Enhancing Enterprise Graph Analytics – Victor Lee, TigerGraph

This is Textron tv. Hey everyone, welcome back to another session here on, uh, Techron tv. We have a returning guest.

I'm really glad to have Dr. Victor Lee. He's the director of Product Marketing over at Tiger Graph, and, uh, he'll tell us all about Tiger Graph and some exciting new news coming out of Tiger Graph.

Victor, it's a pleasure to have you back here again. Welcome. Thank you.

It's a pleasure to be with you again, Alan. Very good. So, Victor, before we jump into Tiger Graph, let's hear a little bit kind of about your journey and you know, your director of product marketing, as I mentioned, but you know, let, let's hear how you got here.

Wow. Yeah, so probably the last time I spoke with you, I was head of our machine learning and ai. Um, so I've moved around within the company.

I've been with Tiger Graph nine years. Wow. I came out after, uh, I worked in industry, originally in electrical engineering, but then went back to school, got my PhD in in graph data mining, which, um, and I've always loved databases.

I've, I've loved making use of data and graph was an exciting new technology. So I've, I've been with Tiger Graph since it was a pretty small startup, been in head product management. Um, I used my academic background to help us jumpstart our, our machine learning initiatives.

And, and now I'm in product marketing. Actually, we all, they all, everybody winds up there sooner or later. I guess.

How, how are you, how are you liking, how are you liking this new role in product marketing? It's exciting. It, it, you know, one of the things that I've always liked to do is to take my understanding of the technology and try to explain it to people with all different backgrounds.

Um, you know, whether they are technical but not familiar with this particular technology or they're not particularly technical at all and just want to understand what does this mean to me, why should I care? Um, so I think that's actually, you know, as a professor I did that, um, working in, in as a technical writer. I've done that and, and now I'm just doing it if for a, a slightly different emphasis in marketing.

Absolutely. You know, what I find interesting is so many, look, we, we serve a technical audience here at Dextro, right? And a lot of our folks out here, you know, they, they don't know you.

They, they'll, they're gonna look at the, you know, the writeup or the ad or whatever for this session. They're gonna say, oh, it's a doctor. He is a PhD, but yeah, he is director of product marketing.

Probably little light on the tech. It's not often you get a PhD in graph Databasing and so forth, right? Who happens to also be the director of product marketing for a graph database company.

And, you know, anyone who would kind of question your technical chops, uh, to, to talk about graph databasing, I think is obviously sadly mistaken. So you can't always judge a book by its title and you can't always judge someone by their, by their title either in, in terms of their technical expertise. Um, Victor Tiger graft, I mean to me is, you know, leader in, in graft databases, but there are people who are watching this out here who may have not heard of Tiger Graph and, and quite frankly, may not even be familiar with graph database and, and, uh, you know, uh, graph analytics.

So why don't, if you don't mind, can we give them a quick primer, get 'em up to speed? Absolutely. Um, we've been trying to explain what is, when we say graph, what do we mean by that?

We've been, we don't have to do it as often as we used to, but there, there's still new people. First of all, people generally understand what a database is. It's, it's a particular software tool for storing and accessing data right now, what makes it a graph database?

In this case, graph is a mathematician's term for the idea of a network where you have, um, nodes and connections between the nodes. You think of a social network and the way we envision that of this person knows, you know, several people and they know several people and it forms a network. Mathematicians call that type of structure a graph.

So a graph database is storing nodes of information and the connections to those other nodes. And it's the connections that are so very important because when you think about analysis, what do I wanna know about something? How can I get a deeper understanding of it?

You wanna understand how it relates to other things. I, I know about, okay, I know about this city, or I know about this company, but I wanna know who lives there or who are its customers, what are its products and, and how does it compare to its, its competitors or alternatives. It's that relational understanding that context that really gives you a deeper understanding.

So as companies are looking to take their data and get more out of that data, what I have this data, I have a lot of data, but I've, I've seem to be somehow plateauing on what I can squeeze out of it. Graph has offered a great way of looking at your data differently by emphasizing the relationships. There are a whole set of, uh, standard algorithms and standard analyses you can do.

And there are things that are particular to your situation. Things you know about your business. You know, I know, you know, we have a lot of customers in the, in, uh, financial institutions who wanna do fraud analytics.

They already know some of the patterns of fraud and they wanna way to be able to describe it. And graph is a great way to describe the relationships of these fraudsters do this activity. They work together, they work with these multiple banks to, you know, fool them.

But if you can see the whole pattern, then the fraud is revealed. Same thing works in other applications for better recommendation, obviously you can model a supply chain or a, a delivery system with a network. So there are all sorts of different applications and Graph is giving people one other tool to enhance the, the types of data analysis that they're currently doing.

