Google’s open source database breaks free from expensive proprietary solutions: Andi Gutmans

You recently launched a new product, AlloyDB on PostgreSQL, an open source database system. Could you talk about the motivation for creatingto create the product and the need it is seeking to solve?

One of the key things that customers have asked us over a couple of years is a out of relying on expensive legacy proprietary databases. Therefore as customers want to move into the cloud, they also want to break free from some of the those unfriendly legacy pricing contracting schemes. The biggest challenge there is that customers want to move to the cloud, not just for the cost and agility benefits or even for the flexibility of operating wherever they can- their demand to us is to help them get to an open API where we still have the benefits of the proprietary databases. Customers need availability, durability, manageability and scalability, that they are used to from their proprietary databases. So we (Google) wanted to embrace an open API, and PostgreSQL has become the defacto open API for enterprise applications or databases. Our product on PostgreSQL, “AlloyDB” fully embraces the open source while meeting the benefits of a proprietary database.

While AlloyDB has not been made publicly available yet, many people tout it to have the key differentiating factors to take on Amazon and Microsoft who lead Google in cloud. What expectations do you have from the product to tilt the competitive landscape in your favour?

Our focus is on what the customers are asking for and not the competitive landscape because it is a large market segment. Customers choose different cloud services for different reasons and 80 per cent of the customers are multi-cloud. We came into this from a customer focus perspective, however, we wanted to make sure that what we were building is very differentiated in the market from day one. From that perspective we basically used a lot of capabilities we have built internally, to be able to deliver very differentiated systems. So for example, our infrastructure is unique where our computing and storage infrastructure is completely disaggregated, which gives us a lot of flexibility on how we build systems in the most available and performative way. We own our network backbone. So the way we think of engineering our systems is probably different from other cloud service providers which gives us a real advantage. The end result is that we are two times faster than Amazon, from a commercial perspective, we are far easier to the consumer than other solutions. Our pricing dimensions are simpler compared to other players, who are more opaque when it comes to pricing.

What has been the feedback from clients on AlloyDB so far?

We have been overwhelmed with the feedback from the market since we announced this product. This tells us that we hit on some key pain points that the customers were feeling. We knew that a differentiated Postgres will be interesting, and customers who grew open source at some point would want something which is more robust and scalable so that did not surprise us. We also knew that customers wanted to move away from legacy proprietary databases, so getting engagement there did not surprise us as well. But two places where we got more engagement than we expected was that we did not expect the customers’ desires to get off proprietary databases to be as overwhelming as it was. So more and more customers coming out of the woodworks with this desire was very interesting. Secondly, we were also getting interested in the cloud outside of the Google Cloud Platform realm, which was interesting.

So AlloyDB aims to create a differentiation between legacy closed systems and open systems, however since your solution will work on your own closed system infrastructure- do you think in the long run it can be opened like Kubernetes and Borg?

So even if we wanted to, it is not really easy to go open source on our solution. Since it is not just a piece of source code or software, but rather the infrastructure it is operated on- the hardware, the data centers and so on. Infrastructure that is very purpose-built over ten-fifteen years and serves the rest of Google. So we take a very different kind of approach. We make sure that our solution is 100 per cent compatible with Postgres. But if the customers want to run it elsewhere, they will be able to run it wherever they want. They might get different performance characteristics but we really give them the freedom and flexibility that they are looking for.

Can you take us through your thought process on how you are innovating future products in this domain?

We can’t talk about future products, but we have a very robust portfolio. I talked about migration, modernisation and transformation, and AlloyDB really fills a very critical role in our portfolio. We really feel good about where the portfolio is today. Looking at the feedback from analysts, and customers, we hear that Google has moved to the right place really quickly. A big part of that is that we have focused on key enterprise features. While I cannot talk much about products, I would definitely say is as you think about a Data Cloud and making it easy for customers to innovate, you know our goal is to really integrate that data cloud. So that customers do not have to do that integration themselves, and they can focus on driving innovation and generating business value.