Hydrolix on the Rise: Funding, Growth and Partnerships with Marty Kagan

Hey everyone. Welcome back here to Techstrong tv. You know, I always enjoy having a first time guest on our show and introducing you all to a new company.

So I'd like to introduce you to Marty Kagan. Marty is the co-founder and CEO of a company called Hydraulics. I hope I got that right.

Hey, Marty, welcome to Tech Drunk tv. It's great to have you on here. Thank you for having me.

Appreciate it. Okay, Marty, we're gonna talk about hydraulics a lot, but let's first talk about you, Marty. You're the, uh, co-founder and CEO I'm, you know, I'm sure there was a lot of thought consternation, you know, everyone who started, who's ever started a company has been through this.

You know, you gotta, they say you gotta be a little bit crazy. But how crazy is the question. But give us an idea of kind of your road leading up to co-founding, uh, hydraulics here.

Uh, my background is mostly in the content delivery space. I worked at Cisco and Akamai and Fastly running, uh, big engineering teams there. But I started a company prior to this called Sedexis that we ended up selling to your neighbors, uh, Citrix down in Florida.

Sure. I remember the challenge we, the challenge we had with Sodexos, where we were just generating huge amounts of data, about 23 billion records a day, and it was like rating a blank check to Google every month to deal with the data. And at the time of the acquisition, my co-founder Hassan said, I think we're doing this wrong.

I think everyone's doing this wrong. I think there's a better way to deal with this kind of volume of data. That there was a, there was a hole in the market when it comes to that kinda scale.

And he had a bunch of ideas that he thought, well, you know, you're probably wrong 'cause it's a big crowded space. But on the off chance that you've sort of looked at the problem differently, it's a big market to disrupt. And so we decided to give it a shot.

Absolutely. And, and you know what, that, that is a big market. I mean, so my background's in security cyber, and you know, one of the things we've had for as long as I've been in security is, you know, the more we wanna track everything, right?

We want logs for everything. But you know, the average security incident, you may not discover it until nine to 18 months after it took place. And so you need to keep records.

I Think there's, there's some sense, I think there's some myths in the industry around that, that, uh, the organizations want less data. What in fact, we really think they want more. I think that there's, you can't treat Lossless transaction logs the same way you treat Lossy application logs where there's noise and redundancy.

And I think that the hot, warm, cold strategies that many companies have adopted were really a trade off rather than optimization that the problem was that you've got finance putting pressure on it to say you gotta cut cost, you gotta contain cost. And the same thing, you've got the product engineering ai, ML data scientists saying, we need more data, more data. And that tension between managing your costs and making everything available creates an opportunity to change the economics and solve some interesting problems.

Absolutely. What we found there, there were lots of, of, of, uh, great modern data stacks like Snowflake and Databricks that do amazing jobs of decoupling compute from storage, but they're not really good at driving interactive dashboards. That's not really what they're built for.

And so you've got a knock and a soc and these other use groups that want to be able to interact with the data. They were sort of, they were often reliant on things like Elasticsearch for only the most recent, you know, 24 hours of data and then everything else ends up in cold storage. And that creates problems that you don't have access to all your data.

Um, so that was sort of the problem we wanted to go after. Uh, it makes perfect Marty. I I live it every day.

You know, I, because I I, I think you're right. First of all, geez, I remember it's gotta be around 1992 or 93, maybe 94. I bought a, it was a hard drive.

It was gigantic. I think it was 40 megabytes. And it was a hard, it was literally a hard box.

It was about this big, you know, remember those days and spinning, Spinning media with you Uhhuh and uh, and I said, my God, this is more memory than I'll ever, ever need. What am I going to do with all of this? Right?

But the very na for what there's, there's some sort of law at play. The more storage you have available to you, the more things you find that you need to store, right? And, but then we, there was this, Well, cost is, cost is a big driver.

Affordability is a big driver. So you say, well, Absolutely. And, and data anxiety and data grit are very real.

So you say, well, I, I wish I could keep everything. I wish I could index everything, but I, I just can't afford it. So we'll just index these columns and we'll just keep this data.

We'll just, we'll, we'll roll it up into our summaries. We'll throw away the raw data 'cause we just can't afford. But really, I don't, but then later I'm gonna regret like, oh, I wish I'd kept that.

I wish I'd indexed that. And so that, that we felt if we could just make it easier for everyone to store everything and keep everything and index everything, then you would eliminate all that anxiety regret. Like, you just wanted to worry about it.

And Absolutely. But there was another thing at play Marty, where people said, you know what? Not all data is created equal.

I am going to keep this in, as you mentioned, cold storage. Maybe I'm gonna put this on a tape. Maybe I'm going to keep this in S3, but I'm gonna keep this somewhere else, right?

Because I need this close by, I need this fast. And I don't really go into that as much. And so that was one of Go ahead.

That's, that's because there's this trade off between fast and hot right? And, and expensive and cold and slow and cheap. And if you can eliminate the difference, then why would you not have everything hot?

