Exploring Open Source Database Services with Percona’s Bennie Grant

Hey everyone. Welcome back here to Tech Drunk tv. My next guest is Benny Grant.

Benny is the COO acting CEO of Percona. Benny, welcome to Tech Drunk tv. It's great to have you on.

Thank you. Nice to see you. Nice to see you.

Hey, Benny, why don't we start a little bit about you. You got, you're wearing a couple of hats right now, but you know, how did, how'd you get to the scene here to wear the hats? Absolutely, yes.

So you will hear from my accent that I'm originally from the uk, I've been in the US for ION 20 years, uh, and I've been with poona for around about eight years now. And so for the last three or four years I've been in the COO position, responsible for all service delivery, post-sale, and various operational things. And in the last few months as well, I'm acting as the CEO during an internal transition that we're running.

Excellent. What about before Pcoa in the, in the IT industry? Abso Yeah, absolutely.

So I kind of grew up as a deep technical, uh, engineer originally in the uk in the early days of the voiceover ip, unified communications networking space. That moved me over to the US in 2008 to kind of with one of the owners of that business at the time, to set up a US operation. And ultimately that business was exited in 2010, at which point I was part of a largest software organization that we all focused on, again, telephony, contact centers, that kind of stuff.

And it moved over to open source with poona in 28 years ago, whatever that might be. 2017? Yeah, 2017.

That's the fact. Yeah. That's, it'll be a second.

Yeah. Very cool. Mm-hmm.

So, Benny Perona Perona is a company I think a lot, you know, our audience is cybersecurity, cloud, cloud native DevOps, platform engineering, digital transformation. Perona is a company, I think that is, at least we've heard the name, maybe a Cube Con or or what have you. But I don't know if everyone truly understands what Perona is, what it does, what it's mission is.

Give a, make us smart. Benny, what, what is Preco about? Well, I'll do my best.

Uh, you know, it's a, we are an open source database services and software organization. We focus on three or four primary databases for us, which is MySQL, Postgres, MongoDB Technology, and nowadays of Val Key slash Redis, which is a new, new emerging technology. And fundamentally, our focus is all around enabling our customers and their customers.

What we, we refer to our customer's customer on all, all the data management stuff, high performance, security, scalability and things like that. And yeah, as a go to business, uh, go to market, the way we do business is we are predominantly on the software, apologies on the services side of the house. So all of our software is underpinning a lot of the largest organizations in the world.

Our software is free and available for anyone to use. Um, it's a drop in replacement for most of those technologies that I mentioned before. And we see a lot of adoption with that.

You are not obliged to pay per Kona to use our technology. And that's one of the things that makes us very unique. So we, we believe our mission is an open world is a better world.

It is completely open source. We fundamentally believe that that's in our DNA. Um, and this year, Piona just turned 19, so we're 19 years old running and operating this way and growing.

And congratulations. Very, very happy. Thank you very much.

So we're very proud of that milestone. Um, in the uk. We're old enough to drink in the US will be old enough to drink in a couple more years.

Um, it's, uh, it's something that every single person at Perona is extremely proud to have achieved. And, and you should be. That's great.

So when, when you say, you know, so the software is free, anybody could use the software. Yeah. Use the software.

You want to use Perona for services. Mm-hmm. Are those like human delivered services?

They're software as a service services software as we're hearing now as well. Um, what, how would you describe that? I, It's a bit of everything.

We, we have three primary lines of business. Um, our largest business is what we call support. And so we are there serving our customers and typically our customers that we are working with, we're servicing their DevOps team, their database infrastructure teams, their SRE organizations and stuff like that.

So if you look at our, the profile of our customer bases, you know, we go from small to medium all the way through to the, you know, fortune 50 customers, the level of expertise that we have, we do view ourselves at that highest level, you know, so we are helping with the most challenging problems for a lot of our customers. A lot of our customers have got a lot of internal expertise. We are there to back them up.

So it is a bit more on the people centric side there within dealing with some, you know, hard problems and significant scale and performance related thing. Um, within, we then offer a managed service where we'll actually look after the customer's environment for you and do it that there is a lot more, uh, software driven. So you have internal tooling and platforms, and we've got some intelligence built into that now, artificial or otherwise, um, that is enabling that, but obviously still looked at and maintained by our experts internally.

And we, we underpin all of it with our, our professional services, our consulting expertise, which the name KOA kind of started 19 years ago with performance consulting. That was kind of the pedigree was consulting. Oh, okay.

