To SAN or not to SAN
One of the infrastructure projects that my team is undertaking this quarter is to expand the storage capacity at our production data center. Currently we run all locally attached storage, and since it has grown by 60% in the last year, it has become a nightmare to manager over the last couple of quarters.
One of the solutions that we are looking at is installing a SAN/NAS solution. We are currently evaluating a number of solutions, but at this point, it looks as if EMC and NetApp are the strong front-runners. I have had experience with both of these vendors in the past (we currently use a EMC CX500 unit at the corporate office), and I have found that their technologies and offerings do not differ a whole lot.
The issue is that the price of disk has come down so much in the last couple of years, we could almost build our own SAN/NAS solution out of a couple of clustered linux boxes and some decent JBOD shelves (Dell Powervaults come to mind). Add to this the promise of GFS (Global FileSystem) and you have yourself a pretty robust and reliable storage solution for a fraction of what the SAN vendors want to charge.
I see the build vs. buy issue come up more and more these days as the cost of hardware drops and the ability to "roll your own" increases. It used to be that if you wanted anything (server, disk, network) that was more then "workgroup class" hardware, you had to partner up with a VAR and pony up some serious cash. To point, when I was with Excite@Home, I ran a project with a capital budget of 16 million dollars for hard/software. If given the same project/budget today, I could do it with a fraction of the budget whilst delviering the same level of capacity and reliablity.
Isn't technology just grand?
7:26:16 AM 
|
