Sramana Mitra: At the same time, that raised quite a bit of funding in that area. Is that correct?
Sasha Gilenson: Yes. Recently Puppet [Labs], an infrastructure as a code company, which automates infrastructure measurement popped up. They just raised $50 million from VMware. There are many other companies in this space. With the increasing pace of change and the growing complexity of the environments, IT has no other way to cope with it than to automate. This is why this space is hot. The automation in this space is still second generation. Puppet is still involved in a lot of expertise and work from IT itself, so there is a long way for improvement.
SM: What do you think is going to happen? Is Puppet, with all the funding and expertise they brought together under their umbrella, going to be able to address the problem, or is there something fundamentally missing in their architecture that they will not be able to get to the solution?
SG: I think they do significant things. First of all they create an [orchestration], and they are managing infrastructure that creates a new level of abstraction. This is an important step. But will they solve the problem? I don’t think so. The environment is more complex, and many things happen according to dynamic processes. The systems needs to be able to react to that. Today’s automation tools are based on the policy of model management philosophy, meaning that if I define a model of how my environment should look and I enforce this environment in this model, everything will work fine. But the reality of the enterprise today is that this model is dynamic. Essentially, you don’t have a standard model that you can standardize. A lot of changes happen that bypass this model. If your system is not intelligent enough to react to all these events, then this model breaks very fast. This is a lack in the philosophy of automation. It is intelligent based on the analysis of what actually happened to the environment.
SM: These companies that are playing in the space – in the IT automation space – are assuming that the environment is static, and they use policies to try to manage automation, whereas your observation is that the environment is dynamic and there are lots of changes happening. [Current solutions are not] taking those changes into account, hence the automation is incomplete. Is that an accurate observation?
SG: I wouldn’t say that they see the environment as static, but they see the environment as more rigid in terms of being compliant with a specific model.