Graph database upgrades are like buses. First Neo4j updated its wares, and now TigerGraph has pushed out new features it hopes will appeal to a broader range of enterprise folk.
The thinking behind the latest release from TigerGraph is that its own graph database - and graph DBs in general - might have more traction if they weren't just the preserve of data scientists and technical specialists. The firm wants to lure business analysts and more lay users into the graph approach, which organises and analyses data as a network of relationships — rather than dropping it in rows and columns — and can be used for recommendation engines and fraud detection.
Top of the list of new features is "No Code Migration", which is designed to allow users to extract data from relational data stores by letting them convert schema and load data to a graph by using drop menus and tick boxes, rather than writing any scary code.
Similarly, TigerGraph 3.0 offers a Visual Query Builder to non-technical users to create and deploy queries by drawing shapes or patterns they want in the same sense visual data modelling does.
TigerGraph chief operating officer Todd Blaschka told The Register: "What we have been developing is the technology to bring graph to the enterprise.
"The challenge has been that graph databases have been seen as a developer tool, not as part of a business application. And what we're seeing is when we're working with, for example, a fraud business group, [is that] the business case is very clear: 'If I can link name, phone number, and email address, with my list of known previous fraudsters you see how they are connected'.
"[Business users] already know the question and they've just been trying to find the right technology to answer that question."
Bryan Betts, principal analyst with Freeform Dynamics, said TigerGraph's strategy is to position its product as an analytics tool, rather than a database.
"Graph is its own subject and part of a well-established database market. Anyone trying to bring in new products into that has a huge incumbency problem to get over. It's the same for Neo4j and TigerGraph."
TigerGraph is pitching its technology to those used to querying relational databases with SQL – its GSQL query language will be familiar to SQL buffs, the company said. So the argument goes that they can lift data they already have and analyse it in a way they understand, Betts said.
But to succeed, TigerGraph needs to sell to business users rather than developers, he said.
On the engineering side, Neo4j was perhaps more mature and programmable than TigerGraph, but both systems were distributed and could scale well, he said.
TigerGraph 3.0 includes other new features, such as faster scaling across distributed environments and user-defined indexing, which lets users speed up the database performance for specific queries, the company said.
The new product will be on limited release at the end of March and generally available there after. Only then will we see if the TigerGraph who came for tea will eat the lunch of its market rivals. ®