Building a cloud-native, horizontally scalable graph database — Our investment in Dgraph
By Patrick Chase and Satish Dharmaraj
If you’ve ever searched on Google, ordered an Uber, or used a credit card — you’ve most likely interacted with a graph database. Graph databases are not new, but they have been gaining popularity in recent years because it is becoming more and more common to build applications on top of highly connected data. For example, at LinkedIn, graph databases were a core part of our infrastructure because we had to quickly serve requests like “Get me all of Jane’s connections.” A graph is the fastest way to do this because it stores the actual relationships in the data instead of computing all relationships on the fly.
The need for graph databases has grown as data complexity has increased. Even though data has gotten more complex, companies must maintain the speed of their products to keep customers happy. This has led many companies like Facebook, AirBnB, LinkedIn, and Dropbox to build their own graph databases because they couldn’t find anything in the market that could keep up with their growing data needs.¹ ² ³ ⁴ However, building your own graph database is a huge engineering effort. Today, many companies have a need for a scalable graph database, but do not have the resources to build one themselves.
After years building graph systems at Google, Manish Jain saw a broader opportunity in the market. He founded Dgraph Labs with the mission of building a scalable, open-source graph database that would serve all companies from startups to enterprises. And they are doing it — companies like Intuit and VMware are already onboard and publicly using Dgraph.⁵ ⁶
Since founding the company, Manish has assembled an incredible team to go after this opportunity with deep domain expertise in distributed systems and open-source. In addition, Francesc Campoy recently joined Dgraph as VP of Product to an incredible reception. He is well-respected in the Go community from his time at Google running developer relations for Go and Kubernetes, two huge open-source projects.
The Dgraph team has already built incredible technology with an amazing open-source following. Dgraph recently passed 10k stars on Github 🎉 and is used in production by everyone from Fortune 500 companies to startups that are choosing Dgraph as their primary backend.
Why does Dgraph’s technology stand out?
It is horizontally scalable. Due to its distributed architecture, Dgraph can scale by adding more machines to increase capacity instead of adding memory and storage to a single machine. This means Dgraph can scale to huge amounts of data, which is critical in today’s world where the amount of data stored worldwide is expected to double every two years.
It is fast. As a true graph database, Dgraph can handle complex queries like “Get me Jane’s friends of friends” very quickly. Many systems would need to first get Jane’s friends, then issue a second query to get all the friends of the people returned in the first query. With Dgraph it is one query, so it’s fast.
It is designed for modern developers. Dgraph was built on modern technologies from day one. It accepts JSON input, and produces JSON output. Its query language is modeled off of GraphQL and GraphQL is on 🔥. We found that many people are turning to Dgraph because it fits with the other technologies they’re using like the missing piece to their backend puzzle. We believe that Dgraph can play a much larger role in how the next generation of applications is built.
Today we’re excited to announce that we’re leading an $11.5M Series A in Dgraph alongside existing investors including Bain Capital Ventures, Blackbird, Grok and AirTree. We feel very fortunate that Manish and the Dgraph team chose us as their partner. We are thrilled to join Dgraph on their journey to build the most advanced graph database in the world.