Graphs on GPUs: 5-Minute Interview with Saul Rosales

Tell us about how you work with Neo4j

The Graphistry and Neo4j relationship is very interesting. The one way that we work with Neo4j is we’re solving a lot of the same problems. The great thing about Neo4j is they’ve been around a little bit longer, so they’ve had a chance to distill some of the new problems that are out there that are being presented to graph databases.

Where Graphistry comes in is we actually sit on top of Neo4, and we’re able to allow customers to view a hundred times more data than they’re able to look at right now. So that’s how we’ve come into the Neo4j relationship, and we’re happy to partner with them.

What is Graphistry?

Graphistry leverages GPU technology. And by leveraging graphic processor technology in our partnership with Nvidia, we allow our customers to see about ten times to a hundred times more data. The more data that you can look at, the more important decisions you can take.

Can you talk about Graphistry 2.0?

What’s amazing about 2.0 is that we’re building upon what we did in 1.0, and we’re actually building it on what’s called the RAPIDS framework. That’s Nvidia’s framework, that allows us to use end-to-end GPU on an ETL setting. So, what does that mean for everybody? It means that you get larger graphs, and now you’re able to process information a lot faster.

What made you choose to partner with Neo4j?

One of the reasons, outside of the technical realm, why we partner with Neo4j, is we’re complementary in many ways. We’re dealing with the same type of business problems, where customers are approaching us and trying to find new ways to explore graphs.

One of the symbiotic relationships that we found with Neo4j is, on our end, we work a lot with cybersecurity investigations, and from what I’ve heard from different salespeople and from some of the business development people, that’s something Neo4j would like to get to. So, I believe that there’s a relationship there, where we learn from their business use cases, and we can transfer some of our knowledge to Neo4j, outside of just the technical integration.

What do you think is in store for the future of graphs?

For me, it’s an interesting question, because coming from the cybersecurity side of the house, I think there is a merge of data science that is starting to happen. But as customers work with larger data lakes and larger datasets, I see that visualization is playing a big part in that, especially in security, where you’re trying to meld different alerts, different logs, different security products.

What customers are struggling with is trying to interpret those large datasets. So, a lot of them are trying to lean down the path of artificial intelligence and machine learning, finding ways to automate that process. And what we found is, the human mind, when it’s processing, let’s say a malware attack, or trying to understand network mapping, or something to that effect, has a really good intuition of what that looks like.

So, what we found with those data teams is, when we present a graph, I think they have a faster ability to be able to resolve a problem than they did before on just looking at static data.

Want to share about your Neo4j project in a future 5-Minute Interview? Drop us a line at content@neo4j.com