A recent post by the Cloud vendor CohesiveFT talks about the potential changes in technical sales cycles when evaluating Grid based products. I’m not sure I agree totally with the article, but the ethos behind the article i.e. making it easier to trial products, try out solutions and build apps /services quicker to be build internal business cases is solid.
Cloud is a game changer, which is the intent of the article, but you cannot apply a broad brush to “Grid on the Cloud” as a unilateral game changer in respect of Cloud replacing Grid (which to be fair is not the intent of the article). For many companies replacing internal Grids, or even planning for new Grids, cannot be done on the Cloud. There are challenge of integration, moving data, securing data (and this is where Cohesive FT’s VPN-Cubed product offering can help), physical location, legislation, SLA’s and availability (see this article for a good synopsis on this as applied to EC2). Many of these will be resolved in time, and some of course can be resolved right now, but with the move by many vendors to enable existing IT infrastructure and Data Centers as private clouds the point is likely to be mute in the future I think. Right now, an internal Grid is not elastic. It does not add more servers or resources to the service as required, but this will change as such internal Fabric enablers become more normal.In fact one can image a future where such companies may sell excess capacity of their “Grid Clouds” to ensure a more economical running of their infrastructures.