Bladewatch.com

I was having a chat with Chris (he’s in a not so sunny Canary Wharf today), he was mentioning that one of the virtualization enthusiasts, their architect was embracing the concept of everything being a virtual machine, that we’ll have one big virtual platform. Chris was asking my views (he disagreed and thought it was rubbish, so here’s a summary of my telephone based response which was had over lunch on the phone).

“It depends, we need to move to a virtual platform, it’s the easiest and most efficient way of providing business requirements at typically the lowest overall cost.

At the same time, I get nervous about such things, and about making statements without business buy-in or sign off:

1. Are we ready for a virtual world – how’s the storage/network – can you provide my virtual machine with that proprietory connection to the markets for the ION application?
2. Is the reporting/billing process ready – many organizations have enough trouble reporting how many servers we have, what they’re connected too and who owns them, how will we cope in a virtual world and how do we respond to ‘virtual server sprawl’?
3. I don’t see this mega virtual farm – will it not be a range of component possible business aligned or regional server farms which are divided in line with requirements, we need to appreciate individual needs, and ‘high maintenance users’ where the savings are less than the emotional trauma of debates about segragation and ownership.
   1. How do we deal with support issues and priorities – we might have individual requirements per business line or application, trading systems might only have a few virtual machines per server to avoid any allegations of ‘slowness’ or to maintain agreed levels of performance/stability.
   2. How do we deal with different subsidiary departments or business needs – what works for fixed income might not work for the electronic trading or HR departments.
4. Who has asked the business what it is they want and where they see organizational and IT transformational changes being made – it’s fine saying one massive pool, but how are we strategically developing – 24/7 support by application teams, or by region or by follow the sun support? How does that all work in terms of access and standards, the build might be the same, but the configuration, the SAN or the network might not be.

We need to be enthusiastic about the platform, meet the business needs using the technology, but we have to do so in line with budgets, operational and application constraints. We need to achieve as much as possible at the least possible political and organizational cost for long term success.

No related posts.

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