And that's still where we, you know, we, we show off. We, we we're very proud of our expertise. Um, a lot of our consultants are, you know, the best of the best of me.

I'm not ashamed to say that. Very cool. Yeah.

Very cool. Um, you mentioned intelligence, artificial or otherwise, right? Yeah.

You can't have a discussion in it today, tech without talking ai, it seems. Um, but you know, I I just before doing this interview with you, I was on the set of our Textron gang show when we were talking about, you know, ai, all the great things, whether it's generative or agentic or, or what have you, any flavor you want, all the great things AI is doing and what jobs it can do, what, how can it help other jobs, et cetera. But one of the things that we arrived at in that gang discussion is it doesn't exist in a vacuum.

You know, it's only as good as the information it has. It's only as good as the foundation it's built on. It's only as secure as, as it's been designed, allowed to be.

Yeah. Designed to be. Mm-hmm.

And a lot of the problem is if you, if it's pulling information from a design that wasn't secure to begin with, don't be surprised if it's Yeah. Right. It's not secure.

Mm-hmm. Um, same thing goes for data, data infrastructure. Right.

Uh, talk to us about that a little bit, if you don't mind. Absolutely. And you know, it's, it's a lot of the stuff that we're helping customers with at the moment.

So there's certainly two, uh, parallel tracks we're seeing with a, with a lot of this stuff in sofar. As to your point, if you've built your architecture's database infrastructure as well as the rest of your infrastructure, transparent by design, secure by design, ai ready by design, if you build it that way, you address a lot of the issues that you have up upstream. So when it comes to trust and performance and the whole garbage in, garbage out, the reality is the AI mo world is moving so quickly.

Most folks are backing into existing infrastructure right now. So you already have data in fragmented places and silos and things like that, that have been around for years, and now folks are trying to put it in, put it all together quickly, kind of put a model on top of it and just kind of learn as you go. And that there is where you see, and certainly where we are seeing customers, you, you, you lose trust, you know, you kind of get the whole hallucination thing or, and there security concerns because it just wasn't designed that way.

So how do you redesign re-architect without stopping the business for the next six months to do so? And, you know, keep things moving in parallel. So it's certainly some of the challenges we are seeing in the organization within our customer base.

And as you say, there's, it's all about trust and trusting that data and can I trust the results it's given me? Am I confident that it's secure? And we, yeah, we have a saying of you trust but verify.

I'm sure you've heard the trust but verify and, you know, the cynics will say, if I have to trust but verify my AI stuff, it'd be quicker for me not to bother with it in the first place and just do what I've always done. And so it's, it is kind of that mindset shift of, well, this isn't time investment. It's a bit like the cloud was 15, 20 years ago, you know, this is a moment where it is worth investing, but you need to sort of get that mindset shift and, you know, work with teams on that.

Mm-hmm. Um, of course they, the, the question is, and, and you know, part of the reason is silos. How do we know the data infrastructure that our AI is operating on is secure, is trustworthy?

Is is correct? What is the Yeah, well, well, that's the, that's the thing, right? So the, the, the security stability, you know, again, it's sort of transparent by design.

Secure by design. You, you can architect that to a certain degree, but you still have to test it. You still have to make sure that what we think we've done is actually doing it.

You know, it's the, you know, the, the analogy is that you, people do backups every day of the week. They backup up all their data, never test a restore, and then one time it needs to do a restore. Guess what?

The backups are never working and the work it's, and so it's kind of the same thing. It's, you need to trust this stuff. You can't just take our word for it.

You need to test it and everything else. Um, and there's ways of doing that. There's tools out there, um, depending on the size of organization, they'll have teams dedicated to it.

There'll be some, you know, regulator regulations that will mean that you need to outsource that, you know, independent testing, certainly on the security side of it. Um, but then there's the trust of it. It can be a very, very secure and it's transparent and it's great, but is it accurate?

And so then again, that's the whole, the other side to this of making sure that we've built this the right way. And how do you test that? There is still, I mean, we're playing with this internally, there is still a bit of what I call eyeballing it.

You know, you just look at it, there's, there's automations, there's tests, there's other things you can do, but fundamentally there is someone smart at some point or do some spot checking and just say, well, do I, does this make sense? Um, and I've found that we've been picking the cynics to do that. The, the, the folks that are more cynical and AI as a thing, as a, as a trusted source.

And they're the ones that are obviously the best testers for it. And so you kind of have internal review and QA and stuff like that. Absolutely.

