Start with Business Requirements, not Technology

This is the feedback I got from someone who used ExpertExpress to discuss the evolution of their data center:

The session has greatly simplified what had appeared to be a complex and difficult undertaking for us. Great to get fresh ideas on how we could best approach our requirements and with the existing equipment we have. Very much looking forward to putting into practice what we discussed.

And here’s what Nicola Modena (the expert working with the customer) replied:

As I told you, the problem is usually to map the architectures and solutions that are found in books, whitepapers, and validated designs into customer’s own reality, then to divide the architecture into independent functional layers, and most importantly to always start from requirements and not technology.

A really good summary of what is all about ;) Thank you, Nicola!


  1. Your business requirements are all fine and dandy until the responsibles see the price tag. Suddenly they forget about the requirements and go with the cheapest. The requirements are only worth the paper it's written on. I don't know in which world you live in.
    1. Have to agree with @Anonymous.

      The architecture is often driven by the Excel sheet. Seen this many times: the product is selected first because it simply matches the budget and we (designers & architects) are required to implement the solution matching original requirements.

      Worked for many big projects both for US-based and EU-based companies. I am talking about non-military projects - those strictly driven by business case.

      I wonder if someone lives in 'ideal' world (the one from the books).
    2. I have seen the ideal World on vendors presentations;)
    3. price is a business requirement too.
      Customer will not choose overengineered solution against business requirements.
      For example technically it is good to have redundancy. But if RTO is large enough customer will never pay for redundancy
    4. When you cooperate with the design & requirements with big vendors very often you are "suggested" to use specific hardware (as a part of the vendor strategy) regardless of the real needs here.
      You are "encouraged" by the better discounts to promote certain solutions.

      I do not want to share any further details (vendor names, project names) - I took part in many big deals and believe me - the reality is far from the ideal imagination.
    5. As @Mikhail pointed out, price tag is a business requirement. The budget the business is willing to invest into the critical infrastructure that runs the business is a business requirement. If you can't persuade the budget owners that your solution is worth their budget you have an obvious problem... and it's not technical.

      For example, "redundant WAN links" is not a business requirement. It's a means to an end, and may or may not be justified based on what the real business requirements are.

      Then there's the "negotiation" phase - based on business requirement you come up with a solution. They balk at the price tag. No problem - you can change the solution **when they adjust the requirements**.

      Now, all this works in imaginary world (at least for Grumpy Anonymous) where the motivation of the network architect is aligned with the motivation of the business (getting the most bang for the buck). If you allow a vendor to design your network, you get the price tag you deserve. Worst example I've seen: non-redundant Nexus 7000 (because there was no budget for two of them) acting as a ridiculously-expensive patch panel.

      That's why some customers go for independent consultant and/or design reviews. I understand that not everyone believes in the feasibility of this approach. I have no problem with non-believers, we're too busy working with people who see value in what we do.

      Finally, there are the customers going for the cheapest. We had a few of them in the past, thanked them for their interest, and moved on. They're a waste of everyone's time.
    6. 'When' is a conditional and nothing more. That's probably why you became a consultant so you don't have to literally deal with the budget owners anymore. It seems you were absent from reality for too long. The management decides to introduce the next big bang (it has to be at least as nice as AWS). If you can't deliver in budget and time you know where the door is. They give a damn about your art of persuasion. And now tell me another fairy tale.
    7. OK, so you're not invited to the table when decisions are made, and you get the ridiculous marching orders from people who have no clue what they're doing (according to your opinion). Welcome to the real world.

      Have you ever tried to figure out why that is and whether you can change that (maybe by moving on)? I would strongly recommend you read some of this:
    8. I would be constantly moving if I would do so. It's not an option. The only thing I do is to take this humanity with a smile.
  2. One of my colleagues purchased VMware vSphere licenses with NSX bundle, which was promoted by some external licenses. Now they ask me whether they can do X, Y and Z with it... Strange world we live in...
    1. external consultants, not external licenses, sorry...
    2. I feel sympathy with you, namesake. They probably ask if they could do the same things as with AWS because they like it that way. Welcome to the beautiful new world of human stupidity. The only possibility to remedy is to take the red pill.
    3. Yeah, been there, seen that. Even worse, I've seen very smart teams (who honestly try to do everything right) start at infrastructure instead of how they'll deliver the services to the business.

      Not new either (as in "stupidity and universe are infinite... well, we're not sure about the universe"). 20+ years ago had to deal with the aftermath of someone selling a customer a bunch of totally unnecessary Stratacom switches because everyone knows you need ATM to run MPLS.

      BTW, you might enjoy this video
    4. Really great example with Deutsche Telekom. They stick with Openstack and Huawei but nobody cares. Customers ask if they could to do the same thing as with AWS (you can guess the answer). The only unique selling point maybe data location but that's also gone. So they partner now with Microsoft because they can't beat them. Open Telekom Cloud is a lead balloon.
Add comment