TigerGraph emerges from undergrowth with 2.0 release in its jaws

Move over, doggy. 2018 is 'the year of the graph' apparently

By Rebecca Hill


Database upstart TigerGraph has launched its latest platform, pitching the ability for enterprise customers to collaborate in real time.

The business was founded in 2012 but only came out of stealth in September. Today it is launching its 2.0 platform, which it says offers more collaboration, and faster performance, than other graph databases.

Having flown relatively under the radar recently – privately-funded Neo4J has had the market mostly to itself for more than a decade – graph databases are firmly planted in the mainstream these days.

TigerGraph founder Yu Xu told The Register that 2018 would be "the year of the graph". Graph theory isn't new, he said, but graph database was a niche market until the age of big data.

"At the start, people were just happy to store data... then they started to think about simple analytics... then started talking about a method for simplified computation," Yu said.

"People realised that to connect the dots, a graph database is best – [so] graph database and big data started to merge. Everything is coming together, which is why we're seeing such a big movement."

This is emphasised by the increase in traditional vendors offering graph databases – Microsoft's Cosmos is ranked second behind Neo4J in DB Engines' graph DBMS ranking, and AWS launched Neptune in November.

TigerGraph claims to out-compete these legacy and traditional vendors by offering a distributed, parallel graph computing platform that runs faster queries and uses less space to store the data. It promises to load 50 to 150GB of data per hour, per machine and traverse hundreds of millions of vertices/edges per second per machine.

"We define real-time as data coming in, being able to process that data in less than a second and being able to take action on that data in that business moment," said chief operating officer Todd Blaschka.

The database is pitched firmly at enterprise customers that they hope will buy into the idea of genuinely real-time processing of very large data sets, for instance recommendations engines and fraud detection.

Customers include major e-commerce firms Alipay, Visa and Wish, along with Uber and the State Grid of China.

Blaschka claims that Alipay has "arguably the world's largest graph database on TigerGraph for its anti-money laundering business". It streams 2 billion events in real-time, he said, with the graph having over 100 billion vertices and 600 billion edges on a cluster of 20 machines.

The latest release pushes more enterprise upgrades, with the main focus being real-time graph collaboration through what it calls multi-graph.

The idea is that multiple groups – like product, recommendations and security teams – can share one master database and benefit from real-time updates to that underlying data.

The biz said that local control and security settings would ensure the customer still meets data compliance regulations, as it will be able to dictate who can access which data.

"With TigerGraph, we're able to provide different views working off the same underlying database, and as data in one batch of the database gets updated, all those groups will benefit in real time," Blaschka said.

The biz says that while multi-tenancy "enforces data silos", multi-graph supports multiple people using the same underlying data set.

"There's a significant performance enhancement because it will give customers more access to data to analyse, and make new business decisions they wouldn’t have before – or that it was very time consuming to do," said Blaschka.

Sign up to our NewsletterGet IT in your inbox daily


More from The Register

Ask your data anything, says data viz biz Tableau as it flings 'patent-pending' tech at natural language tool

Plain language questions, automated data prep promised in first release of 2019

French data watchdog withdraws probe from location data guzzling adtech biz Vectaury

CNIL says firm now collects valid consent, shutters case privacy-watchers hoped would help see off adtech's model

Register Lecture: For AI's sake – taming the wild data frontier

Blighty's metrology experts define terms

Former headteacher fined £700 after dumping old pupil data on server at new school

Why? Er, 'professional reasons'

Trakt app users' personal data exposed: We were hit by a 'PHP exploit'... back in 2014

No payment info, but users' names, locations, email addies etc all 'lost'

Can you trust an AI data trust not to slurp your data?

Comment Data ownership is so yesterday. Give us all you've got... ethically

HPE squirts cold data at on/off-prem object stores, adds capacity to StoreOnce for ever-increasing data firehose

Someone's gotta service that embiggening backup burden

Node your customer, right NetApp? It's only gone and hooked up MAX Data to HCI

Analysis PMem software hustle for the hyperconverged – if all goes to plan...

Keeping up with the kollect-kash-ians: Data manager Komprise more than doubles funding

Thanks to whip-round from WD investment arm and co

Microsoft wants to cart your data away in a box and punt it onto Azure

Ignite 1PB of Fedex-able rack-on-rollers