Why, if you're gonna store that stuff anyway, like regulatory reasons or whatever, if you're gonna have all this data, you might as well make it usable. And, uh, and it's only the cost that is driving you to say, well, let's just putting home storage. Uh, and we found as we open it up, people say, if all your cold storage is all of a sudden you have subsecond query performance against all of your data going back three years, it creates all new business opportunities that were sort of limited that you couldn't get to before because you've been forced to stick it into somewhere that was, you know, Absolutely.

And, and then when you went to retrieve it from the cold blocker, you know, you, you, my God, no one told me it was gonna take this long and it was gonna be this hard Or this expensive. The other thing was people were finding actually getting data back out can actually be really expensive. Yeah.

Yep. So talk to me about how hydro Hydraulics, hydraulics, excuse me, how hydraulics helped solve this. Uh, you know, we felt there really, there was nothing else really built for that kind of one to a terabyte to a petabyte a day scale.

There were lots of solutions where you're talking millions of transactions, but when you're talking billions, you need a different architecture. You need something that, uh, is built on stateless auto scaling. So you can handle DDoS attacks, you things like ETL, traditional ETL, where you can extract all the data, write it back in, that works fine at the million events scale.

It doesn't work well when you're talking hundreds of billions of events. Um, to have like multi-stage pipelines to manipulate your data before it goes into the data lake can all of a sudden cost just as much as the data lake. So you need a fully integrated solution, and you need something that is Kubernetes native, that is designed to, to scale up and down automatically.

Because as much as you want to be able to handle the Super Bowl five minutes after to the Super Bowl, you don't wanna be a database that big. You, you wanna back down to the scale you need. And so building something that was stateless and scaling, um, you know, was a, a, a unique approach to the market.

And Hassan, my co-founder, felt that we could get SSD like performance directly off object storage if we change the way we do indexing and compression and IO and kind of rebuilt the, the underlying, um, uh, storage layer, um, to, to optimize for, for object storage. And we've been able to successfully get kinda that hot data performance off of cold storage. Excellent.

The, the other thing we found is sort of what do you do with this? Now you got a data lake who, you know, who needs a data lake? What are the problems you could solve?

Uh, we felt that there were lots of good solutions for internal observability, but a big gap in the market around SaaS observability, especially for high volume, high value use cases like CDN data, security data, ad tech data where you're talking, again, hundreds of terabytes a day of information. Um, that wasn't, um, it was a different set of problems. Uh, and then lastly, we, um, we felt that most of the tools out there came with their own dashboards and their own query languages.

And that created a lot of sales friction because end users don't actually want new tools. They wanna use their existing tools. So we took a kind of a counterintuitive approach and built a logging platform with no front end, or rather every front end we embrace and extend things like Grafana and Cabana and Splunk and Databricks and Fabric and Click House and Spark and MySEQ and ProQ on my engineering team is working very hard to build more and more, um, uh, interoperability with existing tools so that we can make Databricks faster and cheaper.

We can make Splunk faster and cheaper. We can make click house faster and cheaper rather than having our own, our own front end. That, that has really helped accelerate our business.

We think, Oh, I would imagine it was. 'cause I mean, you look at a, a Splunk for instance, right? A lot of companies are very relying on Splunk.

They love Splunk, it's a great tool, but that, that they're often like, I think, can't Afford it's Prohibitive, right? And so you say, well only put this data in Splunk versus saying, I'm gonna use Splunk for everything, but I'm just gonna leave the data. I'm gonna land the data hydraulics that hydraulics handle the ingestion, the transformation, the enrichment, the indexing compression, and then use Splunk as my front end.

But I've got another team that wants to use Kibana. Another team wants to use Grafana, another team wants to use Databricks, and they can all operate off that one copy of the data in hydraulics, right? And have it to copy the data to six different databases.

There's a huge consolidation benefit in, in having a single platform to service these different needs. Wa I gotta tell you the truth, as someone who's lived this, this sounds almost too good to be true. We, we get a lot of that.

What's the catch? We, what's the catch? Well, so, you know, there's trade offs.

We don't support updates. We're not an asset compliant data warehouse. We're not built for data that is, uh, uh, we're not built for unstructured data.

You we're, we're very focused on particular kind of workloads and use cases, time series, structured logs that are immutable. That's kind of our sweet spot. Um, you know, and, and, and again, scale is really the key there, like the, the folks who are doing this at, at scale.

So we had lot of success. We've, we've signed nearly 500 customers in the last two years. Um, we, uh, we just announced a big, uh, you know, it's hard sometimes to get customers to be referenceable, but, um, we, we did just, uh, publish a blog about what we did with Fox, it, the Super Bowl.

We were ingesting a terabyte a minute during the game for them, um, handling, you know, millions of events per second, um, able to deliver five second, uh, time to glass. And, and moreover also ingesting many different data feeds, different sources, logs, fast logs, uh, many different CDM providers. They're, you know, other, other data they have from their ecosystem.

