During my time at BT I set the Data Centre team the challenge of getting Asset Management right, the team thought I was just being stupid and unreasonable but buckled down and with a bit of hard work got it close to 100% right with processes to keep it right. As a direct result of this effort, we were able to switch off 10% of BT’s UK Server estate and save $7M in electricity costs.
So how was it done? We got the assets right, got the rack layout right and then started focussing on who owned the servers and what they were doing. I took a very hard line approach and promised to switch off all servers that did not have an owner and a purpose. There was a lot of brinkmanship, change control items and angst but everyone got the message loud and clear that kit was being switched off and removed from the premises unless it was in active production.
The programme started well, with some good gains in the first few weeks and then we had an outage for a set of development servers. I guess that there was an expectation that my team would get the blame but it was the owner of the servers who got into trouble. I had taken the trouble to engage our CIO and explain that some of his team were not updating support information for critical applications. He got right behind the initiative and supported us fully. After that it was pretty plain sailing, we just issued batches of change tickets to take out servers without proper owner data either the data miraculously appeared or the servers got shut down.
Proper Asset management is critical to a well run IT Operations shop. Without the right data you can’t support the businesses applications, can’t manage the costs of providing IT and are always on the back foot.
With proper Asset Management rack planning, change control, incident management and financial controls are so much simpler to achieve. In the picture above Glasshouse Technologies integrated their Reflector product with Rackwise Rack Management.