Yet today's IT infrastructure is highly complex, typically multi-platform and supports 24x7 business critical, end to end processes creating new demands for the efficient data centre.
CIOs will never achieve their cost reduction targets if the data centre requires an army of highly paid, highly skilled individuals to keep it up and running. It is only by leveraging highly intuitive, cross-platform system management tools that the CIO can remove the reliance on the IT specialist and run an effective, efficient data centre with a small team of less skilled individuals.
Consolidation and Virtualisation Trend
There is a considerable shift away from the distributed, server based infrastructure towards centralised, consolidated data as CIOs look to address the top two business concerns ? costs and green issues. Without a doubt, the data centre can support significant reduction in headcount: replacing the diverse skills required to support a distributed infrastructure with a core data centre team can contribute massively to achieving new budgetary targets. Furthermore, multiple maintenance contracts can be replaced with a single, lower cost contract, shaving additional cost off the budget.
Viewed from a high level, the environmental claims of the data centre look compelling. According to one hardware supplier, within 15 months the cost of electricity used by one server will be equivalent to that hardware's original price. Consolidating onto one or two centralised machines should, in theory, result in a dramatic drop in power consumption, supporting CSR policies and driving costs down further.
But in the headlong rush to meet corporate objectives, are CIOs recognising the new demands that will be placed on the 21st century data centre? Will the new mega data centre be able to meet the volume of business in an acceptable time frame? Will the CIO be able to leverage this consolidated infrastructure to extract more volume from existing resources? And, critically, will the cost and power consumption objectives actually be achieved?
While the downsides to virtualisation are few, those that do exist should be carefully considered. The "eggs in one basket" approach will naturally offer lower maintenance costs than if you were to run numerous systems. But if one single server fails it's simple to "plug in" another. If a larger system goes down containing hundreds of virtual servers there has to be a resilient contingency plan in place. Virtualisation is the reason data centres are investing heavily in automated management solutions to reduce human error, provide early warning systems and lower the cost of management.
Are business growth and strategic business requirements in danger of being compromised by short term economic goals? Or will the cost of highly skilled, highly paid IT specialists employed to meet those requirements fundamentally undermine the data centre cost benefits?
Driving down running costs and improving operational efficiencies remains the key objective for the CIO. Yet the mega data centre carries a significantly higher risk strategy, creating a very vulnerable single point of potential failure. Disaster recovery plans must be robust and organisations will require a hot standby ? a strategy that will add cost and could therefore ultimately undermine CSR goals.
With these additional green and financial considerations to address, it is therefore imperative that the CIO can deliver highly effective, highly efficient data centre management. Organisations moved away from data centres over the last ten years to distributed computing for a reason: the need to move fast, develop tailored solutions that responded rapidly to business needs and achieve high availability to support the 24x7 economy.
And whilst the hardware technology has progressed significantly in the last few years, delivering unprecedented reliability at a new, low cost point, the overall environment has become far more complex, adding new systems management challenges that could undermine both the quality of service and cost objectives.
Critically, there has been a dramatic shift towards multi-platform operations to support end to end business processes that integrate the online customer with every element of customer service and the supply chain. Indeed, this shift in approach is now so entrenched that the new IBM systems support different operating systems on one platform.
This new environment, therefore, offers unprecedented challenges in system management, capacity planning and support. It is perhaps, little surprise that according to a recent IBM survey on data centre use, 58 per cent of respondents said the biggest challenge for the CIO is the cost of operation in terms of power and people.
If organisations are to achieve a reduction in headcount whilst still delivering a robust data centre environment, there is a pressing need for effective systems management tools that can support end-to-end business processes across the multi-platform environment. Highly skilled, data centre specialists and platform specific utilities will have no place in the increasingly cost conscious data centre.
The mega data centre of the 21st century should begin to resemble the Marie Celeste. With the right, cross platform systems management tools, organisations can manage the entire data centre from a single console using exception reporting to highlight problems or potential problems as they arise.
With this easy to use, highly visual data centre management approach, organisations can also improve day to day operations: from real-time alerts of escalating temperature or processing power usage to checking available disk space. Critically, by automating these data centre management routines, the CIO can move away from reliance upon highly skilled, highly paid individuals and de-skill the data centre to achieve further cost savings.
Whilst organisations will struggle to achieve the expected cost benefits without robust, cross platform system management, it is important to truly consider the power consumption equation. Take the much vaunted blade technology, for example. The footprint may be impressively small but each blade generates a huge amount of heat. While blade technology enables organisations to add huge capacity to the data centre, each requires around 15 Kw of power for cooling ? as compared with 2Kw per rack of traditional servers.
This is a massive increase in power consumption and will significantly dent the green credentials of any data centre. IBM has attempted to improve the situation by adding power management software to machines that enable unused processors or disk drives to be switched off or powered down. However, the inherent latency adds a small but significant delay when systems are then switched back on, undermining performance.
However, there is a very simple change that organisations can make: reduce the amount of power hungry cooling technology. Today's hardware can operate effectively in greater temperature extremes ? typically between 10C and 35C ? although some HP kit will run at 50C. Yet most organisations are using huge amounts of electricity to keep the data centre at 22 to 24C, in the mistaken belief that systems will shut down if the temperature goes any higher.
Simply reducing the amount of cooling within the data centre and letting the temperature increase by 5 to 10 degrees will result in a significant reduction in power consumption and cost, without compromising performance in any way.
Without a doubt the data centre trend is based upon sound principles and reflects current corporate demands for financial prudence and environmental credibility. But it is not a strategy free from risk; get it wrong and organisations will fail to achieve either cost or green objectives and will risk undermining the quality and reliability of service delivery across the business.
Get it right and the mega data centre will offer economies of scale hitherto unforeseen; it will enable a long overdue deskilling and a move away from over-reliance on ?IT heroes' and it will deliver significant cost savings. Whether it will really reduce power consumption, however, remains to be seen.
Three top tips for effective systems management
1 ? Adopt automation across platforms with good systems management tools
- - This will control business processes in data centres and satisfy customers
that you're delivering on the SLA
2 - Challenge the setting on your air cooling systems.
- Servers are now far less sensitive to heat with much higher operating temperatures. It is vital to know the limits of your hardware but you may not need to cool the computer room to such an extent and could make significant energy savings
3 ? Adopt virtualisation wherever possible
- Consolidate software tools as well as hardware. Using an Enterprise Edition software tool allows you to have multiple instances of products and applications on a single server, saving time and resources
Tags: Applications, Hosting & Colocation, Design & Facilities Management, Networks & Telco , Security, Servers, Storage Networking