Comments on: Utility computing and the drive towards a more sustainable future http://www.hybrid-cluster.com/blog/2010/03/utility-computing-and-the-drive-towards-a-more-sustainable-future/ Next generation web cluster technology in the cloud Thu, 04 Nov 2010 21:14:44 +0000 hourly 1 http://wordpress.org/?v=3.0.1 By: Kieran Simkin http://www.hybrid-cluster.com/blog/2010/03/utility-computing-and-the-drive-towards-a-more-sustainable-future/comment-page-1/#comment-695 Kieran Simkin Thu, 04 Nov 2010 21:14:44 +0000 http://www.hybrid-cluster.com/?p=541#comment-695 I think this all part of an overall and gradual shift from the main cost in computing being hardware and associated maintenance costs, towards a situation in which these are provided as utilties in the same way we receive water, electricity and gas - the main cost becomes the software which makes use of the available computing resource. I think of this as analogous to the introduction of electricity and electric lighting - at the beginning the real struggle was the provision of electricity to the home as a utility, the cost of the light bulbs in people's homes was neglegible compared to the cost associated with the provision of a national grid. Gradually as the electricity infrastructure became available, the market for appliances that made use of this new utility blossomed and eventually overtook the size of the market for the utility itself. This is what we're expecting to see in the cloud computing sector - as the infrastructure for providing computing as a utility becomes widely utilised and available, the size of the market for software or "appliances" to make use of it will grow, potentially beyond our wildest expectations. In the days of Thomas Edison, who would have predicted the incredible variety of devices we'd be plugging into the national grid today? In terms of what I think it will do to the nuts and bolts of computer hardware - I imagine that the devices we use to access the internet will become thinner and lighter both physically and in hardware capabilities, I suspect they may even plateau at a certain processor speed deemed "fast enough" to render the internet satisfactorily. Most of the computing resource will be moved into the data centre - internet devices will become dumb terminals, effectively just a window on the internet. This move towards consoldating computing resource allows for economies of scale - there may be a significant divergence between server hardware and client hardware - currently clients and servers run hardware that is basically very similar. However, as the requirement of the client changes from being as fast as possible to being as lightweight and power efficient as possible I think it's likely we'll see a gradual fork in architectures. I think this all part of an overall and gradual shift from the main cost in computing being hardware and associated maintenance costs, towards a situation in which these are provided as utilties in the same way we receive water, electricity and gas – the main cost becomes the software which makes use of the available computing resource. I think of this as analogous to the introduction of electricity and electric lighting – at the beginning the real struggle was the provision of electricity to the home as a utility, the cost of the light bulbs in people’s homes was neglegible compared to the cost associated with the provision of a national grid. Gradually as the electricity infrastructure became available, the market for appliances that made use of this new utility blossomed and eventually overtook the size of the market for the utility itself. This is what we’re expecting to see in the cloud computing sector – as the infrastructure for providing computing as a utility becomes widely utilised and available, the size of the market for software or “appliances” to make use of it will grow, potentially beyond our wildest expectations. In the days of Thomas Edison, who would have predicted the incredible variety of devices we’d be plugging into the national grid today?

In terms of what I think it will do to the nuts and bolts of computer hardware – I imagine that the devices we use to access the internet will become thinner and lighter both physically and in hardware capabilities, I suspect they may even plateau at a certain processor speed deemed “fast enough” to render the internet satisfactorily. Most of the computing resource will be moved into the data centre – internet devices will become dumb terminals, effectively just a window on the internet. This move towards consoldating computing resource allows for economies of scale – there may be a significant divergence between server hardware and client hardware – currently clients and servers run hardware that is basically very similar. However, as the requirement of the client changes from being as fast as possible to being as lightweight and power efficient as possible I think it’s likely we’ll see a gradual fork in architectures.

]]> By: Peter http://www.hybrid-cluster.com/blog/2010/03/utility-computing-and-the-drive-towards-a-more-sustainable-future/comment-page-1/#comment-682 Peter Mon, 01 Nov 2010 11:47:54 +0000 http://www.hybrid-cluster.com/?p=541#comment-682 A very well presented posting. I would like to add, that to support this model, much faster, symmetric internet access is required, and inter-server data transfer, at lower cost. This is developing, but universal availability is still a way off. I am interested to know if you have any views as to what will happen to the costs of nuts and bolts - the PC and servers - and how changes will affect how money is made by companies/people. Perhaps a more orientated towards revenue from software. For companies like yourselves providing enabling software tools and services. A very well presented posting. I would like to add, that to support this model, much faster, symmetric internet access is required, and inter-server data transfer, at lower cost. This is developing, but universal availability is still a way off.

I am interested to know if you have any views as to what will happen to the costs of nuts and bolts – the PC and servers – and how changes will affect how money is made by companies/people. Perhaps a more orientated towards revenue from software. For companies like yourselves providing enabling software tools and services.

]]>