resource clustering on the device level for a massive computer. This is doing a sort of RAISE (Redundant Array of Inexpensive/Invaluable Server Environments) level 1 (mirroring) setup. If it could happen like I wanted, I guess it would be RAISE level-0 (striped, resource doubling of slower/smaller/weaker components in the mix).
I would say though, this is a really tight setup for its focused audience group. I wonder how you could get VMware Server or MS Virutal Server to run like such a product. One of the obstacles that I know at least can’t happen with common setups of Virutal Server and Windows Clustering is Termincal Services fault-tolerance of Sessions.
One thing also in your comment in the paragraph below last: “The disadvantage is that both physical servers have to be absolutely identical.” I read in the FAQ that they said that for FT only the CPU and RAM configuration have to be the same, the network should cards need to run at the same speed, and that the hard drives SHOULD to be of the same size to not loose storage space on either side. Good day.
Yemi Bedu
Does it require a shared storage? Is it compatible with window server 2008? Or does it support VRRP?