Why are you still running your own data centre?

The time has come for all businesses to ask the question “why am I still owning and running my own data centres and server rooms?” rather than “why would I outsource my data centres and server room?” In other words I think you need to justify your continuing ownership and operation of such facilities.

Let’s be clear, this is not the same as owning, managing and operating the IT hardware and software. There may be a number of very good, strategically important reasons why you would want to continue to do that, and the decisions may well vary depending on the business you are in, the nature of the applications you are developing and running and will probably encompass elements of both ownership and outsourcing. But whatever you decide there will be a physical data centre underpinning, and you need to understand some of the key implications of the own vs rent decision.

Agility and Speed to Market

The lifecycle of a data centre is 15 – 20 years. Think back 15 years and reflect on how much your IT has changed in that time. It’s not just that the technology itself has changed, putting different demands on the data centre infrastructure, but the technology has also driven and facilitated different business models and extended most companies’ geographic reach. The decision isn’t just about building adequate capacity; it’s about being flexible enough to adapt to change; how quickly can you provision and deploy new applications globally and also, how probable is it that any data centre you design today won’t be what you need in 5 years’ time, let alone 15.

Availability and Resilience

All applications are not born equal. Some are absolutely critical to your business, others less so. But if you are building your own data centres you almost certainly don’t have the luxury of multiple locations like Amazon or Google, so that the data centre will need to be designed to meet the criticality of maybe one key application. Nice to have a Tier III or Tier IV facility, but a very expensive way of running the variety of development and less critical applications you are bound to have.

Connectivity

We live in a globally connected world, yet much of enterprise data connectivity was built to manage the Corporate WAN. The availability of high density, low latency networks in the major co-location facilities provide a level of price performance and flexibility no single company owned data centre can match.

Security

Organisations often cite perceived security issues as a reason for not using 3^rd party operators. Yet, go and look at the physical security that is in place at the vast majority of 3^rd party data centre sites and then look critically at your capabilities and what it would cost to provide the same level of security yourself.

Capex v Opex

This perennial business debate is complicated in IT by the fact that, too often, the CIO is not held responsible for the capital expenditure on the data centres themselves and the associated mechanical equipment. This tends to be in the hands of the Estates function. If there was an integrated view of the costs at the CIO level I think there would be less inclination to build your own facilities up.

…and the answer is?

Each business will have different models, priorities and factors to take into account. Some legacy applications, and the hardware they run on, may not be able to find a suitable 3^rd party operator to take them on. The risk profile and appetite of some organisations may militate against some organisations taking an outsourcing decision. But in general the issues highlighted here, and the continuing growth of cloud, hosting and co-location services, suggests that most businesses should be looking at a “rent first” strategy when it comes to data centre facilities.

In future blogs we will look at the sort of capabilities and features you should consider when looking where and with whom to place your IT equipment.

This post first appeared on the old Cassini Reviews website.