Bladewatch.com

I was discussing with a CIO friend who’s been evaluating getting rid of their hardware support contract and having their on site engineers do the first line/second line hardware support whilst relying on the standard warranty they buy with the servers to cover the rest.  This got me thinking as we discussed his drivers and how he saw it, and got me bullet pointing some thoughts below:

• Explore the scenarios and what procedures should be followed – for systems out of warranty do we rebuild/repair call out the vendor or simply offer a replacement
• What parts do we keep on site and therefore what is ‘supported’ and what is not
• What migration or handover period is there between now and when the support contract expires
• For applications that are not capable of running on newer platforms how do we accommodate business or application specific service level agreements
• How does IT respond to service level agreements and expectations, if it is in warranty the turn around might be in hours or days, if it’s out of warranty could the actual fix be weeks?
• How do we communicate this change of policy and gain user acceptance of the operational risk
• In environment where there are different teams with different ownership of the infrastructure, how do we handle a component failure, is it the server owner that deals with the repair or a central infrastructure team? Does the global Windows team fix the hardware of the Windows build validation team or not if they are separate cost centers?
• What perceived reduction or transference in costs is going to be realized and how do we equate that with operational risk
• Do we have a process or on going project that identifies systems that are about to go out of support and co-ordinate their replacement and migration on to the new platforms. If so, how is this handled or combined with an operating system revolving upgrade path?
• How do we manage the economically viable repair debate with that of disruption, my file server that’s on a DL380 G3 if the system board fails, is it really worth swapping it out if it’s not harming anyone and how do we explain or force this through?
• How is this managed globally and how do we manage the gold stock, what is the cost of the gold stock against having on site hardware cover?
• As we move to virtualized environments does a system failure not start to have bigger impact than when it was a simple standalone box, one ESX server with a faulty memory dimm could result in a platform or application outage.

Related posts:

1. Operating System Support on HP Proliant Servers Chris called me up and asked if I had a...

Related posts brought to you by Yet Another Related Posts Plugin.