How hard is it to replicate Vault Enterprise features in Vault open source?

Transcript

The Vault is a very complex product in its implementation. It is very simple to use but that's because a lot of the complexity associated with key management as well as the security of Vault is removed from the user, and instead placed in how we have effectively built Vault. And when I say built I mean at a very deep architectural level.

So certain types of features like replication are inherently complex. They're complex in such a way that if one did not have a deep level of understanding and time, one would not be able to implement those features without a certain amount of expertise as well as just a certain amount of time spent on proofing those features from a security point of view.

The Vault team is very dedicated and has a deep background within Vault and secure systems. But we are by no means rocket scientists. So this is not simply a function that we are smarter than anyone else. It's really a function of—we just spent a lot of time building out this feature that allows you to basically ensure that no one is going to hack it and break into Vault using a vector derived from replication. That's very important. It took us as a company around two and a half years to build replication, and that includes work with the founders of the company who actually built Vault in the first place. So what you're paying for is not having to spend two and a half years building replication—plus.

The second feature that people run into is really just complexity of the subject matter. What I mean by that is that Vault is pushing the boundaries of not just the technology but the science associated with how we protect data.

Features like replication, for example, are literally at the cutting edge of the whole field of multi-party computing. So when we build features like replication or we build tools where Vault has to operate now in this global environment in a very secure way, we do so in a way that requires us to have a very deep, intimate understanding of topics like distributing computing, topics like secure multi-party computing, cryptography, in some cases formal languages and formal methods. These are areas that are, frankly, very difficult to understand. Not just as a function of the complexity of the subject matter but also because they are generally really rare to have a skill set in.

So we are very blessed to have engineers at HashiCorp who have a deep background in some of these really interesting theoretical areas of computer science and applied mathematics. But their application is necessary to really understand and overcome challenges associated with replication, with transit, with many of the major features of Vault, especially the features within Vault Enterprise.