Capacity planning maturity seems like it must be a pipe dream for the majority of IT departments. Unlike other process maturity questions, when it comes to capacity planning process maturity, I am wondering if thinking in terms of levels even makes sense. Instead, perhaps we can consider a different maturity model.
Goal of Capacity Planning
I tend to think about capacity planning in one simple overarching idea:
“Just enough to be good enough – just in time.”
Translating that phrase for many CIO’s means the goal of capacity planning is to have sufficient capacity today that is fully utilized and supporting the service level requirements of the enterprise while anticipating the near term and farther out requirements.
For others, like your CFO, the goal of capacity planning is to achieve a balance in asset utilization levels and the time-to-value of an asset in support of achieving the committed business case and financial returns.
My simple view of capacity planning means we manage the utilization levels of everything from Internet bandwidth to servers (physical and virtual) and storage. It means we focus on procuring additional capacity to meet growing demand without reducing utilization to an unacceptable level.
Overbuying Capacity Planning
It is not always clear in a case of overbuying capacity planning maturity if the issue is one of the IT department simply over purchasing or one of the vendor overselling. If you are confused by this, try thinking about the case of a patron at a bar who may have purchased more drinks than they needed, while at the same time the bartender was over serving them.
There is no clearer illustration of overbuying as one client who spent nearly $2M on a new SAN. Now 18 months later, they are still only using 39% of the capacity with a 970 hours of implementation work remaining. The real shame is when complete, the SAN may only get to 60% or 65% utilization after almost 2½ years.
From the CIO’s perspective, he has planned for growth and a possible tightening of capital funds. From the CFO’s perspective the asset utilization and time-to-value is absolutely unacceptable. Somewhere in all of this is a reseller that may have lost a client and a strong stream of future sales because they oversold the solution and most certainly knew it.
The good news about overbuying is you can’t criticize it for not being proactive.
Over Provisioning Capacity Planning
Over provisioning is the close older cousin to overbuying. It was thought that in the age of virtualization and push button provisioning of resources the practice of over provisioning would have ended.
And yet, despite the capability of data center automation and orchestration technologies, the power of systems management tools, and the ease of provisioning entire systems including storage in minutes, system administrators are still not managing their systems within a utilization band – not too high to affect performance, but also not too low to be wasteful.
Perhaps the effort required to manage utilization within a band is just too much compared to the effort required to simply give it more memory from the start. This smacks of a “set it and forget it” mentality despite the implications for data center power, operating costs, and lifecycle costs.
As with overbuying, over provisioning capacity planning can’t be criticized for being reactive. Over provisioning is also usually repeatable and consistent in any shop that uses this approach.
Throwing Money At It Capacity Planning
“Throwing money at it” capacity planning is the least mature approach but still very common. I consider this the least mature because it is reactive and usually triggered by a problem which means something negative has already occurred with a service.
Throwing money at it can be the most pleasurable form of capacity planning because IT departments use the crisis of the day to get a blank check to solve the problem. This can be like sailors in an exotic port looking for that friendly bartender.
In many cases this leads to overbuying (overselling) as well as over provisioning. Not surprising is the cyclical nature of Throwing Money At It and Overbuying and Over Provisioning methods of capacity planning.
The real problem of being in the throwing money phase is that bad decisions are much more likely as you take a tactical approach to the problem at hand and sometimes try to do more than is needed. When you find yourself in this place you can be assured of remaining in the cycle for some time.
We have all been there. I’ve been there. You make the best decision you can from the information available at the time with the resources at your disposal. You size systems based on the best guess of future demand and likely business changes.
But afterwards, you have to continue to revisit the sizing as part of considering near term and future needs. Ideally, this would be a recurring activity with the ‘owner’ as part of a service management process.
My best advise is to simply start – today. Write down two or three questions you think should be answerable about future capacity and performance requirements. Then set about answering them. Start with what you know using data you can look at using existing tools. And only then, should you consider adding additional tools.