Although this is a logical continuation of IBM s consolidation effort, one view of the catalyst for this particular project is the projection that IBM will need to double its computing power by 2010, but is challenged in terms of its floor space and energy consumption. Rob Hailstone, Software Infrastructure Practice Director with Butler Group, Europe s leading IT research and advisory organization, says it is encouraging to see a firm commit itself to a radical change in the way IT is delivered.
IBM has already made great advances in the consolidation of its IT infrastructure, many of these as a result of its decision to run its IT as a commercial service to the rest of the business. Over the last 10 years it has, for example, reduced its number of data centres from 155 to 7, and the number of corporate applications from 15,000 to 4,700. This new phase of consolidation represents a very significant investment for IBM, but as well as providing a significant, long-term cost saving will enable IBM to build a team of very credible resources to pass on the experiences learned to its customer base. IBM already has a team that provides consolidation support to its customers, and elements of that team will become part of Project Big Green. From the customer perspective, there is a big difference between ‘Do as we say’ and ‘Do as we do’, where the supplier is able to demonstrate its own commitment to consolidation and its own associated cost savings.
Altogether IBM aims to reallocate US$1 billion per year to accelerate its development of green technologies, creating a 1,000-strong ‘green team’ of specialists, and building a client roadmap based on its experiences. Project Big Green represents an important part of this learning curve as well as delivering considerable in-house benefits. The green credentials of the project come from the anticipated 80% energy reduction and 85% floor space savings through the planned consolidation. These contribute to the overall 40% cost savings for the consolidated applications, working out at an estimated US$250 million savings for IBM over five years. The other substantial savings come from labour and software costs. These estimated savings allow for the cost of new hardware, and the migration effort.
IBM already runs around 1,500 of its 4,700 applications in System z, and its use of Linux on System z has increased by 300% over the last year. IBM s own figures show that 33% of its applications run on 1% of its number of physical servers.
While the use of System z to consolidate multiple Linux servers has been available for some years, several factors are now coming together to suggest that it is a good time to embark on a more wide-scale deployment. These factors include confidence in the stability of the environment, greater awareness of the potential savings, and the business imperative to reduce the energy and carbon footprint of applications.
For reasons of disaster recovery, organisations will still wish to have multiple physical sites, and so the use of Series z as a consolidation platform will inevitably appeal only to the larger enterprises. Not all workloads will necessarily be obvious candidates for this type of server virtualisation. IBM has acknowledged that the applications that have most to gain are those that are transactional in nature, and where the periodic variation in workload can benefit the most from being able to dynamically reconfigure the resources allocated. Organisations that do not fit this profile might have more to gain from the use of load-balancing clusters or grids, with the use of high-density blade servers, and hardware designed from the ground-up to have low power requirements. There might also be scope to reconsider the Series z as an appropriate platform for on-demand computing where the raw computing power is delivered across the wire.
Overall, we find it encouraging to see vendors that are prepared to put their money where their mouth is and commit the company to a radical change in the way IT is delivered.