Feature: Delivering on data management

Reliability of access to data will be a core consideration in most law firms’ risk-management strategies. Magic-circle firm Allen & Overy decided to outsource its production centre outside of London – but not without considering the costs and alternatives.

By Andrew Brammer, head of global IT operations, Allen & Overy

In an increasingly competitive professional-services marketplace, clients are demanding greater service levels and engagement fees are under pressure. To protect profitability, professional-services firms are taking steps both to drive up revenue and reduce operational costs. At Allen & Overy LLP, this has led to a number of strategic initiatives, including ‘Valuing your Time’, with a focus on increasing revenues; and ‘Valuing your Support’, which is focused on driving down operational costs. This article looks at one aspect of operational cost: the provision and maintenance of the data centre providing the computing power needed to run the majority of the firm. The drivers behind outsourcing the data centre are explained, alternatives are examined, and outsourcing is set in the context of a wider sourcing strategy. Finally, the benefits and experiences gained at Allen & Overy are outlined.

Technology drivers

Traditionally, organisations have maintained their computer equipment within their own data centres. When operation of technology offered a key competitive edge, or even differentiation in the marketplace, this was understandable. However, technology moves apace and the reliability and availability of the technology, along with the commoditisation of the necessary management skills to provide infrastructure services, mean that quality third parties may well outperform internal data-centre management from both service and cost perspectives. Coupled with this are some key technology trends that make it increasingly difficult for organisations to make a business case for continuing to provide data centres internally:

The volume of servers in the marketplace has grown by 12 per cent¹;
To accommodate the ever richer forms and sheer volume of data, the amount of storage required in organisations is increasing on average at over 30 per cent compound per annum;
Data-centre footprints (their floor area) need to grow at ten per cent a year to accommodate this;

The IT manufacturer’s solution to this is to produce smaller infrastructure components (e.g. blade servers), but these require more power to cool their greater heat output.

Environmental pressures

Computer equipment requires electrical power both to run and provide sufficient cooling to keep running. This simple equation drives data-centre design, size and costs. Computer equipment is normally housed in racks (shelves). As computing capability has increased (i.e. chips have become faster and computer disks smaller and faster), the key economic metric within a data centre is shifting from the real estate to the energy bill. In the US market, floor space in a data centre is around $12-15 per square foot compared with an energy cost of $60 to $80 per square foot for high-density computing racks.

In some large organisations the power bill now outweighs the IT hardware bill, and with rising energy costs and increasing IT infrastructure densities, in the medium term energy costs are projected to equal 30 per cent of a larger organisation’s IT budget. Data centres built within an office will have a high real-estate cost, and so may well require additional cooling technology and infrastructure to allow higher-density infrastructure to run within a confined space. Allen & Overy has a disaster recovery (DR) facility within the London office, which has in-cabinet cooling to cater for very high densities in a small physical footprint.

However, longer-term regulatory pressure may well come to bear on energy-intensive data centres in city centres in the medium term. A balancing act will need to be struck. If IT is to remain an agile provider of service to the business, the agility of the data centre is just as critical to the business as the IT applications that support key business processes.

Business risk and regulatory pressures

Over and above the technology and environmental pressures on the data centre, however, there are business risk and regulatory factors to consider. When Allen & Overy looked at London in business-risk terms, it saw a global capital city with relatively poor power and transportation infrastructure. London transportation strikes, although not that frequent, cause losses valued at £100m^2,3when they occur. Added to this are the disruption risks linked to the UK’s robust foreign policy. In pitches, clients are now frequently demanding reassurance that these risks are being managed.

Service delivery and IT cost

As with most organisations today, clients are demanding greater levels of service, which is mirrored by the levels of service required of the IT department. More and more organisations now require 24/7 operations, with minimal or no interruption to IT service. Couple this with the need to reduce IT costs through consolidation of infrastructure and you realise data centres can start to take on a significant scale.

Moving to 24/7 operations requires a lot from the IT infrastructure, as well as from the data centre and its supporting building facilities. Redundant power supplies; power sourced from separate providers; and engineered solutions to avoid the mandatory building power-down tests, all require large investments and significant levels of expertise to manage them. Absorbing these costs, and managing the risk and expertise needed to ensure services meet service-level agreements (SLAs) mean a lot of organisations decide that hosting a data centre is outside their area of core competence.