Excellent. I expected no less from you, Victor. Thank you.

Um, alright, so now we've kinda laid the baseline, right? Tiger Graph is, is as I mentioned, leader in in the space. You guys recently, uh, announced a, a new release, new feature set.

Why don't you kind of give us the, uh, the inside scoop on that? Sure. Um, thank you for, for naming us as the leader in, in graph.

We of course do, do have some competitors, but what distinguishes us is our performance and scale. We are the, not just the fastest, but we can grow. We can handle enterprises that have a lot of data now, or businesses that are starting smaller and, and expect, well, they may grow someday because we're the only graph out there that's a distributed database.

That is, you know, if you need to expand to multiple servers because your data's growing, we can do that. We're the, we're the only graph that can do that. Um, so we've always had this leadership and performance and scale and in complexity of analysis, what our emphasis of late has been on in improving the, uh, enterprise support that large businesses require and making it easier to use for everybody.

3 has four, four key improvements and, and three of them relate particularly to enterprise use. One is giving them more options for workload management. Businesses are using their database not just for one task, but for multiple tasks and they have multiple users concurrency.

Um, so there's lots of things going on and you want to be able to control who is getting, um, how much resource so that you're prioritizing things. Um, so we're adding, giving one, adding to our suite of existing workload management techniques. One more this, this one lets them say, um, look for the who's least busy right now and assign the job to whoever is which server is least busy.

So that's, that's a nice technique. Another improvement, um, is giving them more real-time analytics, excuse me, real-time monitoring and, and consistency checking when they're doing high speed data loading. Um, we can handle multiple terabytes of data and people, they've got a lot of data they're loading and they need to know as soon as possible if something is not going quite right.

Maybe it's a, it's a new loading job. It, and they haven't quite, um, got the logic right for, for how to take their data files and put it into the graph. So they need to check right away if things aren't going right.

So they don't, you know, oh no, I've, I've loaded 10 terabytes of data and it wasn't right. So, um, this is a, a realtime monitoring to help them see, um, more quickly if, if things are going as they intended. Um, third more Kubernetes support.

Um, Kubernetes has established itself as, as the standard way to do container based orchestration. That is thinking of your software resources, you know, bundling them into these separate packages called containers and being able to build your overall system as a set of containers. Kubernetes is the standard way in the industry for saying declaratively what you want.

And then the Kubernetes will automatically say, oh, okay, this is what you want. I've got these containers to work with, let me build it for you. Let me continuously monitor to see if you're getting what you asked for.

If something needs adjustment or correction, I'll take care of it. So that's what the Kubernetes framework can do. And Tiger Graph is, is, is making additional steps to enhance our Kubernetes support.

Excellent, excellent. Anything else? Did we mention Open Cipher Have four improvements, those three or all for, well, they could be for any user.

They're particularly for our enterprise users, things that they've asked for. And so it's benefiting our current customers and it's helping, um, new customers. The fourth is we're adding support, even more support for the Open Cipher graph query language.

Um, again, graph is, has sort of been growing over the years and different, different vendors like Tiger Graph has its G SQL language. Neo four J, another well-known graph company has Cipher and there's a version called Open Cipher that is, um, run by a, you know, a public consortium. And, and so a lot of people start learning about graph using Open Cipher.

And so we've added support for Open Cipher. This enables the people that already know it that to, to get started using Tiger Graph immediately. The differences between Open Cipher and G sql Open Cipher works with a lot of graph platforms.

It's more general purpose. It's for typical graph querying. If you want to do analytics, uh, and particularly on a distributed parallel platform.

G SQL is the language we you want to use. So what we've done is we've embedded Open Cipher within G sql. You just take, you can take an open Cipher query, put AG SQL wrapper around it, just a header, and then it'll run within our G SQL engine.

And then as you say, okay, that's great for basic querying, now we wanna do some analytics. You can, you can easily switch over and, and start to, you know, add on the analytics to what you already have in, in Open Cipher. I love it.

Victor, for people who want to get more information that, you know, this is just a quick interview, but they want to take a deeper dive, where can they go on the web? Sure. com.

Um, we also have a YouTube channel with lots of videos that have introductions to the product. com. Um, I've, I've got time to chat with people and I, I love to do it, um, to, to get in touch with, with, you know, real people out there.

Excellent man. Victor, thank you so much for coming on and telling us about the new features in Tiger Graph, continued success and keep us posted. Thank you, Alan.

It's been a pleasure. Alrighty, Dr. Victor Lee, director of product marketing, newly minted director of product marketing over at Tiger Graph.

We're gonna take a break here on Text Junk tv. We'll be right back. I.