Research results obtained by Freeform Dynamics towards the end of 2009 (http://www.freeformdynamics.com/fullarticle.asp?aid=902) shows that, despite the considerable efforts of vendors to promote the benefits of ‘consolidation', organisations of all sizes continue have large numbers of servers in their estates. Organisations are undertaking server consolidation projects, especially employing virtualisation technologies, but there is still a long way to go before such projects encompass the majority of systems. With so much attention being placed on infrastructure consolidation, essentially to save money rather than yet to enhance service flexibility, should CIOs think carefully of the implications to their procurement policies going forward?
It is clear that infrastructure refresh is an area that continues to be a focus of attention. Even if only four per cent of respondents state that their x86 server estates are out of date and in need of investment, the fact that only slightly more than one in three say they have very modern platforms indicates that server estates will continue to evolve. Indeed, the figure below lists the many challenges that organisations face with keeping their server infrastructures running. These challenges are acting as secondary factors influencing approaches to server estate refresh along with the usual suspect of reducing expenditures, both capital and operational.
Click on the image to view in full screen
Similar developments are also to be expected in the storage arena as organisations attention moves on from the low hanging server consolidation projects to address the other high cost area of infrastructure budgets. So with infrastructure refresh likely to be a focus area in the coming years, what effect will the acquisition and merger fashion amongst suppliers have?
Past research has indicated that many organisations have taken some time and, not inconsiderable, effort to streamline the number of suppliers with whom they do business. Clearly there are potential cost benefits to be had by buying greater volumes from fewer vendors. There can also be other tangible, but sometimes more difficult to quantify, benefits through having closer links to suppliers in terms of better insight into coming developments and higher levels of support. On the operational side there can also be significant gains to be had, especially by standardising on fewer platforms thereby reducing the need for broader skills sets and thus potentially reducing operational risk.
But there may also be a downside. The fear of being ‘locked-in' to a supplier and any ‘apparent' dependency making it difficult to change ship or resulting in having to pay higher prices is one that often raises its shadowy head. An internal discussion amongst the analysts at Freeform Dynamics addressed some of these issues, especially in light of recent vendor acquisitions and partnership announcements, most notably the acquisition by Oracle of Sun Microsystems and the news that Microsoft and HP will develop ‘optimised' Exchange appliances. Are these good developments that will help organisations streamline service delivery from the application stack all the way down to the server and storage tin or are they backward steps to the past?
One view expressed is that after 20 years the IT Infrastructure was finally getting to the point where everything was beginning to open up, namely that applications and platforms were being designed for interoperability and work reasonably well anyway without vendors and customers needing to do much to improve service and performance levels except in extreme scenarios. This is has been the open, or at least widely accepted, standards approach that has been adopted by the x86 and UNIX communities.
So the question becomes will the ‘new' approach of designing for optimal running in a nearly proprietary configuration deliver benefit or hinder the flexibility that has been established? Clearly it is easier for a vendor to develop pre-configured systems from the storage to the application. There is an argument that customers may benefit from an immediate cost perspective, but the longer term consequences of reduced choice and even vendor lock-in may outweigh this. Is there a worst case scenario whereby if you want entire systems pre-integrated and optimised with everything locked together that the overall cost could be higher, never mind long term contraction of choice at all layers of the infrastructure and management stack?
It will be interesting to see how vendors address the ‘industry standard' / ‘pre-built and optimised' debate. It will be difficult for a vendor in one breath to trumpet the importance of adherence to industry standards and then talk about engineering things to work together optimally, and therefore presumably making things less effective if you wish to employ alternative components anyway in the stack. In effect, the latter might only be achieved by taking advantage of proprietary, or at least undisclosed, information. Such an approach to integration of complete stacks may also attract the attention of the many regulatory bodies active in the markets as has happened in the not too distant past.
There are arguments on both sides and, as ever when industry consolidation occurs, new entrants will seek to exploit developing niche areas and for ‘best of breed' suppliers to demonstrate their merits. In the middle, as ever, will be the CIO and IT manager trying to work out the best approach for their organisations IT operations and future procurement policies. Things will not be getting any simpler anytime soon.