Why We Invested: pgEdge

Fully Distributed Postgres for Low Latency and Ultra High Availability

Rally Ventures
6 min readNov 19, 2024

--

PostgreSQL, the world’s most popular open-source relational database, has been a cornerstone of modern applications for decades. pgEdge takes Postgres a giant step further by offering a fully distributed implementation that empowers businesses to build scalable applications with multiple, geographically dispersed database copies.

Below is a short Q&A with pgEdge Co-Founder and CEO Phillip Merrick and Rally Venture Partner and pgEdge Board Member Ben Fried, where they talk about what makes Postgres so powerful, why open source technologies provide a competitive advantage and how they see the demand for distributed Postgres evolving over the next few years. Welcome to the portfolio, pgEdge!

1. What is pgEdge and what core problem does it solve?

Phillip Merrick: pgEdge is a fully distributed implementation of the popular open-source Postgres database. Our software sits between the operating system and applications, enabling customers to build distributed applications with multiple, geographically dispersed database copies. This approach addresses two key challenges for businesses building modern, scalable applications.

Today’s applications need to be always on, always available and be very responsive, even when users are geographically distant. Users expect no downtime and quick response times. For application builders, a distributed database helps meet these demands. With multiple redundant copies of the database kept in sync across different locations, if one node goes down — like when a cloud provider has an outage — traffic can automatically route to other nodes without affecting users.

Furthermore, having database copies closer to users, wherever they are globally, significantly improves response times, enhancing the overall user experience. For example, a global application with users in Europe, Asia and Australia can ensure a much more instantaneous response from the application by having database copies in each of these regions.

Ben Fried: Postgres is the world’s most popular relational database, widely used in businesses of all sizes. No other solution offers the level of reliability that pgEdge provides for the things Phillip describes. While some have tried DIY solutions, pgEdge offers a world-class, production-ready implementation that rivals features found only in costly closed-source databases. Having served as a CIO and technology leader in multiple industries, I see these features as essential for delivering reliable, world-class applications, and they will drive even greater demand for Postgres.

2. Postgres is an open source technology. Why do you like open source and how is it a competitive advantage?

Phillip: Postgres is popular because it’s open source and not owned by anyone. We’ve been very careful to maintain that openness, allowing customers full access to our source code with minimal restrictions. This gives companies flexibility to use and modify it within their organizations, avoiding vendor lock-in.

With proprietary software, it’s harder to switch vendors — the dreaded “vendor lock-in”. But with open source, even from a specific vendor, companies can choose to support it themselves or find alternative support. This freedom is key in the open source community. On the commercial side, especially if you’re doing something fairly technical, enterprise customers will seek a direct relationship with you as the developer of that open source code. This relationship might be in the form of a managed cloud service or support for self-hosted software, and this is how vendors typically monetize.

Ben: Purchasing technology from a small company is a risk, as you’re betting they’ll be around long term to support you. With closed-source products, you rely on source code escrow as a way of protecting yourself. This means that if a company goes under, you inherit the source code and you can then try to support the code yourself. It’s a terrible situation to be in and I’ve seen it fail more often than it succeeds.

Open source eliminates that risk, allowing self-support from the start. Additionally, oftentimes a potential technology will get flagged internally when the vendor is new or small. Open source alleviates much of the risk and concerns you’ll face from multiple departments at a company, making onboarding a new technology so much easier.

Several customers we spoke to during due diligence viewed pgEdge’s embrace of open source as a key competitive advantage, and we strongly agree with that.

3. pgEdge was founded by industry veterans with decades of experience in Postgres. How has your prior experience influenced the way you’ve built pgEdge?

Phillip: Postgres was created by Michael Stonebraker, a renowned database researcher and creator. The open source project has been around for about 25 years, and some of our engineers have worked with it for almost that entire time and have made major contributions to it.

Overall, we just have a great team and I love working with them every day. Many of us have worked together before and that long, shared history allows us to move quickly. We have a common understanding of what it takes to build a successful subscription-based software business, which has shaped both our market approach and our technology choices.

As a team, we strongly believe that since Postgres is open source, we should be too. We also strongly believe that Postgres’ architecture is capable of handling nearly any major enterprise data management problem, from AI to geospatial applications. These previously required specialized databases, but Postgres has a fantastic extension architecture to address new requirements. If the core Postgres engine doesn’t do something you need it to do, you can readily enhance its capabilities by adding an extension. We respect the Postgres architecture and our distributed database technology is implemented as a set of standard Postgres extensions, unlike some competitors who add limited Postgres compatibility on top of what’s essentially a proprietary or mono-vendor NoSQL database engine.

4. How do you see the demand for distributed Postgres evolving over the next few years?

Phillip: The demands on applications are only increasing, with expectations for zero downtime, high levels of user responsiveness and ultra high availability on the rise. Going fully distributed has proven to be the best approach to meet these needs. Organizations deploying to the cloud have realized it is not enough to simply host their application stack and database in a single cloud region. As these demands on applications increase, the demand for distributed Postgres is only going to increase as well.

AI is a big driver of that increasing demand. For global organizations deploying AI at scale — such as using machine vision on factory lines for defect detection — the AI inference needs to happen close to where decisions are made. Our fully distributed Postgres technology allows databases to be positioned near users and decision points, enabling AI applications that wouldn’t be feasible with centralized architectures. This makes AI, especially AI inference at the edge, a key driver of distributed Postgres growth.

Ben: Postgres already has a rich ecosystem of plugins and extensions that add capabilities you’d otherwise have to purchase and manage separately. There are popular extensions that provide essential capabilities for AI applications, making it likely that enterprises already using Postgres will prefer to stick with it rather than invest in specialized products. For example, there is a Postgres extension that offers vector database capabilities, which can compete with standalone solutions like Pinecone.

For large enterprises, using existing databases with extensions can be more efficient than introducing new products and vendors. Additionally, Postgres offers advanced machine learning capabilities that will attract enterprises for similar reasons.

5. What company-building step are you most challenged by right now?

Phillip: Our biggest challenge right now is scaling. We have a growing customer base and an increasing volume of leads coming into our pipeline, but we still have a relatively small engineering team and we don’t want to stretch them too thin. We want to manage growth carefully without significantly increasing headcount or burn rate, and we’re actively discussing the right level of investment in growth during our board meetings.

In Conclusion:

Ben: Postgres has become foundational technology across countless applications and businesses. We are so excited about pgEdge, a technologically sophisticated company that’s found an open and standard way of adding an essential missing feature to Postgres. Given Postgres’ widespread use and significance, we see a tremendous opportunity for pgEdge. We made this investment because we see that opportunity, and we love the sophistication, power and experience of the team.

--

--

Rally Ventures
Rally Ventures

Written by Rally Ventures

We unite a thriving ecosystem of emerging business technology. For more information visit www.rallyventures.com