Mm-hmm. Um, so again, I'll go back to a conversation I was having over on the other set. O one of the things I think we're seeing is sort of reallocating budget from services, from people services to ai, right?

We don't need a person to do this. The AI could do it. And generally it's the, the, the, the, the less sophisticated kind of test, the junior developer, the junior security person, the intern, right?

Those kinds of data ga gathering and stuff like that seem to be the, the, the jobs, if you will, that AI is, is performing. Mm-hmm. I'm wondering if you've seen this at Preco, at Benny where people are saying, look, I I just want your people to do like this really high level stuff.

Sure. But you know, we'll, we'll feed it up to you. We'll, the AI will feed up to you the data you need for that.

So yes and no, actually, it's really interesting for us, I think the, our customers are still very sensitive, rightly so on where their data goes. So I'll, I'll talk about both what we're doing internally with in Piona and also the service. We could live into our customers, but you know, obviously we have a lot of commitments and process we make to our customers around compliance.

You know, what customer data they give us, sometimes that data they're giving us, they need to make sure it doesn't include their customer's data again. And you've got the HIPAA and GDPR and all the stuff that goes with that. So we have and do talk with customers about automation of kind of agents, primarily using some agents to, what would you automatically share, how do you wanna play it?

How do you wanna do it? We have the facilities to do some of that stuff. However, our customers are very, very cautious.

You know, they are concerned about not giving us as a third, third party, we are a third party vendor. To them, they are conscious of, well, how much information are we giving this third party vendor? And still at the moment, there is still a, a check, but balance, a check and balance within our customer before it comes to us.

Most of the time that will change. I mean, there is no doubt in my mind that over time that trust will get there. There'll be automated sharing of certain things and redactions and stuff like that.

But at the moment, the trend we're seeing within our customer base is they still typically want a bit of human intervention before they share information with us. Um, and that's fine. We respect that.

And actually that's the same policy that we take with our vendors, you know, the same thing. We still just a little bit cautious of it internally when you're talking about how can you automate or streamline, make things more efficient. The term we use a lot, and I I I is one I believe in is the co-pilot.

You know, we, there is a fear across the technical community, this AI stuff is gonna take our jobs. And you know, the saying is, and I I do believe it won't take your job unless you are resistant to AI and partnering with it. I think if, sorry, go ahead.

Yeah, I I I've heard a different take on it, right? AI won't take your job. The next guy who knows how to use AI will take, will take your job.

Yeah, a hundred percent. Yeah. I think if you're gonna sit there and keep your head in the cloud or head in the stands, say, no, no, no, this is just a fad.

This will pass. We'll go back to the way we used to do things. I think you're kidding yourself.

I think you need to get with it, right? Um, and, and that's okay. You know, folks make the decisions that, that they will make.

It's a personal choice, but ultimately it's a technology you have to embrace. The next generation are gonna grow up or are growing up with this. It's, it's just a, a default.

It's just a standard. It's, it's not a weird thing for them. That's just all they're ever gonna know.

And if we're not there, somebody else will be. So yeah, we are certainly not looking at it. We are looking at it from a, an efficiency perspective.

How can we offer a better level of service? How can we enable more innovation quicker? Um, but we are looking at it in the context of a co-pilot, you know, as, as a, as a bootstrap self-funded company, which has been for 19 years.

We are conscious of making sure we are doing the right things with our resources. And we absolutely see AI as a way to kickstart us and get us further forward. Um, but yeah, if we can automate away some of the menial tasks, if we can script and use gen AI for certain things, we'll absolutely do that.

But we will not do that in the context of replacing staff. We're doing that in the way of freeing up bandwidth for that staff to do much more Interesting. All stuff.

Great. Yeah, agreed. Hey, Benny, we're almost outta time.

You know what? I realized we didn't even mention the Percona website. How do people like contacting can engage with Perona?

Yeah, absolutely. It's, thank you. com.

com. We are a global organization, so you can pretty much contact us 24 7 through the website. I love it.

Benny, I wish you continued success at Perona. You guys got 19 years in counting. You owe it on your 21st.

You owe me a drink. We'll, we'll, we'll toast through it. Uh, but until then, man, keep doing what you're doing.

Rock on. I, I appreciate it. Adam, thank you very much for your time.

Thank you. Appreciate it. Benny Grant, COO acting CEO Perona here on Tech Drunk tv.

We're gonna take a break. We'll be back in a moment.