This article is more than 1 year old
Neo4j graph database boss: 'The mainstream is always under attack'
CEO bullish as biz pushes Cypher QL and new DBaaS
Emil Eifrem, chief exec of Neo4j, has said that graph databases are at a turning point for so long that you'd imagine he's bored of the line by now. But 2018 might just be the year his predictions come true.
Speaking at his firm's Graph Connect event in London last week, Eifrem once again asserted that the graph industry is "at a very interesting inflection point".
As evidence, he pointed to a recent report from Forrester that said more than 50 per cent of enterprises were using graph databases in 2017, dwarfing the 25 per cent analysts had predicted back in 2014.
The last year has seen more of the big database vendors offering their own graphs – AWS promising Neptune and Microsoft launching Cosmos. And there's a bevy of startups looking to get in on the action.
"I think the mainstream is always under attack," Eifrem told The Register after his keynote.
"But I don't look at the other graph players – the other graph players look at us. We have 10 times more adoption than the others combined... so I don't tend to focus so much on that."
Eifrem said Neo4j's focus is not on winning over the people who already want to use graphs, but on getting the "5 million developers out there to understand when to use graph".
Neo4j moves up stack: Ahem. We're a graph company, not a graph database companyREAD MORE
However, he is quick to add that the biz has "never lost for performance reasons to a graph database. Never in the history of the company."
He is also bullish about TigerGraph – widely touted as one to watch in the space – waving away the hype and their benchmarks as one of "a bunch of small players" and pushing conversation on to the entry of the likes of Oracle, AWS and SAP. This, he said, proves "the category has arrived".
Eifrem acknowledged that the company will end up losing deals – "then some of it will become more real" – but maintained that the "leader of the category will be always disproportionately rewarded".
Attracting the enterprise
Nonetheless, as competition hots up, Neo4j needs to keep up. Last year it launched its Native Graph Platform to add analytics, data import and visualisation on top of its database, and last week announced it was launching a Database-as-a-Service.
"We increasingly hear from users, customers, prospects that they... really don't want to run any servers," said Kurt Freytag, director of product management. "They don't really want to manage it. Not everybody has a DevOps team. And that they want a lower cost way to get started.
"Not every project has the ability to extract the value out of Neo4j immediately that lines up with the enterprise business models. They want to go with a pay-for-what-you-use model."
The solution is Neo4j's service on public cloud, which offers a cloud-hosted database with automated backups, upgrades and management that can be set up "in less than five minutes" and is now being offered as a "very limited preview".
Neo4j is trying to push adoption of its graph query language, Cypher, as no language has yet asserted its dominance in the market.
It has already created openCypher, which allows it to be used for graphs built on databases like SAP HANA and for Redis Graph and AgensGraph.
Now it has launched Cypher for Gremlin – another query language that is run by a number of alternate graph platforms.
"Cypher for Gremlin allows developers to access Gremlin databases using Cypher," said Freytag.
"What this allows you to do is use Cypher and invest in Cypher and be sure that when writing in Cypher, you can be confident that any work in Cypher that you do will be able to run across any number of graph technologies."
Freytag asserted that it was about "promoting confidence that there is a language out there that everyone can run" – but wider adoption of Cypher is sure to benefit business.
And Eifrem said that it will help Neo4j persuade enterprises to take the graph plunge.
"Years ago, I was talking to architects, now I'm talking to Fortune 500 CEOs," he said. "And they wants assurance that, if they commit to this [Neo4j's database] and train 10,000 devs [in Cypher] at their bank, they want to know that it's portable."
But this highlights a fundamental question about graph databases: Neo4j has been around for nigh on two decades. So why, when other non-relational databases have taken off with some fervour, has graph been seen as the niche option?
"Essentially, it comes down to one thing," Eifrem said.
"With a lot of these new categories of databases – document, key value and so on – it was always up to the leader to evangelise. With graph, clearly that's been us."
And that message hasn't been spread as widely for graph, he said, because of Neo4j's different approach to business, which he says merges the values of the West Coast and his home nation of Sweden.
Panama Papers graph database cracked open for world+dogREAD MORE
"What a lot of other people chose to do, in the Hadoop world, in the document world, is to say 'we're going all in on adoption'. They spent all of their dollars trying to get it out, trying to get it used – not commercially used, just used.
"We didn't put as much into adoption... and I think people thought graph databases were niche because we weren't as obvious as Hadoop."
Eifrem said that Neo4j had "chosen a different path" that he describes as "a marathon, not a sprint".
"At the end of the day, we can all raise a lot of money from people, but we want to grow with our customers," he said.
"That's a much healthier way of building a company – we're not just a Silicon Valley company, we're a 'Sveeedish' Silicon valley company." (Swilicon Valley? suggested your portmanteau-loving correspondent.)
But, however much Eifrem waxes lyrical about social impact – and he loves to talk about the firm's work with journalists to unpick political and business misdeeds (it's the engine behind the Panama Papers, Paradise Papers and recent investigations into Russian trolls) – he has to acknowledge businesses need to make money.
"I don't think our entire job is to make money, but we have to make money... and there is a trade-off; the more people use if for free, the less money [there is]."
Nonetheless, he remains confident of his slower approach. "I want to build a massive company, with tens of thousands, if not hundreds of thousands of staff," he said. "And you don't do that overnight." ®