After listening to a recent Packet Pushers podcast, the question was raised about network design and where the network design should start. When I first started designing IT, I believed the correct starting point was to define the required outcome through the deployment of technology.
For example, the requirement from the customer could be to “Upgrade our mail server” which was converted to “Deploy Microsoft Exchange 2003” by someone in the supply-chain. I would therefore start the design by looking at what hardware I would need, what software was required, who was going to use the system, etc, etc.
These are perfectly valid ways of approaching the design of any IT system, but I do not believe they are the correct starting points.
Regardless of whether you work for an integrator, supplier, or internal IT team, the starting questions should be the same. The questions are business-related rather than technical, and sometimes it takes multiple questions to get the right answers.
- What do you want to provide to your users?
- What are you hoping to achieve with this technology?
- How will this increase productivity/profitability?
Most projects will have been given to you by sales/management/the customer with what is believed to be the correct business or technical requirement. On closer examination, you may find that for various reasons, what the customer actually wants is not what you are tasked with deploying.
In the above example, had I gone back to the customer and further analyzed what they believed was their requirement, I would have understood that the disks in the mail server were too small and that simply putting in more disks would fix the problem.
Caveat: This analysis is not always possible and sometimes it requires you to speak to a multitude of people before you can get the correct answer. It is also very probable that you might not ever get the answer you require.
BUT, if you are able to ask the questions above to the correct persons, then it can pay dividends for you in the long term.
This may be positive (save your internal IT budget and get you a promotion, Yay) or negative (loose a contract and get you fired, Boo).
You will however, be able to console yourself with the fact that as you have successfully qualified this project, your chances of having to re-design at the 11th hour and not sleeping for 3 days straight will be severely diminished.