If there is any piece of history in the IT business that seems to not only continue to hold true but has both a high and long-standing impact on the development of IT and the services it renders to business, it is to be found in the company that has the skills, the nous, or just the luck to create a technology that everyone wants – even if they don’t know they want it till they’ve got it. And that `everyone’ includes not only end users but also other vendors that can exploit it.
Two obvious examples are the IBM mainframe, which has used a great deal of commonality in its architecture and technology from the 360 series machines of the 1960s through to the current z-Series machines, and Microsoft’s MS-DOS, still to be found deep in the soul of the latest iteration of the wide range of Windows systems running on everything from commodity servers in datacentres through to mobile phones.
The PC is perhaps the best example of all, where MS-DOS spawned a wealth of applications that got developed quickly because it existed as the bedrock on which the ISVs could build. A good half of the development problems were solved simply by writing the unique parts of the application to run on the standard part – MS-DOS.
That same scenario now exists in the cloud. Take Data Centre Infrastructure Management (DCIM), for example. Or more specifically take the next iteration of that, which for the purposes of discussion at least I’ll call Service Platform and Delivery Management (SPDM). The arrival of a reliable and reasonably comprehensive SPDM could create an environment where everyone, from applications developers through to service orchestrators and Cloud architects, would have a common management platform to write code for, build services with, and specify compliant new applications and tools with confidence.
Services are built up from applications, even sub-systems of applications, that are brought together to run a business process or function of some kind. They can be re-used in different combinations to provide different services. In that way services are unlike applications – which are by their nature fixed until the next upgrade or patch is issued. These can and do change, sometimes dynamically in response to other monitored events.
So the key here will be the development of platform management systems that are analogous to the glue used on Post-It Notes – it never quite sets. Individual notes can be moved when required and the glue will release them. But when in position they stay stuck. That is how users will want the Cloud – applications can be brought together to create services that can be broken down when no longer required – but don’t break otherwise - and where the applications can be re-used to create other services, reliably and effectively.
One of the players in this area is nlyte, which already has a track record in CDIM systems. CEO, Jon Temple, currently sees the company’s role in areas such as ensuring the Cloud lifecycle management piece is fulfilled properly, how users are planning for the capacity required and that sufficient capacity is available. He does acknowledge, however, that this put the company in the mix as a precursor to the development of an SPDM environment.:
To achieve this, whatever is running as part of a service has to communicate in real time to nlyte information about what resources it needs, which is a start point for a building a real time, dynamic re-planning environment. This is needed to cope with the fact that in the real world, no plan is `fixed’. That means more and more data about the state of companies’ resources and the tasks they are running will end up in the Cloud as feedstock for the Cloud management systems:
There is growing evidence that the widespread availability of management information is not only needed but being seen to be of value. My recent interview with Peter Coffee, Vice President and Head of Platform Research at Salesforce.com, showed how performance and service reliability data can be used to build detailed service risk assessments that can be in turn be used to build service insurance policies as a more business-relevant alternative to Service Level Agreements. Temple argues:
The issue therefore is timing. He sees it being a little way off. For my part I see it starting to happen this year, if only on the basis that everything is driven by vested interests, and users, especially the SMB sector, could make use of the capabilities companies like nlyte offer. More to the point, their service providers could certainly make good use of – and be an excellent sales channel for – DCIM and SPDM management tools, especially delivered as a pay-per-use SaaS service.
The need is going to be driven by the on-going economic situation. There will continue to be the fact that some 20% of on-premise infrastructure will come up for replacement this year. Some of it will be capable of soldiering on, but the capital investment costs saved are likely to be used up in ever-growing running costs. Many companies will have already made this move once, last year, so this year could see 30%-40% of enterprise infrastructures coming up for replacement.
There is indirect evidence for this in that Temple admits that sales growth this last year has been `better than expected’ – arguably a sign that on-premise based enterprises have been seeking ways to better manage their existing infrastructure.
Another lever here is going to revolve around people. There is growing talk about skills shortages in IT which, if true, could put enormous pressure on all on-premise sites. Shared services, and perhaps more importantly shared management resources will become the only sensible alternative.
It is also likely that the service provider community will want to take up DCIM tools not only to manage their own resources but to allow individual customers to manage their own environments. Temple acknowledged that service providers are already talking seriously with nlyte.
He also acknowledges that the time is coming for nlyte to start building an ecosystem of partners that can contribute useful additional services to the core DCIM offering. The company already partners with systems software companies, but the need to look elsewhere is close.
Post new Comment