Breaking news, years in the making. That’s how we would describe today’s dual announcement by graph database startup TigerGraph. TigerGraph, which came out of stealth in 2017, but has been in the works since 2012, has been making strides.
Today TigerGraph announced the general availability of TigerGraph Cloud, dubbed the first native graph database-as-a-service, as well as $32 million in Series B funding. The investment, led by SIG, will boost TigerGraph’s global expansion, which TigerGraph notes is fueled by TigerGraph Cloud. This market is estimated to be worth more than $6 billion in 2022.
$32 million is a hefty amount, but it’s not the biggest we’ve seen in this space. A few months ago Neo4j scored $80 million in Series E funding. We’ve also seen Dgraph score $11.5 million to pursue its unique and opinionated path. And, hint, we do expect more funding news to be breaking soon. This breaks down to a couple of things.
First, there is substance beyond hype in graph databases. VCs are not in the habit of repeatedly opening their wallets to chase windmills. The total amount of investment in the last 12 months is in the area of $150 million. This may not sound like much compared to other red-hot technology areas, but do expect to see investment grow even further.
Besides, this only represents a fraction of the overall effort dedicated by vendors evolving their products, users building applications, researchers advancing the state of the art, and marketers spreading the word. This space really is booming, the Gartners of the world have taken note, and this is just another affirmation.
The other thing to note is that this market is still up for grabs. When discussing with a VC a couple of months back, one of the key questions was whether Neo4j can already be considered the winner, based on their overall better funding. Our answer was that these things change. Though we did not have any inside information to base our estimate on, and Neo4j still is better funded, the gap may be closing.
Last point on funding and market analysis: when connecting with TigerGraph’s COO Todd Blaschka and VP of Marketing Gaurav Deshpande, we did ask whether they could share any information regarding the relationship with the new investor, such as board member placement for example. Although that was not disclosed, we speculate it’s highly likely – $32M is quite an amount for a vendor like TigerGraph.
TigerGraph Cloud, a cloud-based solution for analytics
But funding is not the only way this announcement is shifting the landscape. Perhaps more importantly, TigerGraph is the first graph database to offer what looks like a turn-key cloud-based platform for analytics.
To be clear, the majority of graph database vendors today already support cloud deployment. And we also have AWS and Microsoft Azure with their own offerings there, Neptune and Cosmos DB, respectively. But although having your database run in the cloud sounds great, in most cases, you still have to do at least part of the provisioning and managing, and connect it to your data pipeline.
With TigerGraph Cloud, you don’t. That’s what TigerGraph promises. Blaschka mentioned all it takes to use TigerGraph Cloud is an account. TigerGraph Cloud is a managed platform, and one that also incorporates key parts of modern data pipelines such as Apache Kafka and Apache Spark.
The idea is actually quite close to what Databricks does, according to Blaschka. Databricks is the vendor offering a commercial cloud-based platform based on Apache Spark. More like an iPaaS, on which the point is not so much to have someone run your infrastructure for you, but more to have a platform you can use to get insights. Managed data pipelines are just a part of this.
The key part is what you can do with this. TigerGraph has been putting effort into making its product more approachable, by means of a visual environment, out of the box support for graph algorithms, and getting people up to speed with GSQL, its graph query language.
In addition, TigerGraph Cloud comes bundled with Application Starter Kits. There’s more than a dozen out-of-the-box kits for fast application development, for use cases such as customer 360, fraud detection, real-time recommendation, enterprise knowledge graph, machine learning, explainable AI, and more. We did have a quick demo on those, and they seem easy to use.
Start in minutes, build in hours, deploy in days is TigerGraph’s motto for its service. TigerGraph Cloud also comes with a free tier, meant to enable new users to learning TigerGraph, prototype and develop applications. TigerGraph will grant users a free instance which Blashka said is ideal to for learning TigerGraph, prototyping and development. Pricing is elastic: users only pay for hours they use, billed monthly.
Beyond analytics: performance and interoperability
Analytics, however, is not the only goal for TigerGraph. Blaschka touted TigerGraph as an HTAP (Hybrid Transactional Analytics Platform) solution, capable of handling analytics, transactional and real-time workloads. TigerGraph promises 100,000 real-time deep link analytics queries per second on a single machine, with a number of benchmarks having been released – and commented on.
The other technical aspect TigerGraph has been emphasizing is query language. The landscape as far as graph query languages go is quite fragmented. Lately, there has been ongoing effort under the auspices of W3C to come up with a standard query language for graph databases. Or, to be more precise, for property-graph-based graph databases, since there already is SPARQL for RDF graph databases.
Recently, it has been announced that ISO/IEC JTC 1, a key international standards body, has approved a 4-year project to set a Graph Query Language standard – GQL. TigerGraph has been a part of this effort since 2018, and Blaschka said they are thrilled to see this big step forward for the GQL team:
“Though still in draft and incomplete form, GQL is drawing heavily from existing languages and prior academic work such XPath, StruQL, WebSQL, Cypher, GSQL, PGQL, G-Core and SQL/PGQ. Each of the major languages in existence has made some valuable contributions and innovations and we’d like to acknowledge our fellow vendors and partners that have been involved in the journey including Amazon, ArangoDB, DataStax , Neo4j, Redis Graph, Optum, SAP and others”.
We agree with Blaschka on this one, and it’s a milestone also hailed by others in the domain. This is just the beginning of what we expect to be a long and complex process, but one we do hope will result in the synthesis of a query language based on the best each contributor has to offer. Interoperability is key to usability and market growth, and this is something vendors in this space seem to be aware of.
NOTE: An earlier version of this article has been corrected to remove the erroneous reference to a limit of 10 hours of use for TigerGraph Cloud free tier.