So Steve and I had a briefing call with Kaavo yesterday who have some interesting technology. And it set me thinking about whether there is a market for pure play application services management in the cloud.
Kaavo automates the job of application configuration and management in the cloud. The product – imod, is rules and workflow-based and manages the life-cycle of application provisioning, including deploying and configuring the software components or services required to create the environment in which applications execute.
I hope I’m not dumbing it down too much to say that I think of it as a data centre automation tool that understands how to manage virtual IaaS instead of physical infrastructure. Kaavo’s CEO and founder Jamal Mazhar would I’m sure also point out that Kaavo takes a top-down, application-centric approach when compared to other solutions in the space. The IaaS deployment environments that they currently support include Amazon, Rackspace and GoGrid amongst others with support for the Eucalytpus project coming soon.
The product naturally fits into at least two of the categories of the market landscape / taxonomy / market segmentation model that I’ve developed. They certainly appear in
- Infrastructure Services/Services
But I could also make a case for them in
- Cloud Software / Cloud Management
- Cloud Software / Cloud Management / Application Services Management
But whilst the business model of Kaavo remains service-based (they charge per CPU hour of managed application) then that pretty much excludes them from the last two software-based categories.
As I’ve been looking at the application services management category in some detail, one pattern that I’ve seem amongst vendors such as DataSynapse (TIBCO), Appistry, and 3Tera is that whilst they offer the management services to automate the deployment of applications, they appear to major on deploying those applications and application components to their own infrastructure as opposed to infrastructure provided as a service by a third party.
A number of these vendors have come to Cloud Computing via Grid Computing and as such it makes senses that the virtual infrastructure that they deploy to is their own grid. They would rightly point out that owning the management and the infrastructure leads to many benefits such as tighter control, better monitoring and better support for the scaling the infrastructure up and down to match demand. In fact some of these vendors do appear to provide the option for deploying to third-party infrastructure services such as Amazon’s EC2, so it suggests that this sort of hybrid infrastructure may be being endorsed.
But I guess I’m left wondering two things.
Firstly is there really a separate market for pure-play application services management where the infrastructure is always provided by a third party? Don’t get me wrong I can see the need and I can see the benefit but it looks a little too much like the existing discipline of application services management already present in today’s data centre automation tools. So if these existing tools add the capability to deploy to, monitor and manage virtual infrastructure as a service then they will be well placed to get the business. But then again perhaps adding this capability is not a trivial matter. Hmm. Not sure.
Secondly, assuming that there is a separate market – what is the route to market for this sort of pure-play, services-based ( as opposed to licensed software ) offer? Could it be taken to enterprises directly? Yes, but it would require significant resources. To me, it looks like a more natural proposition for aaS providers to help them manage the massive number of deployed applications that they will be looking after if the predictions for the impact of cloud are accurate.
Either way Kaavo has an interesting approach that I’m sure either Steve or I will revisit as they develop.