Standardising and integrating in the data centre

By Roger Keenan, managing director, City Lifeline.

  • Monday, 19th January 2015 Posted 9 years ago in by Phil Alsop

Modern organisations have become defined by the information technology they use, and as a result the data centre is where the key assets of any business now reside. Companies in all sectors now routinely devote resources to collecting, analysing and using the information in a data centre, even when that is not the obvious product of the organisation.


Every business is different, and their information technology needs can vary wildly, especially when it comes to how they define and prioritise an IT project. Organisations all have different appetites for risk, integration, stability, urgency and spend. So how do decision makers choose which projects to do and how the user needs should be structured?
 

One way to start is to consider the extent to which the organisation needs or wants to integrate or standardise business processes in its data centres. That has a very substantial effect on information strategy and the degree of globalisation or centralisation the company achieves. Integration involves making everything happen together as a seamless whole. Standardisation on the other hand means making things happen in the same way, but not necessarily together. The two are independent of each other, but not mutually exclusive.
 

Airlines, for example need a high degree of both integration and standardisation – everything has to work together in the same way in every part of the world. The customer experience of someone buying a ticket from New York to London has to be not only the same as the experience of a person buying a ticket from London to Lusaka, but completely integrated with it. There can be only one ticketing system (although it may exist on different servers in different data centres in different parts of the world), otherwise co-ordinating flight connections would be impossible. Both standardisation and integration levels need to be high.
 

As a contrast, take McDonald’s, selling hamburgers all over the world through a franchise operation. The need is for a very high degree of standardisation and a Big Mac in Moscow must be identical to a Big Mac in Okinawa or Kalamazoo. It is not only the product which must be standardised – so too must the customer experience and the information systems. However, there is very little need for integration as all the McDonald’s outlets are separate franchises, run by independent entrepreneurs. So the information systems can all be independent of each other, provided they produce standardised management reporting. The business processes involved in serving a Big Mac in Okinawa have no need to be connected to those in Kalamazoo and there is certainly no need for the systems to run in the same data centres.

While an airline finds it both easy and important to introduce a single worldwide customer loyalty card, McDonald’s would not only find it impossible, but would gain little benefit from doing so. While the standardisation need is high, the integration need is low.

A bank data centre environment represents a different combination. In a bank, there is a high need for business process integration, as regulatory regimes require close monitoring and control of what happens in every corner of the business. Good risk management requires that remote subsidiaries are not allowed to commit the bank to dubious loans which are not visible from the centre. Likewise the processing, recording and monitoring of all loans needs to be the same and capital allocations and ratios are set for the bank as a whole. This means a bank requires highly integrated systems running in its data centres.

The business of lending on the other hand is highly personal. Every loan is different for each customer’s individual needs, and every business plan presented to the bank for evaluation is different as well. Unlike someone buying a hamburger or even an airline ticket, a loan customer needs a highly personalised service. In other words, the need for standardisation is low, but for integration is high.

Finally, there is a fourth combination. A large industrial conglomerate, such as GE, Jardine Matheson or Mitsubishi has many subsidiaries which are intentionally operating in different ways in diverse industry sectors and cultures. There is no need to integrate them, but also no point in standardising their operations either. So long as their management reporting into a single data centre is consistent, they operate independently of each other, so a conglomerate needs neither standardisation nor integration.

The number of information technology projects running in data centres around the world is vast. All those projects serve organisational needs, and all organisations are different. Looking at all those projects in terms of their differing needs for standardisation and integration is a simple, yet powerful start point for ensuring that the information technology projects stay focused on the needs of the organisation they serve.