The practice of locating data centres remote from primary business operations is becoming increasingly commonplace. Alastair McAulay examines why it may be time for all data centres to leave home.

This trend is, in many ways, analogous to the situation of an adolescent who grows up in the family home, and ultimately needs to move out and find a place of their own, because the accommodation in the family home isn't suitable anymore. And, as with any maturing dependant, the reasons for flying the nest are many.

From the parental side, there is a reluctant admission that it is time for the dependant to move out, coupled with a concern about the dependant's future welfare. Even though they know that in the long-run it could be detrimental to keep the dependant at home, they still find it difficult to accept that the adolescent needs independence. In the instance of the data centres it is more a case of ,'Well maybe we can squeeze just another server in if we get rid of that old DLT tape changer. Also, it is a really bad time to move to a new data centre as we have so much else on with the SAP upgrade.' And perhaps a more pressing concern is the cost - after all, removal vans cost money.

These 'parental' concerns have validity and need to be addressed but the decision point is driven fundamentally by a need for growth and flexibility, and a decrease in the requirement for dependency. Taking each of these in turn:

  • Need for business growth is one of the greatest IT challenges and even though the size of hardware is decreasing, increasing processing power and storage requirements are driving the need for more space with adequate supporting infrastructure. On-site data centres are struggling to meet the space demands that result from this surge, and the supporting infrastructure (such as power and cooling) was simply not designed to cope with the increases in processing density. While data centres remain at home, there is an ongoing battle to either find more space or justify existing space.
  • Greater flexibility of IT is required. As organisations become increasingly agile, they expect the IT function to support rapid reorganisations of the business without any disruption in service. The IT function needs to act as a business enabler, rather than a hindrance - it is expected to act as a professional service unit, supporting global business users around the clock. For many, this shift to a service-oriented function can only be realised when the supporting infrastructure can flex to meet the demands of an agile business. With the space to meet future capacity and features that improve the reliability of the infrastructure, moving off-site will often allow the IT function to better plan for capacity fluctuations and support the business through times of change.
  • Decreasing dependence is evident from the increasingly hands-off relationship that most support areas have with the physical infrastructure. Remote monitoring has evolved considerably in the past five years so that physical checks are minimised with more descriptive alerting and monitoring, allowing incidents to be resolved remotely. In addition, the increasing speed of communications and the falling cost of network bandwidth mean that having equipment at distance no longer requires highly tolerant applications.

It may be time for data centres to leave home

Increased attention to operational risk, as a consequence of service outages from relatively minor IT failures, is resulting in some challenging questions being asked about the ability of in-house data centres to deliver the levels of services required of an enterprise that is now completely dependent on IT. It is difficult for many organisations to legitimately argue that hosting data centres, at their business premises, is a sensible or beneficial approach. Ultimately, logic suggests that precious space at head offices being used for data centres has a number of unfortunate consequences:

  • the utilisation of expensive space, both in terms of financial and opportunity cost, for IT equipment, as opposed to head office staff;
  • lack of flexibility for future office options as these are complicated by additional consideration required for data centre provision;
  • little in the way of organisational resilience in the event of major incidents;
  • potentially more hands-on meddling, circumventing of change control and increasing risk to operations.

In addition, there are many benefits to be gained from locating the data centre remote from business operations:

  • Flexibility - ability to expand and upgrade in the future without affecting business operations. Optimum designs can be progressed for business operations and data centres independently of each other.
  • Enhanced business continuity - reduced risk through the separation of people and systems and through locating the data centre in an area that is not prone to natural disaster (such as flooding) or considered a prime terrorist target.
  • Enhanced security - heightened security measures are more easily implemented (such as access controls, bomb-proof walls, perimeter fencing, faraday cages).

There is clearly a lot more to data centres than unplugging the existing IT kit, loading it into a white van and setting it up in a warehouse somewhere. Choosing (or potentially building) a suitable data centre, commissioning it and managing the migration are all complex activities that take time, money and careful planning to accomplish successfully. This is a high-profile undertaking that the business is critically dependent on being done right; therefore, a strategic approach is required.

Developing a data centre strategy involves taking a complex set of requirements and mapping them to a set of scenarios:

  • outsourced or in-sourced?
  • a single large centre or multiple regional centres?
  • shared with other business functions or standalone?
  • manned or un-manned?

Getting a good understanding of the financial, and other, benefits from each scenario is the obvious step prior to planning a move to a new data centre facility. However, this is not a task that should be underestimated as it requires a complete understanding of current running costs, future service demand, operational and security requirements, constraints imposed on any migration from applications and other IT service availability, as well as an up-to-date understanding of the current data centre marketplace. This information is required to develop a comprehensive cost and benefit model that ties together these variables, thereby allowing an accurate assessment of each of the candidate scenarios.

Even though there are considerable challenges in developing a model and an implementation plan it is worthwhile doing so properly. Data centres, whether they be in co-located facilities or built from scratch, are big cost items. There can be millions of pounds' difference between first and second place scenarios and the right implementation plan can ensure that any move occurs without disruption to the IT services provided to the business.

After all, you wouldn't leave home and move into the first place you came across without a good understanding of what you were letting yourself in for.

Alastair McAulay is senior consultant at IT infrastructure. He has 20 years experience in building and specifying IT solutions in the finance, energy transport and health sectors. He specialises in implementing IT infrastructures that are integrated into an appropriate service management framework.