Any conference prefixing its name with the words ‘Future of …’ inevitably features presentations and discussions about trends and strategies that may or may not come to pass. Sometimes, these potential trends are spoken of in a way that makes it difficult to distinguish fiction from reality. Fortunately, at The Future of the Data Centre 2006, the event was rounded off by a timely reality check from two of the world’s leading experts on advanced data centre technology development.
In their joint presentation, ‘Plus ça change: the users’ verdict’, the 451 Group’s founder and principal analyst, Will Fellows, and research director, Steve Wallage, provided an up-to-date view of what is really going on in the leading data centres of North America and Europe. Based on the 451’s Grid Adoption Research Service (GARS), Fellows and Wallage were able to draw on interviews with more than 200 commercial early adopters of grid and related data centre technologies.
Steve is a director of research at The 451 Group, and is a leading member of the 451 Grid Adoption Research Service. Before joining The 451 Group, he was a principal analyst at Gartner Group, tracking the voice, data and IP services markets.
Will is a founder and principal analyst at research company The 451 Group. He is an expert in identifying emerging technology trends, including virtualisation, grid computing, mobile enterprise and next-generation distributed computing.
From this large number of interviews it is clear that new approaches to distributed computing architecture are already making their impact felt among commercial IT shops in a variety of business sectors, and in a variety of different ways. The “grid” approach – where companies seek to create a shared and seamless fabric of computing resources – is one of these approaches, and may yet turn out to be the final destination for many organisations. However, it is by no means the only approach, nor necessarily the most significant.
At 23% of companies surveyed by 451, virtualisation, rather than grid, is considered to be a better description of their current architecture strategy. Others (19%) prefer to call their approach a utility model; 23% speak in terms of high-performance computing; 15% refer to the adoption of service-oriented architecture; and 19% see themselves as focusing on clustering.
Whatever the preferred approach or nomenclature, the sample companies have much in common: a desire to aggregate and share heterogeneous resources across multiple logical and physical domains, and to create self-managing, autonomic platforms that deliver – in the long-term – a better quality of service.
More immediately, 451’s sample companies expect to realise a variety of benefits from the various architecture investments. Many, 77% and 57% respectively, expect their efforts to generally improve systems performance and save them money. Almost half (49%) also believe their efforts will enable them to do “other things”, with 41% also hoping to improve their competitiveness and another 25% specifically aiming to achieve faster time to market.
Not unnaturally, all these goals are being pursued with varying degrees of success, and like most early technology adopters, all of the 451 sample are running into obstacles. The most common of these, cited by 46% of organisations, are connected with software licensing issues – particularly the reluctance that some vendors show towards abandoning old license models in favour of those that better reflect today’s technological realities. But if licensing is the “big issue” it is far from being the only one: bandwidth constraints, cultural intransigence, a lack of standards, security, skills shortages and the paucity of grid-enabled applications or the tools to create them are just a few of the others bedevilling today’s data centre pioneers.
Yet for all these problems, Fellows and Wallage believe today’s grid and virtualised platform adopters are confident that their efforts will be rewarded. As one of them commented to the 451, the ultimate benefit of the modern data centre architecture should not be measured in terms of short-term operational or other tactical gains. “Using grids,” they said “will not save us money, but it will enable us to make more money.”