The recent announcement by HP of the C-class blade servers got me thinking. Could it possibly be the marker for something more than just another new blade architecture?
The interesting aspect about the new range is the way it represents a closer marrying of what technology can offer and what business users require, namely an ability to define the systems they require in terms map more directly onto the business processes they need to employ.
It is, for example, possible to see C-Class as a move towards matching the available technology to the same level of granularity as the individual business processes users link together.
The motif that seems appropriate is that of the whiteboard, where a business manager can draw out the detailed elements of the process, together with the flow of business logic that connects them. Up till now, however, it has normally only been possible to map between the business process and the technology at very broad-brush, major sub-systems level – for example, a server cluster running a mega-application such as a large database system or an ERP system. It is only within the confines of the mega-app that the detailed elements have been identified.
The practical result of this has usually meant that the act of mapping between the business requirement and the technology provision becomes the sole preserve of consultants and other high-fee earners.
Nothing intrinsically wrong with them, of course, but such a role of intermediation between business and technology – once absolutely necessary as a translation service between one and the other – is now, arguably, simply helping to perpetuate the need for such translation at all. Everyone agrees it would be better if business and technology could speak the same language – some sort of Biztech Esperanto – but till now there has not existed the reality of technology and technology models at a suitable level of granularity in which to implement it.
The technology model that now seems to be appearing is that of the server `appliance’, single function servers that provide services matched and focused to the specific needs of a business function. This is certainly an approach that maps well onto the other development trend for this year – server virtualization.
But with the arrival of the C-Class range it is possible to see the development of hardware that allows a similar level of granularity in the types of server that are available.
Going back to the whiteboard analogy, it then could become possible for IT managements – and even technology vendors – to map their offerings onto the business process boxes drawn on the board by business managers. This should become much more likely as more communications servers, and other support functions such as SAN and NAS servers, get added to the mix. Business process `A’ could be defined, specified and designed-in by simply writing : “run software appliance `B’ on hardware server `C’” over it on the whiteboard.
OK, so none of this is going to happen over-night; it will take a good deal of time for HP to build the ecosystem – as well as the users’ perceptions of it as a `good idea’. It could be argued that this will be a harder task for the company than providing the technology – after all, it still seems to struggle to get beyond its `Sushi = cold, raw fish’ approach to hyping its technology.
That being said, one of the most important components of such an approach is the availability of infrastructure management software to run it all. HP is, of course, one of the big players here, with OpenView, but it will have a fight on its hands against IBM, which has BladeCenter and the Tivoli
So it will be the relationship between finer logical granularity and increasingly powerful management tools that will be the key to creating enterprise infrastructures that are both infinitely scalable and infinitely flexible. If you want another analogy, what business users are seeking is something like the Sahara Desert