Get email updates every time we post!
http://www.quocirca.com/pages/analysis/articles/view/store251/item4496/
Having looked at all the information available around gird, having noted that your existing infrastructure utilisation is running at around 30% or so and with the CFO and CEO of the company breathing down your neck with the ongoing mantra of “Do more, with less”, you’ve decided that grid is for you. You bring all your existing hardware assets together through the application of all the various virtualisation techniques that are available, get hold of some software for managing all these assets, and off you go – you are running on a grid. So far, so good for a simple cluster-based, single application grid – everything is available to the application concerned as a single logical resource, and the application just thinks that it is running on a very fast old-style supercomputer.
Well, hold your horses – what about more complex environments, where we are looking at multiple applications? It may be a grid as far as the hardware goes, but can the software that’s going to be sitting on top of it actually make the best use of it? Surely, providing that we can just provision the software (essentially, build the stack from bottom to top that the solution requires), then the grid will manage everything else?
Unfortunately, this is a bit of a problem for some existing applications. The need for certain applications to be fully stateful, in that they know exactly where they are at any time, can cause problems in a virtualised environment. If the underlying infrastructure is inherently dynamic, an application that is dependent on static resources can only be provisioned to a set of static resources – which negates the main advantages of a grid.
An interesting article, grid is a solution for many applications and can be a real business enabler, just look at the success customers have had with their Platform or DataSynapse grids in the Financial Risk arena. The issues many are having though, tend to be more non-technical, that sure I can deploy a grid infrastructure by tomorrow morning, but you know what, that grid is only going to be as effective as your application, your billing and business processes. That the application code is optimized, that the business and the developers have buy in is what drives your success most importantly – that’s not to say the vendors aren’t doing their part, but understanding that putting a grid solution in is just like deploying a new application or a new infrastructure and therefore requires the same level of buy in, support and key to everything communication. That the reports are not late today due to a grid failure, they are due to an issue with the application, the database or engine7, sure from the user point of view, it’s all IT, from the developer, the support and the management, it’s a different ball game.
No related posts.
Related posts brought to you by Yet Another Related Posts Plugin.