IT suppliers are lining up to play at the Data Centre Casino, the biggest IT game in the world, to stake a claim as a supplier of IT services as utilities. There are just two tables with limited seats; complete systems on one, essential components on the other. Are your IT suppliers in the game or not? This is all about rewiring …
Sun/Oracle is not limited on Containers
There is a thing called QLayer laying around in the assets of SUN
Public Clouds - it is a more complex story
Chris - good zip thru at Autobahn speeds through all things cloud. However, I must respectfullly disagree on a few key points.
1)You commented that a small number of very large players will own the public cloud compute/storage market. True at the lowest common denominator level - but I think the largest telcos and hosting companies are going to join Google and Amazon, HP and Sun. There is a science to service creation and delivery, and while Amazon can deliver services for the small web outfits and consumers, there is more to be done before enterprise SLAs are met.
2)You observed that Mimecast, ParaScale and Nirvanix will be relegated to niche clouds. That is a broad brush stroke since Mimecast offers hosted email archival, ParaScale offers enabling software to help service providers build public storage clouds and enterprises to build internal storage clouds, while Nirvanix competes with Amazon S3 cloud storage. Mimecast is clearly specialized from the get go, and Nirvanix obviously has the challenge of competing with the 800 pound gorilla. ParaScale is a different beast - we are arms merchants to create storage clouds - niche or otherwise.
For another perspective on storage clouds - here are a couple of reads:
Sajai, CEO, ParaScale
thank god for a new buzzword
I was getting so bored of saas
This article whilst interesting seems to have missed some fundamental points and ploughed head long in to prescribing what each vendor must do for world domination.
The biggest point completely avoided is, What is Cloud Computing? The reason its glossed over here and also in the recent Cloud Computing Manifesto document is there are two different competing views.
They all agree its about turning data centers into a virtualised utility computing providers. Where you simply pay for the resources you use and can vary the amount consumed just like turning on a tap. Thats the theory anyway.
The problem is there a two different ways of doing this.
On the one hand you have Google and Microsoft who have proprietary technologies which your infrastructure must be built to use, such as the Big Table DB. They operate on top of the cloud infrastructure.
On the other hand you have Amazon, GridScale and everyone else who seem to just be providing VPS servers, which can be configured automatically to scale out. So if your web site got dugg today just fire up 10 more Instances based on the same original image and all will be well. The problem with this is you need to write or purchase the software to actually do the scalling out.
You have also fundamentally forgotten about the software side of it. Sure we need VMs to manage it, but what images are these VMs going to run. If you read the various Google papers on scaling out its a seriously complex problem. Its also not helped by all the various different vendors each with there own vision and matching API. Ideally we really need one standard to control deployment to the cloud.
Another interesting technology you didn't mention is Eucalyptus, its going to be strong part of the Ubuntu Server cloud play and effectively provides a private Amazon EC2 on your own hardware. And the most important thing is its API is compatible with Amazons. Its still in fairly early stages, although they have just received funding to commercialise it in much the same way XEN did.
Of course many companies don't need to worry about scaling out, since they can happily just run one instance to meet all there needs for that particular service. But then thats really just server visualization with a trendier name and not cloud computing.
The really attractive prospect is mixing the public and private clouds, so you can do disaster recovery to Amazon or simply purchase there service when you really need them and supply the very day base load via your private cloud.
Public Cloud - Vendor Lock-in
Companies will be locked in to Microsoft and Google once they build apps for the proprietary clouds. Do you think companies be willing to accept this lock-in for the built in scalability that will be lacking in generic clouds such as Amazon's?
What you really want is blanket fog instead of clouds. Are companies not at the mercy of the ISP's with traffic prioritisation etc? Which I guess would be the air between the ground and the clouds.