Client side data, bring it all together in a single pane of glass, uh, with, you know, five second ingest latency and subsecond query latency. Um, you know, something that those are easy things to do at, at the low end. It's very hard to do with that kind of volume at scale.

Um, so it was very, very successful outcome for us. Good for you guys. You guys also announced a, a really large, especially in today's world, right?

A really large series C. We closed our, uh, $80 million series CA couple days before the market, uh, crashed. Um, we're very grateful for the timing there.

Um, and timing Is everything We're, yeah, we had a lot last year. We grew eightfold. I think the investors were excited about the growth and, and they saw the overall, uh, market opportunity, um, to, to go after.

So very Excited. Who, who are the investors, if you don't mind sharing? Uh, found out of New York called QED investors.

Uh, they're, um, the founder of QED was actually the guy who founded, um, capital One. So very strong presence in FinTech and, and financial services. And, um, Nigel was excited about Capital One was really built on a thesis around data.

At, at one point, I think they were the largest Oracle deployment in the world. And so, uh, they really saw the opportunity for us to go and to break into the financial services market. There's a lot of problems there.

They're very data centric and, and disruptive technology like ours that can have a lot of, a lot of, um, uh, opportunity. Love it. Congratulations.

I mean, in today's world, that is a very, very healthy round. So congratulations. You guys also nabbed, oh, no pun intended.

NAB best of show award. Yeah. That we had, we had, we had a, a, we had, we had a booth.

We did some events. Uh, we had I think a really enthusiastic, uh, we had about 20 employees there. And, uh, I think, yeah, it was a, it was a nice, nice recognition.

Uh, we sell primarily today through partners, with partners, uh, that's most, most of our sales is through partners like Akamai and Amazon. We're working on, um, a number of additional partner, uh, integrations this year. And so we had a, uh, did a lot, uh, with, with Akamai, with Amazon at the event.

Um, that, you know, it was a, it was a, it was a great opportunity to get there and meet, meet with customers, meet with partners, meet with prospective new partners. Uh, it was, uh, you know, I didn't love spending five days in Vegas, but, um, it was, uh, it was a good, good show for us. I don't like spending five days in Vegas, either my friend three days and I'm done.

I really need to get home. Um, but that, congratulations on that again. So, so things are, you know, the arrow's pointing up as they say, right?

Yes. What, What else? Yeah, things we expect To see.

Things are, things are really busy. We, we, you know, we've expand. Last year we grew the team.

Uh, we started out last year with 40 people, wrote 200 people now. So we've done a lot, a lot to scale up, to handle, you know, we also grew from two customers to 500 in, in, in less than two years. So we are, we've been in this sort of keep up, uh, you know, chasing, uh, the opportunity and making sure that we're not getting in the way, uh, uh, of the opportunity.

And so, uh, for us, it's, now that we've kind of reached the scale we think we need is about execution. We've, we've, uh, we've built out a team now in LA in Latin America, where we do about 10% of our revenue. We've expanded, expanded our team in Europe, but we're only doing 15%.

But we think we can really grow there. Uh, we've got, we do about a quarter of our business in across a PJ and we've got new sales presence in India and Japan. Um, and we do about half in the US and we continue to expand our presence here as well.

Fantastic. You know what, we didn't even mention the website. com.

And if you, if you're in the Amazon marketplace, excuse Me, dot io right? io. Yeah.

Right. If you're in the Amazon market marketplace, a WS marketplace, and you, you search for hydraulics, you can, you gimme all of our, our, uh, services available, uh, for companies to, uh, to sign up for and do a free trial and, and check it out. I love it.

Um, what's next for hydraulics Marty? We continue to look at new verticals and new new use cases for our technology. So we're very strong today in things like media and ad tech.

Um, and with our new funding partners, we're looking now expanding into financial services. We're building some new serv new tools around, uh, advertising ad tech, uh, data for publishers. So we we're very vertical focused and very partner focused.

So I said the OC ine Amazon Drive most of our business today, but we think that there's, um, similar partnership opportunities. Again, it's, it's all around these SaaS services and building observability for these services. So if you're an Akamai customer, you need dashboards for all the Akamai products you're buying the Amazon customer needs dashboards for those products.

Um, we are looking at, um, replicating that success, uh, across, across the industry. I love it. Well, Marty, I wish you a lot of success, continued success here with hydraulics.

What a great story this is. This is really, I could see why the success. This is a real problem and it's been a real problem.

So let's You know. Well, thank you Alan. Thanks for your time.

It was lovely to meet you. Thank you. Marty Kagan, co-founder and CEO of hydraulics here on Techstrong tv.

Go check it out. io. I know you've been fretting over how much your storage costs and what are you storing the right things?

Are you throwing away things you need to be storing? Hydraulics can help you there. We're gonna take a break here on Techstrong TV will be back in a moment.