TigerGraph 3.5 – Building a Cloud-Native Graph Database Platform

I’m delighted to announce the release of TigerGraph 3.5. This release introduced several improvements, many of which were added thanks to direct feedback from our customers and partners. 

Included in the release were some manageability and security features, data streaming support, and a preview release of Elastic support that puts us on the path toward making our product a cloud-native graph platform.

Elasticity Preview Release

Elasticity is the biggest differentiator in the TigerGraph 3.5 release. Even if it is currently only a preview for early adopters and proofs of concept, the support to work with an Elastic read-only cluster brings us one step closer to making TigerGraph a cloud-native graph database platform.

Our customers often don’t know how much hardware or compute capacity they’ll need ahead of time. This becomes an issue when they encounter big queries that require more resources than they’d planned for. Elasticity eliminates the need to plan for big demand spikes and balances the tradeoff between capacity planning and cost efficiency.

As we expand on our elasticity support, our goal is to eliminate our customers’ requirement to pay for a huge cluster up-front and, ultimately, to build a solution that will automatically scale to handle spikes in traffic.

Manageability and Security

Our customers were vocal about the need for two specific improvements to manageability and security, areas we addressed in the TigerGraph 3.5 release.

Datasets are never perfect. In a graph database, this can lead to encountering failed or ‘dead’ nodes. With our latest release, these nodes can be removed from a cluster to ensure the dataset remains highly available for use. Using the expansion feature released in TigerGraph 3.2, after node removal the user can restore the cluster back to its original size.

On the security side, we heard especially from solution providers using TigerGraph that they needed enhanced role-based security to help control who can read, execute, and edit queries. With the newly released granular access control features in TigerGraph 3.5, these customers can now allow or restrict access to these queries on a role-by-role basis.

Data Streaming Connector

To provide faster, more scalable, and more reliable data streaming between TigerGraph and other data systems, we introduced support for Google Cloud Platform (GCP) streaming via a data streaming connector.

This fault-tolerant data streaming capability makes data stored in GCP buckets available to TigerGraph without needing to worry about a stream failing midway; if the stream fails, it will pick back up where it left off when it resumes. We will be working to introduce support for other cloud storage solutions in future releases.

GSQL Syntax Default

Finally, specifically for developers working with TigerGraph, we’ve updated our products to use GSQL V2 as the default options for building queries. V2, introduced in TigerGraph 2.4, has since been augmented with performance tuning and other improvements. 

V2 allows for more powerful pattern matching and simplifies the language, removing lines of code building queries when compared to V1. GSQL V2 is much more precise and flexible as a result. The timing for this change was also driven by the launch of the Graph for All Million Dollar Challenge, ensuring that entrants in the challenge were developing their solutions using GSQL V2.

The feedback we receive from our customers and partners helped drive the enhancements introduced in this release, and we have even more coming in the months ahead as part of planned releases.

For full release notes about TigerGraph 3.5 and previous releases, visit docs.tigergraph.com.