Supporting openCypher – Jay Yu, TigerGraph

This is texturung TV. Hey everyone, we're back at text John TV here. Our next guest on today's show is Dr.

J. U. He's with tiger graph.

Jay welcome to texture on TV. Thank you. It's honored to be here.

It's our honored actually have you here? You know Jay we always like our audience to understand a little bit who who's talking to them? So why don't we start there?

Why don't we go a little bit in your background and then we could jump into your role at Tiger graph and what tiger graph is about? Thank you. My name is jeu I'm the vice president for product Innovation.

Tiger graph my background. I started as a technology. I got my PhD in highly scalable share nothing database systems from University of Wisconsin-Madison backing the late 90s early 2000.

So I'm actually pretty old and I I actually worked for teradata. So that's how I learned all how to apply Advanced Research into the production and then I decided to do something different I moved to into it which is the you know, personal finance Turbo Tax QuickBooks. I was there distinguished architect director in charge of the Enterprise data architecture strategy and design and then over there.

I was responsible to unify multiple daily silos together. That's Discovered graph and that's how I discovered tiger graph. So I have more than 20 years of experience in the industry starting from database and then go to application and Enterprises not back to back and deliveries actually and and that's actually a great segue into tiger graph.

Tell us about tiger graph. Yes. So tiger graph founded actually 10 years ago was the mission to bring graph technology to everybody as you know graph technology has been using big Enterprise like a Google like a Facebook social network, right Knowledge Graph but never was accessible by small meeting company or other prices.

All mission is build the world most scalable what we call native parallel share nothing architecture MPP apply to graph technology and make sure you can put all your Enterprise data into a connected intuitive data model that can be easily used to discover much much deeper inside any other technology And the company has you know, again has grown to the current size. 8 and we have more than 100 Enterprise customers really benefiting all scalable architecture to bring their graph use case to production. Absolutely, and you know Jay our audience is pretty technical.

And I'm gonna assume most of them know what graph database. Right is all about it's not about Graphics. It's not you know.

But for those who aren't maybe or maybe they think they know and they're not sure for those who don't know or not. Sure. You know, what what's the special sauce behind graph databases?

What's the use cases? That really stand out what makes a graph database such a uniquely powerful tool great question. So graph database is very different than relational database as you know, everybody knows relational database.

So I Oracle Microsoft SQL Server all these things based on SQL relational database organized data into roles and columns called tables and they use foreign keys to join them together. All right SQL queries who drawing different tables together? Graph database took a very different approach, but the more intuitive approach basically say the income was Data should be considered in network graph.

In the node on a graph is entity could be a person and then relationship is an edge. Right so you can actually visually think all how data are connected to each other. So that's how graph data.

Is it based bring that kind of data model graph very intuitive graph model install data that way inside database not like a tables not using drawings. So in this way graph Elevate the relationship between data as a first class citizen. So what are the use cases when you do that you can focus on what we call deep relationship analysis.

You can Traverse the relationship. This is the social network how inflation is particular person how Alan and ji connected together how many different path and those things can be easily analyzed if you organize data in the graphic way? Oh most popular use case the best use casing production is fraud detection.

Know fraud is about traces connectivity how Foster their transaction relate to other ones for detect for all Rings detect illegal money transfer. These are all interconnectedness among the data and graph is best suited for that. So that's actually what graph database is really about.

All right, Jake. First of all, thank you, right we got I think we laid a really nice. Foundation here next I want to introduce our audience to this concept of it's open Cypher.

Yeah. Yes. Yeah, so open sci-fi.

So it is like any database you need a query language so graph database have because it's different than relational database. We can all use SQL we have to use our own graph languages. So there's a bunch of data about database product in production today two of the biggest one on neo4j in Tiger graph.

So open Cipher is actually a open source graph query language first created by neo4j and in the attempt is to make it very simple for anybody to start exploring graph. So it's been created about more than five years ago. There's a lot of people when they first start learning about graph.

They actually use tooling from neo4j or other vendors to use this open source graph query language is designed to be really simple way intuitive. So that's how people use that as a beginner to learn about graph database. So that's what's open software is about Got it, and it's open sight for?

So it's an open source project. Obviously who manages that. Yeah, it's a Consortium.

It's a there's actually a bunch of event vendors but neo4j is the major event behind it because open Cypher is a model closely from their query language and that's actually one of the reason we want to be part of it to bring all database technology to this many a big open source Community to to leverage all scalability and performance from Tiger graph. Got it. All right.

So let's now I think we've done a great job of it's kind of like playing chess, right? We got all our pieces now lined up. Let's talk about tiger graph and open Cypher and recent announcement.

Yes. So before we talk about announcement this another piece of puzzle. I like to share with everybody sure actually part of the whole activity is wrong.

When craft database becomes mature, there's actually two indicator the new graph database technology becomes mainstream. The first indicator. Is there a unified query language is a standard just like a sequel for relational database.

