I thought I would leave alone for at least a while the Hostway data center migration saga. Some time has passed and yes eventually people will forget about this incident as they have with many others (maybe not of the same magnitude) – but did everyone really think it could be successfully pulled off?
When I heard about this incident all I could say was “WOW”.
We have been in the infrastructure, and utility computing business for a few years and early on thought that we were extremely smart folks and could migrate 225 servers overnight, within an 8 hour window with little to no extended interuption to any of our early clients. Boy, were we wrong!! We had provided the clients their new IP’s, setup the new infrastrucutre, configured the racks, routers, switches and all the space, and the plan was to unplug and then plug back in at the new facility all of a 15 minute drive away. Let’s just say that didn’t work out so well…. That 8 hours grew into an extended 48 hour workday!!
The “WOW” response when I heard about the HostWay incident was when I read:
“The planned July 27 data center migration at ValueWeb, a Hostway company, involved moving more than 3,700 servers 270 miles, from Affinity’s Miami hosting facility to a Hostway data center in Tampa, Fla., according to Rich Miller, reporter for Data Center Knowledge, in Lawrenceville, N.J. “
What were they thinking?
These guys are not dumb guys. I believe after merging with Affinity Internet a few months back they now have operations facilities in over 10 countries with thousands of clients and managaing thousands of servers.
But come on. While hardware is running, and deemed to be reliable it is often left best untouched. The old saying applies here, “if it ain’t broke don’t try and fix it”. When you power down and try to then bring back up boxes there is always going to be a % that will fail.
I am sure they planned for this but with 3,700 servers being moved at once, I don’t see logically how they expected to deal with the consequences all at once or within the time frame they allotted or thought to be tolerable from a client’s perspective.
Anyways, I feel for Hostway and for the clients that it effected. The 8-12 hours of planned downtime, seemingly turned into a nightmare for many clients, whose business reputation relies on their provider’s uptime and reliabitlity.
After our experience a few years back, as I stated above, we swore that we would look, build and develop something that would never make us have to move a physical box from DC to DC ever again. **(unless a client specifically has requested a move)
A main point we focussed on was being able to move data and applications from one server or facility to another without the requirement to migrate the underlying gear. With the launch of our VPDC (Virtual Private Data Center) powered by Applogic we have enabled clients to do this. In the age of SaaS companies, thousands of new applications being put online each day and uptime being an essential part of conducting daily business the expectation of anything less is sub par.
A great example of this is Albert Wu who was faced with a similar facitlity to facility migration.
“The migration turned out to be an extremely simple process, that basically involved issuing a migrate command on the old grid, instructing it to migrate the app to the new grid. The command took some time to finish executing, as it had to move quite a bit of data. But once completed, all we had to do was reconfigure the app on the new grid with new ip addresses for the app, the gateway, netmask and dns ip address, which only took a few minutes. Amazingly, we were then able to fire up our grid app on the new grid! No hardware issues to deal with what-so-ever!”
We continue to focus of how we can provide offerings to our clients that allow them to focus on their business, and not worry about the underlying infrastructure, where it is located, is it up or down, is there a bad drive etc. Our clients need to be able to provision additional resources on the fly when their applications calls for it, be able to shift and server data and applications to and from mutliple geographic points while not having to worrying about migrating anything physically. We are making the hardware agnostic, and providing a scalable platform where migrations and growth are only a few short keystrokes away.
I expect to see many more players pushing hard in the marketplace – bringing many new virtualization and utility offerings into the market this year. Ifour competitors do not I believe they will be falling short of clients expectations and requirements.
The dedicated server landscape is changing and everyone involved needs to innovate, adopt and grow.
All good things to come….