Blade Watch

As we move to a more just in time way of doing business, on demand computing, on demand storage, or grid or application availability, we need to not only transform our processes, move to ITIL framework 4, develop new processes for managing service, availability, workload and customer experience. We need to work on internal processes, internal issues surrounding ownership and delivery. In the olden days where it might have been “there’s a problem with server9, can you investigate?” we need to establish when it is worth investigating and whether with the ability to re-deploy a server, an application or both in minutes even seconds, what is best long term?

If our web application version 4.1 leaks, do we care if it will get fixed in 4.5 which is launched in three months?

With the just in time business, the IT infrastructure moves to a commodity, the server, the layered components and database should all be re-deployable on demand, I should be able to say to IT, “I need another 600 blades for Credit, build 3 please” and have 600 blades built with the credit build, Windows 2008, Oracle client, credit services and DataSynapse. In that world how do we manage what is change, what is a request or an incident or even a project. If I can in theory rebuild a blade in 45 minutes, would it be a project to upgrade my blades from Windows 2003 to 2008 when an engineer can simply click cabinet 7, enclosure a, rebuild with credit Windows 2008 job and simply monitor it?

• How granular do we make the billing?
• Do we really want to be debating 15 minute intervals?
• When a business line takes over their own infrastructure, how does that work in terms of reporting, accountability, billing and delivery?
• Does the blade guy for Credit IT have to manage his help desk queue, his SLA like the generic infrastructure teams and should the two be comparable? 
• How do we manage the want for individual business units to take the elements of the IT function they want whilst still providing all the bits you might want or need but not want to pay for, without being emotional?

You might not want to pay for a domain controller, an active directory infrastructure, SMS and Exchange, but somebody has to, regardless of who and how’s it provisioned. In essence how to we keep good people both sides of the organizational fence, keep evolving the process in line with our business and what new processes or best practice are we working on?

As we move on to organization 2.0, carbon neutral, ISO certified and so on, how about measuring end user expectation and rating that? It’s all fine saying we meet ISO certification 9001 for example, but what value does that represent if?

• A) A new joiner might take three weeks before he/she gets a desk, a phone, a working computer with a log in and the applications they need?
• B) When it might take two business days or more to rebuild a pc, an activity that with the right technology AND PROCESS in place need take an hour or two?
• C) If those designs and strategies are just those, designs and ideals, but not representative of the infrastructure we have.

Related posts:

1. The cloud is ready for the enterprise but is cost centre driven enterprise ready? Zdnet Blogs More enterprises are looking to the cloud compute...
2. Server sales to continue regardless of the economy Business Wire The global IT industry continued to be plagued...
3. The IT might be ready for cloud + virtualization but is your business? What is core to your business? Network World One-third of 1,200 organizations (33%) plan to convert...
4. Four things to think about with blade server deployments What technologies are you going to be using to...
5. Emulex LightPulse 8GB adaptors HP BladeSystem Matrix ready Emulex HP Technology@Work – BERLIN, April 20, 2009 – Emulex...

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