Second. What is is there a database benchmark? Customer can use to compare between different vendors.

Both things are happening. So graph database now is entering mainstream. Now, there's a effort called gql.

It's actually sister organization to SQL right after 50 years of SQL people realize graph is very different. We need a standardized language. That's great.

Guess what all the major vendors are working on that. I I so standard called gql. It's designed to combine the best form all the query languages.

So they're both from neo4j and open Cypher as well as found tigerra. So the language itself is a combination of all the great features together. So this is title why tiger graph want to support open site for?

Because we are on a path towards the gkeo the gql standard will be finalized in 2024. But before that's happening, we want to introduce a flavor that early in all language. Guess what the unification between tiger graphs gsql, which is our query language.

We call graph SQL with open Cypher. The combination of both will get us closer to gql. So that's actually a strategic move for us to get you know, get on board get on board our customer earlier to the towards the standard instead of have to wait until 2024 now, I love it.

Part of that is now let's talk about what's the difference between gsql graph SQL language versus open Cypher. They are similarating nature meaning. They actually focus on we don't do relational table drawings.

We allow people to actually specify which graph path you actually use for example from a person they knows another person and then they may actually connect them from through Professional Network. We allow you to specify that grass growth traversal path instead of doing drawings. Both of us both open sign for gcco actually.

Spent a lot of people to do that. Unfortunately use a slightly different syntax. That's what we're gonna do is when we say we open open support open Cypher in all gcco language.

We're going to unify them the way how people can specify the path is exactly how they will specify in open Cypher. Right with that then we can bring a large open source community of more than 200,000 developers. They can easily get on use tiger graph without going through a steep learning curve to learn a different syntax to learn different language.

So all approach is a following in our next version of G sequel language. We gonna have support open Cypher syntax that means in the middle of the query language. You can reuse familiar patterns.

You already learned from open Cypher. Okay, but itself is not directly say bring open Sacred query directly run on top of us because we G SQL one of the key differentiation between us and open Cypher is we add a lot more capability. So for example, one of the key reasons people use graph database is to execute graph algorithms.

Those are actually being database machine learning. For example, one of the famous algorithm is pagerank. That's actually invented by Google.

Based on the interconnectivity of reference of web pages try to calculate the authority of authentication or the most influential page. That is actually craft traversal problem that algorithm can be coded in all G sequel language was only less than 20 lines of a code. Now open Cipher can do part of it the graph traversal calculation part but cannot do this, you know iteration through the entire graph.

This is where you know G SQL is more advanced. That's why what we decided is we want to retain these SQL Advanced capabilities, but in the middle, whatever the overlap we have we're gonna support openside for syntax. So in some ways we want people can bring their open suffer stuff over to Tiger growth, but all so allow them to do more advanced beyond what they can do in open Cypher.

So that's all strategy but to facilitate that we actually release a tool called open side for playground. All web page is openly accessible. com, you can play with that tool with that too is doing is actually give you a translator.

You can just write open Cypher query as this we translate into the advanced version of G SQL query you can see part of it is actually retain the open cycle syntax, but also allow you the ability to do law more then what's in the open Cypher. So that's what we're doing in terms of supporting open Cypher in Tiger graph. Jay defense.

I feel like we just hey learning with Jay. It's a new TV show where you got here on Tech stroke TV. That was pretty amazing man.

I want to thank you for that. So it was open cypherd that tiger graph. That was that it.

com. If you go to that website, you can experiment experience what that look like. We have a sample database called movies database a lot of people learning open sign for that's one of the first Totoro they learn over there.

We have a list of already pre-built query people can select from to see the open Cypher Singh text, but who is a click of a button they can see how it's translating the G sequel query with openstack racing texting the middle and then they can run this query against all tiger graph to see the result. So we like people to play with it and give us feedback again. Well, we actually release so far is a small subset well, Mentally building on more but that's one way we would like to get feedback from developers to say tell us which feature is more important.

I like to see this to be added next all the features you have here may not be exactly right. I like to provide my feedback. That's all way to reach out to develop a community early.

To bring them as part of the process for us to build up open Cypher inside tiger graph. Excellent. com slash playground.

That's great. That's great. And that that's where we'll get them.

We'll try to put that in the notes as well. Hey Jay, this has been fantastic. Thank you for coming on Tech strong TV and number one educating us.

Right? We have a tech audience, but they always want to learn more they want to understand more. This was a great a great interview for that best of luck at Tiger graphing come back and keep us posted.

Yeah. Thank you. I just want to add one more thing sure tiger graph really is known for scalability.

First time. We're able to crack 36 terabytes version of that ldbc Benchmark while other vendors struggle at the one terabytes and then within two years we want to reach to petabyte was over. Offer we believe law more developers can benefit from our scalable platform.

Please. Try it out. Thank you.

Absolutely. Take care doctor j u from Tiger graph here on techstrong TV. We'll take a break.

We'll be right back. Thank you.