Different models

The data-centre outsourcing market in Europe is very significant⁴, with the players in Gartner’s Magic Quadrant driving a market of 350 data centres, 250,000 servers and a turnover of around €10bn. Organisations considering outsourcing their data centres need to analyse their requirements carefully and consider the different models available. After an analysis of internal costs, the risk profile and service-delivery model, the options range from making no change – where outsourcing does not make business sense or clashes with core values – to full third-party centralisation, standardisation, and management of infrastructure and business processes. The shades between these two extremes cover approaches such as:

Data-centre hosting. Building, power and cooling facilities are provided by a third party while the customer continues to run the equipment. With this model it is common for additional services to be offered (remote hands, monitoring, building of infrastructure etc.);
Remote data-centre management, where the customer retains their own data centre but it is run by a third party. This protects the existing capital investment but customers need to be sure their data centre will support requirements in the medium to long term;
Platform or application outsourcing. A slice of hardware (i.e. a mainframe) business process or application is outsourced.

The geographic dimension also needs to be considered.

Onshore. Where an operation is moved into a cheaper area, i.e. from city centre to the suburbs. This is a low-risk approach offering wage savings of between ten and 20 per cent, and property savings of between 20 and 40 per cent. The main risks are the lack of availability of relevant skills in the cheaper locations and, if sited too near the main business, a lesser reduction in business-continuity risk;
Near-shore. Where an operation is moved from the central business district to a location within the country boundary or just beyond it. Again this is a low-risk approach. Wage savings may vary from 15 to 50 per cent; property savings from 30 to 60 per cent. In addition it provides business-continuity-risk mitigation. The main risks are the lack of relevant skills;
Offshore. Where an operation and its related positions are moved a considerable distance from the central-business district to another country (e.g. from New York to India). This is the most radical change. The wage savings offered range from 20 to 90 per cent and the property savings range from 50 to 95 per cent. There is also a large pool of resources to draw upon, but the geo- and socio-political issues in the region can pose a moderate risk to business continuity.

Ownership options

In terms of ownership, the outsourcing options include:

Wholly owned. An option used when significant scale can realise the greatest cost savings. This has the advantage of absolute cultural fit and control is retained within the organisation. On the downside, it has significant start-up costs and requires local/regulatory issues to be overcome. If scale is not significant, this is potentially the most expensive option;
Joint venture. This requires an absolute understanding of what the other partner(s) bring to the deal. It requires significant management and there are risks surrounding IP protection;
Outsourced. This is the lowest-cost option. The main advantages outside cost are ease of set-up and the need to have no knowledge of local/regulatory market conditions. Risks to watch out for include possible training requirements and poor cultural fit with the customer.

Arriving at the best solution for the business will only be achieved through adopting a structured
approach to the analysis of the business drivers; service and technology requirements; detailed risk and market analysis; and a full examination of all the options. Outsourcing the data centre is the most mature outsourcing option but it must be carried out as part of a strategic business-driven approach, rather than a tactical reaction to operational or short-term challenges. Outsourcing arrangements that fail tend to be those that were focused solely on short-term costs, or those where the implications of the decisions were not fully worked through. It is essential to prepare and present a full and detailed business case before proceeding with any option. Ultimately the decision to outsource the data centre should be viewed within the context of the firm’s strategic objectives, and preferably within the context of a broader strategic sourcing policy and supplier-management framework.

A sourcing strategy⁵ starts from a clearly-articulated business strategy and looks at how this can be executed effectively and efficiently through the optimum blend of internal resource and external suppliers. Building this strategy requires organisations to understand their own internal capability; their external supplier capabilities; the organisation’s maturity in sourcing new services from the marketplace; and the ability to manage new supplier partners.

Approaching the delivery of business strategy in this fashion will result in a decision as to whether to outsource a data centre as part of a broader set of decisions or to enable a particular set of strategic business goals.

Bibliography

A Message from Data Center Managers to CIOs: Floor Space, Power and Cooling will Limit our Growth, 21 August 2006. Gartner ID G00142393;
One-day rail strike ‘cost London millions’, Press Association, Wednesday 30 August, 2006, The Guardian;
The best public transport system in the world? Tell that to London’s long-suffering commuters, Wednesday 30 August, 2006, The Independent;
Magic Quadrant for Data Center Outsourcing Services, Western Europe, 2006, 17 March 2006, Gartner ID G00138334;
How to build a sourcing strategy, September 2002, Gartner R-18-1009.

Andrew Brammer is head of global IT operations at Allen & Overy. He can be contacted at andrew.brammer@allenovery.com