Storagebod Rotating Header Image

frAgility

I keep hearing things about making infrastructure agile! Now is it just me or is the word agile replacing the word planning? For the record, I do believe that a good infrastructure should be able to meet the demands of the business; it should be able to ebb and flow with demand. That potentially is where the cloud has a place but I am not sure that it is entirely best practise to be able to meet random demands of badly run projects who turn up and demand infrastructure to support a project that absolutely must go live *NOW*.

Unfortunately, I think that is where the agility road is taking us. I have for my sins run an Agile Development team; we used XP (Extreme Programming) and it worked really well but I'm not convinced that the disciplines introduced in the various Agile Development methodologies are entire appropriate in the infrastructure world. The iterative approach taken during development does not lend itself well to a well-managed infrastructure but yet I keep hearing the words our development is agile, why isn't our infrastructure.

Infrastructure by its sheer nature cannot chop and change in a iterative fashion; it will always take time to deploy, it always requires planning even if it is to flex with business plans. Unit testing, regression testing in a large, complex infrastructure is not especially feasible. And as for 'mocking up' interfaces, best of luck!

 Even if you go cloud, you need to get contracts in place with appropriate provider; if you go to a Capacity on Demand model, you need to get the correct model in place with your supplier.

Agility does not simply mean that you can throw infrastructure around without planning; it does not mean you can simply change your infrastructure without planning. Agility should not replace planning if it does; Agillity==frAgility and we are all doomed!


2 Comments

  1. Dave Rooney says:

    There are a couple of approaches that may be able to help you out in the infrastructure space.
    First, in order to deal with the “must go live NOW” issue, the infrastructure people must be included as at least peripheral members of the development teams. They could attend the Release and Iteration planning meetings and even the daily standup meetings. That greatly reduces the risk that the infrastructure folks will be blind-sided by random request. Same goes for Facilities, HR…
    Another approach that can help is taking a page from Lean and using Kanban or pull-style management of infrastructure work. Have a look at http://blog.crisp.se/mattiasskarin for some ideas.
    Hope that helps!
    Dave Rooney
    Mayford Technologies

  2. Martin G says:

    Ahhh, your approaches in a sane organisation would work! And yes, it would be really good to have infrastructure people involved at an early stage and during the iterative process of development. Not always easy in large and dispersed organisations.

Leave a Reply

Your email address will not be published. Required fields are marked *