First, understand that what I am talking about is a complex system with 100's of components and 1000's of technical details, i.e. a medium size IT operation.....
The system is complex because there are dependencies among the components. The Network infrastructure is really a lot appliance computers dedicated to networking, with often remote services from outsource providers. Storage is often handled centrally by a SAN or a NAS and there are usually multiples of them at various stages of life cycle. Then there are other shared resources, authentication servers, time servers, database servers......some can be outsourced as well.
Each of these groups have multiple parts or future possibilities: speed, size, expandability, power consumption, space consumption, robustness, life cycle (depreciation) and technical compatibility with other components.
Here is an example of what I mean.
We want to expand one type of storage. This storage needs to be in a specific location because of networking constraints. That location has space but no power available. There is a set of servers in the space that are being retired and eventually replaced by servers in another location. That's where the power will come from, retiring these servers. To do that we need to schedule downtime with the users. Their next window is three weeks out. In the mean time, we realize that this server we are moving can't move because it's dependency on a local SAN storage is not moving with it. That's not moving because another server is using it that can't move in the same time frame. We now have to schedule a replace storage expansion at another location. That's going to put us beyond the first three week window, and the next window is a month away. So can we get this storage expansion (not the same storage as we started with) done before the next downtime window? And guess what? We are back to the beginning of this process, but with a whole different objective. And on it goes.....
No comments:
Post a Comment