Cloud Computing Market

Click on the image to get access to the larger version which is a bit more legible.

]]> http://blog.cloudsplit.com/2010/02/21/cloud-computing-map-of-the-world/feed/ 0 Joe Cloud Computing Market http://blog.cloudsplit.com/2010/02/19/cloud-computing/ http://blog.cloudsplit.com/2010/02/19/cloud-computing/#comments Fri, 19 Feb 2010 14:58:42 +0000 Joe http://blog.cloudsplit.com/the-cloud-computing-definition-we-use/ ]]> When we talk about Cloud Computing in CloudSplit this is what we mean:

· Virtualization: All the components in the system are configured in software and have APIs to allow access. The people component has essential been eliminated. People still exist but a Cloud Computing consumer never sees them. Instead they see an API or a web-page interface to all their services.

· Fine Grained Allocation: In Cloud Computing the units of allocation are small enough to make sense to even one-man businesses. Sub-cent pricing for transactions and zero upfront costs can make Cloud Computing very attractive for high growth start-ups and/or cost constrained IT departments.

· Bi-Directional Scalability: The ability to turn stuff off as fast as you can turn it on without incurring setup fees, extended contracts and/or waiting periods. This scale up/scale down model ideally suits spiky or seasonal workloads (e.g. the storage requirements of a University are significantly reduced out of term).

· All You Can Eat: The ability to consume resources in an unbounded fashion. Amazon will never tell you they are out of disk space. One of the trickiest problems to deal with prior to cloud computing was what to do when you hit your existing resource limitations (particularly disk space).

· Pay as You Use: Only pay for use, not configuration. If you do not turn on the node or store data, then you do not incur costs. Setup fees become a thing of the past.

The ones to keep an eye on are:

• DeltaCloud : An initiative from RedHat. This a a generic RESTFul service that mediates access to the Amazon EC2, RedHat Enterprise Virtaulization Manager (RHEVM) and Rackspace (though its not clear whether slicehost.com is supported. Its targetted at spinning up compute nodes so there doesn’t seem to be a storage, queueing or entity store API yet.
• LibCloud was originally developed by the guys at CloudKick but it has recently been promoted to an Apache incubator project. Its a Python library which is again focussed on compute server management. It supports a lot more providers including Amazon Ec2, Slicehost, Rackspace, Linode, VPS.net, RimuHosting, GoGrid, Terremark and VCloud (phew!).
• SimpleCloud is an mega corp initiative by the big boys (Microsoft, IBM, Zend) and is focussed on enabling cloud APIs in the PHP space. The website talks about blob storage queuing and structured storage (what SimpleCloud call document storage, which makes no sense to me). However the download they supply only supports the storage API at the moment. Surprisingly they have no support for managing compute instances. The APIs supported for blob storage right now are Microsoft Azure, Amazon, Nirvanix and local storage.

Let me paraphrase that and say “Private Cloud and Public Cloud”. As organisations move their internal IT operations into virtualised Environments and begin to see the benefits this will only serve to accelerate the adoption of public cloud vendor offerings from Amazon, RackSpace, Joyent, GoGrid and others.

We already know from a a Forrester survey last year that Enterprises are the surprise adopters of external clouds and we can expect to see increasing adoption of the tier 1 cloud providers (Amazon and Microsoft Azure) as these two companies are the only vendors with the global reach that will allow trans-national organisations (think Pepsi or Cola-Cola) to place their data within the appropriate legal boundaries for their various legal  jurisidictions they reside in.

Amazon lowered the bar for European cloud adoption last year with the normalisation of prices between Europe and the US and we an expect to see this trend continue as other vendors make the move into Europe.

Pat Phelan originally intended to join and we booked a Monster House on Fillmore and Fulton. Pat (the consummate deal maker) had to pull out at the last moment which left Eamon and I (of CloudSplit) sharing this mansion with Paul, Moneesh and Sean of VidSchool.

As failed entrants to the onstage event we had both been offered slots in the DemoPit. You get one day to showcase your company in a separate area from the stage through which the attendees have to pass to get to the auditorium. We chose Tuesday (as opposed to Monday) this gave us a chance  to attend on Monday and suss the place our prior to our full day.

Tuesday duly arrived and we headed down to the show. Its a brutal schedule with the show opening at 7.00am and running until 7.00pm the following evening. There is lots of advice to absorb on how best to pitch at TC50, but we followed some simple rules,

• Bring a pull up stand and put your message and pitch all in the top third. Nobody can see below this point. Many of the TC50 companies only used the default table logo provided by TC50. I found myself ignoring companies when I could not discern what the offering was.
• We brought along some tic-tacs that happened to be in the company colours. I don’t thing they got people to come up to our stand but being able to give something to people who listened did leave the whole presenting transaction with a nice soft end.
• You need two people. Otherwise food breaks/toilet visits leave the stand unmanned. There is always action in the DemoPit area so someone needs to be on stand at all time.
• Be prepared to get you pitch away in a few minutes

The Good

We went looking for validation of the CloudSplit offering and received that whole heartedly. We met key influencers at the investor, partner and customer level. This level of exposure to people WHO-REALLY-KNOW the software sector was invaluable.

Just the opportunity to present 200 times or more to genuinely insightful individuals who could really grasp what we were doing was a fantastic education. We now have a crystal clear vision of what we need to do in V1.0 and a goto market strategy honed by hours of feedback.

It also helped that the universal feedback was that CloudSplit was genuinely breaking new ground in a valuable and emerging market.

I would definitely target and time the launch of any new company so that it aligned with TC50.

The Bad

The DemoPit works as a competition in which the conference attendees get chips which they donate to the most interesting projects. The two with the most chips on each day get to present the product on stage. Its a nice idea but is open to all sorts of gaming ranging from booth hotties simply trading on their looks and accosting people for chips without pitching to wholesale buying of chips. Basically you can forget getting on stage on merit alone.

I can’t fix the booth hottie problem but it should be easy enough to fix the chip buying problem by making the containers piggy banks rather than jars so that once chips are donated they can’t be retrieved to be resold.

I also think there would be more liquidity in the market if the attending companies were compelled to donate their chips to other companies rather then bunging them into their own jar. This could be achieved by only giving chips to DemoPit companies on the day they are *not* presenting.

On plus side we got our first chip quite early on from Mark Kvamme of Sequoia so we really did care too much about winning or losing the competition after that piece of validation.

The Ugly

The awards ceremony was a shambolic disgrace. Mike Arrington threw all his toys out of the pram and stormed of stage. Why ? Who cares. It was an insult to the winners of the awards and made Arrington the story instead of the winners. It soured the whole event for me.

If I was Mike I’d be keeping a pretty low profile as well.

Exhausting but very rewarding work. Both companies recieved huge validation for their offerings and like most SV events I have attended our calendars are full for the next few days.

The straight juice, there is no better place to launch a tech company.