Blade Watch

I would like to thank ‘Mark’ from one of financial institutions which shall remain nameless. He’s based in London in the City/Canary Wharf area.

As ever, I’ve removed any references that might identify him, he met me for a Diet Coke and munchies in All Bar One.

Can you describe your role?

Sure, I manage the Windows server estate which comprises 500 production and development servers.  This is mainly comprised of rack servers (DL380’s/360s) but we’ve also started buying blades mainly to test the platform.

What services do these servers run?

We look after a mixture of Windows infrastructure (like DCs, DNS and email/file servers) as well as those application or batch servers. They run a mainly the Windows 2003 operating system.

Explain a typical day.

The first thing to do is to read the overnight report/check the helpdesk queue, this shows the number of issues we had overnight, from this I can usually tell what the day will be like, to be prepared for any meetings/conversations. Next to see how the morning checks ran, if there were any issues that might need to be highlighted or resolved.

Additionally there are the administration tasks that I need to perform, organizing resourcing for the week or day, arranging the workloads and continuing the message to the team to meet the SLA, to respond and resolve our helpdesk queues within the service level agreements.

The main challenge we have is one of resourcing, we have the people, it’s mainly a chargeback thing. I have to account for the team between new projects and support, thus keeping track of requirements – who’s needed to build what servers, work on which project and for which business line can be time consuming. As we scale up the workloads and servers we deploy (particularly as we move to a virtual world), I can see this becoming more of an issue.

What are your current challenges within the IT business line?

Several things, we get pressure from the IT Management teams to improve the response times to incidents, to be more service focussed. At the same time there is also pressure to try and reduce out of hours work as this starts to become an expensive exercise, there is an understanding that weekend work is the cost of doing business, I think it’s more a limit our exposure to it where possible.

We’re now getting pressure from the datacenter manager and the CIO about data center utilization, they were asking if we could evaluate how many ‘infrastructure’ servers we had, if we had too many and could decommission/recycle them as an effort to be more efficient.

We sometimes get involved in conversations about what should be paid for by IT, what should be paid for by the customer. Who pays for it is a mute point. You’ll pay for it directly/in-directly anyway – increasing the server count for infrastructure servers, or additional costs through extra monitoring license costs.

What are your current challenges from the customer?

Mainly cost and meeting expectations. I understand this, we need to focus on delivery and at the same time manage our cost base, doing this can be a tricky thing to achieve, we can’t keep everyone happy all of the time.

Cost is a difficult one. I can strip out costs. I can have my engineers work 9-5, to avoid or reduce the amount of weekend or out of hour work we do, but doing this can be seen as being obstructive. Obviously there are those systems both infrastructure and application that I can’t take down during the day, email for example, or the batch server. There needs to be an understanding that we can provide the level of service, of commitment that the business need, 24/7 operations as required, but this requires investment in our people and the infrastructure.

Managing expectations can be tricky, we had a business line phone up one Friday asking if we had anyone to reboot our server at 4am on Sunday morning. There was general surprise when I said we’d need to get back to them. “But you’re on call 24/7? How can you need to check?”.

In that respect we are always on call 24/7 indeed, but I need to manage my people, I need to see who’s available, see if anyone wants to do this work, before I ask/request an engineer be available to do the work.

Related posts:

1. Please end the email debates – life really is too short I had a chat with a service delivery manager who...
2. VMware’s vCenter Site Recovery Manager VMware PALO ALTO, Calif., October 5, 2009, VMware, Inc. (NYSE:...
3. Talking about Windows Server 2008 Foundation zdnet With a nod to the new economic realities pushing...
4. Choosing your next server – some tips Check out this article talking about how to choose the...
5. A few things for HP Systems Insight Manager to make it better I was asking Chris in sunny Canary Wharf how he’s...

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