pgEdge, the fully distributed PostgreSQL database company, is emerging from stealth with the launch of its fully open, fully distributed database optimized for the network edge and based on the open source PostgreSQL database. Backed by its $9 million seed funding led by Sands Capital Ventures and Grotech Ventures, pgEdge is driving low latency, high performance, and global distribution for applications accompanied by source code transparency.
Leveraging the power of PostgreSQL’s open source architecture, pgEdge Distributed PostgreSQL is a multi-active (or, multi-master) distributed database system that allows compute functions and PostgreSQL to be deployed at, or close to, the network edge, according to the company. This enables application developers and database architects to benefit from high availability and low latency while adapting to data residency requirements without impacting the application's code.
“Our mission is to make distributed databases and distributed applications easy,” said Phillip Merrick, co-founder and CEO at pgEdge. “In today's world, applications need to be always on. They typically have to be globally distributed in many cases, and are at least available in multiple regions around the world. And because so many people now have gigabit internet connections into their homes and their offices, they expect everything to be a whole heck of a lot faster. We think there's tremendous pent-up demand for distributed databases, and it just needed a company to come along and make it easy.”
pgEdge’s multi-active architecture enables read and write processes to be enacted at any database node on the network, where nodes in a single pgEdge cluster can extend over several cloud regions and data centers.
“The beauty of multi-active architecture is that every node in the cluster can take read or write traffic. You might elect for some of those nodes to be read-only nodes, but in general, every node can take reads or writes, and that gets us away from the pretty monolithic architecture that most applications are running on,” explained Merrick. “Say the proprietary Oracle distributed database stuff, like most applications, still has a single monolithic copy of the database running somewhere on the network, often in Amazon's US East 1 region. And of course, when Amazon's US East 1 region goes down, as it sometimes does, it takes out a good chunk of the internet because everybody's got their database there and it's not replicated to some other region.”
“So multi-active technology lets you solve that. Not only getting the faster page load times in your application, because you've got databases closer to your users, but you've got something that's more resilient. If you've got a failure in one cloud region, then you can failover traffic to other cloud regions where you've got your nodes,” continued Merrick.
Supported by a configurable conflict resolution and avoidance algorithm, node write conflicts are mitigated for you. pgEdge nodes run standard PostgreSQL v14 or v15, including the “Spock” extension.
pgEdge can reduce long webpage load times, aid enterprises in reducing costs due to a reliance on proprietary databases, as well as better comply with data residency requirements via pgEdge’s “geo-sharding” for configurable data sharing options.
Source code availability, a feature highly emphasized by pgEdge, is available on Github and covered by the pgEdge Community License.
“Unlike our competitors, you can actually look under the hood because the source code is available,” explained Merrick. “Under the hood, it's asynchronous, logical replication of the data. So, you're actually replicating at the row level, not at the physical block level, which, of course, is physical replication.”
pgEdge offers pgEdge Cloud, a database-as-a-service (DBaaS) version of the solution (monitoring databases in any of over 100 zones across AWS, Azure, and Google Cloud), and pgEdge Platform, a downloadable software for self-hosting either on-prem or in the cloud (available for deployment on AWS, Azure, and Google Cloud).
Existing users of PostgreSQL can also benefit from pgEdge; application workload management agility and rapid failover capabilities (if each node can withstand both read and write traffic) can be applied to either edge deployments or just a few cloud regions.
“Because it's 100% Postgres running at each node, it absolutely acts like Postgres, because it is Postgres. And so all the customers, applications, tooling, and expertise—that all just comes across,” said Merrick. “Existing Postgres applications can come over into this distributed world of pgEdge with either no, or very little, code changes.”
pgEdge Cloud is now available in a limited private beta program. The pgEdge platform is in open beta and available for free download; enterprise support subscriptions are also available for the platform.
For more information about pgEdge’s emergence from stealth and platform launch, please visit https://www.pgedge.com/.