Cloud Computing Needs to get on the Bus
Posted on July 14, 2008
Filed Under Uncategorized |
Irrespective of any other event or factor, I think cloud infrastructures are the quiet revolution in the tech industry. Well not so quiet but my point is that it’s hard to overstate the dramatic change they have brought about for how entrepreneurs build new products and companies, as well as how established IT providers are taking these infrastructures into account as they look to the future.
Engine Yard, a hosting and support company for the popular Ruby on Rails programming framework, has raised $15 million in a second round of financing. Chief Technology Officer and co-founder Tom Mornini says the San Francisco startup will soon expand its offerings with the release of Vertebra, its platform for managing Rails applications in the Internet cloud.
[From Engine Yard gets $15M for Rails hosting, and a new platform » VentureBeat]
One aspect of this that doesn’t quite sit with me well is that the tech stack will ultimately homogenize and commoditize, meaning there will be well defined “camps” that fall within known parameters and depending on what your technology needs and predilections are you will select a cloud that fits your parameters. Assuming reliability achieves some telecom like constant, then the industry moves to a mainly price driven competitive model.
Right now there are the following “as a service” permutations:
- Infrastructure: IaaS is the lowest level set of services and the primary abstraction is provisioning and racking, everything else you get to manage, including the virtual machines. This is what Engine Yard does but the most well known example is AWS; other IaaS providers include 3Tera, Elestra, Xcalibre and Nirvanix. VMware and Xen are also IaaS but you have to build your own IaaS using their virtual machine technology, which means it’s more work but still entirely doable.
- Platform: PaaS is middle of the road, meaning not IaaS but also not SaaS. If you already have a working app that you need to scale quickly, this is likely the best scenario for you. Google Apps Engine, Mosso, Morph and Heroku are known PaaS providers who can take your Python, PHP, Rails, etc. bits and scale ‘em up PDQ. The big tradeoff here is flexibility because you are restricted to what extensions and libraries the PaaS provider is making available, meaning your existing app could require some porting.
- Software: SaaS is the granddaddy of cloud computing and most restrictive. Salesforce and Google Apps are the best examples of SaaS, each provides 3rd party developers with the infrastructure and the app code hooks to work with but in each case you are building to their app rather than a generic set of services. I’ve always thought of this as more of a route to market problem solver than a technology solution.
So why do these aaS variations need to “get on the bus”? Simply put I think that achieving interoperability is a more strategic longer term goal than simply commoditizing hosting costs. OpSource may not be the best example but it’s one that comes to mind, their app hosting service provides the technical infrastructure as well as add-ons like billing, end user support, and analytics (which is kind of a general term…) that provide business solutions to software providers in addition to technical advantages.
OpSource is taking advantage of Mule ESB and that means that one provider on OpSource should be able to take advantage of other hosted app even if it is from a different provider. What this means, as an example, is that my CRM app could embed functionality from another company that is providing some specialty integrated online marketing module. I’m not limited to what I build and host.
It turns out that ESB functionality is making it’s way into cloud services, AWS SQS being one example that also takes advantage of Mule’s most awesome ESB. Going forward, and as clouds more intensely target enterprise accounts, I am sure we will see more ESB like functions make their way to the cloud. It’s all goodness for entrepreneurs and startups, and ultimately the cloud will change the way that enterprises buy, build and manage technology services.
Comments
-
About
My name is Jeff Nolan and I write Venture Chronicles. These are my thoughts on technology, innovation, management, public policy and pretty much anything else that I find interesting.
-
-
Links
Get My Feed!
