Sovereign DBaaS Platforms – Vinay Joosery, Severalnines

This is texturing TV. Hey guys. Thanks for the throw.

We're here with. The name Juicery who is CEO for several nines and we're talking about the emergence of sovereign debas databases and service for those of you unfamiliar with the term. Hey vinay welcome to show.

Thank you. Nice to be here. Well, I guess the first question that everybody's mind would be well, just what the heck is this Sovereign debas?

So what do we need it? And how does it come about? Yeah, so well debas, right database Services has been around for you know more than a decade.

You know, they are very popular. Right all the hyperscale is provide them but what the Enterprises are telling us is you know, they are sort of limited right in a couple of ways a you know, organizations need to have you know, you know, they they need to have authority over the data, right? They need to comply with with you know, rules and having data sitting just in a hyperscala doesn't allow them to comply with these rules.

Right? Sometimes you need to have them on-prem. They need to have them in a specific region, right and and you know, so so, you know from that perspective they need another type of debas that allows them this this control over where data is located, right and a I think that's probably the biggest, you know thing that we you know sort of challenge that we see is to be addressed.

I just by this, you know Sovereign Divas is to give these, you know, these Enterprises the ability to choose where they have their data and have complete control over that data. What makes a sovereign debas different from a standard device and some people would say I could deploy a Divas on premise or there are certain attributes about that or we just describing more of the use case. I think yeah, so, you know the the problem is you know, who when you most of the common D masses they are hosted in the cloud, right?

And typically they'll they are from a hyper Scala right? So from that perspective when you store, you know, even if you store data on a hyper scale in Europe And you actually the data is subject to us laws. Not to you know EU privacy laws and that's the biggest issue.

Right? So and that's a problem when you have you know, gdpr data by gdpr compliant data and you're storing it in in the EU with for example, you know, you know, like like, you know, like like a hyperscala but then being being a us hyperscaler the data is shipped over to the US and suddenly it falls under the US jurisdiction and under the US Cloud Act Right, which means you know, some of these authorities could access the data without telling the data owner, right and that's that's a big problem for even you know for like, you know, the public sector in Europe and where you can do that right and you need to tell the authorities you need to tell, you know, the the, you know, you need to get a warrant from from the EU. Whereas if you're storing it out with a us hyperscala, you know a US warranties enough and you don't know Whether your rate is being accessed, right if you're in in Europe, so so having a sovereign debas gives you control over.

Where where do you put the data because some data is okay to be right with a hyperscaler but sensitive data and you know pii data right for you citizens, you might want to put that with you know with a with a you know company which is entirely based in Europe and entirely under the jurisdiction of of the European country, you know, where whether the data is generated. So will this be more of a Federated approach over time? We'll have a debas maybe on premise and they'll be the best in the cloud and maybe the same it may be different but I'm gonna have different data stored in different places based on its sensitivity.

Is that the goal? Correct? Correct?

I mean, you know, let's let's face it right today the Enterprise and it's not that small, you know, it's not just a couple of databases in a neat data center somewhere right Enterprise have been around for a long time. They've both companies, you know, they've they've started off with their own data centers. They've utilize the public Cloud, you know, they typically we see Enterprises using you know, at least two to three of the clouds minimum, right and so data is not tidy and neatly stored somewhere right?

It's just all over the place and the issue we've Enterprises is how do they can you know, how do they get control over all that complexity? Right and because at the end of the day if there is a data leak They will be responsible and they will be fine and they will be held accountable, right so they need to have control over all that data. So being able to do it in an easy way is key.

Right? So so from that perspective Sovereign debus, we're not mandating that all your data should be in your data center. Now, that's not that's not you know, that's not possible Right.

What we're saying is you can put data in your data centers. You can put data in two free public clouds, but you can have a single control plane right to manage all that, right? And as part of that it's also all the existing workloads running which you cannot forklift into your new shiny device, right?

Because many Enterprises they have Legacy stuff running and these also need to be managed. Right so so for us, you know, the concept of sovereign debas is to have this control plane that allows you, you know, that that sort of allows you to manage your existing data as well as data that you've decided to In you know on-prem or in public Cloud a public Cloud B, right and typically Enterprises. They have compliance teams making those decisions, right?

So and it's the business that decides where can this data go and then it's the role of the infrastructure to enable that business. I would argue that most organizations have not been especially good it managing data. A lot of them would not get a Good Housekeeping seal or approval for the way that data gets managed.

Do you think that these compliance regulations and data privacy regulations that we're starting to see you're forcing that issue in organizations are revisiting the way they think about data management and the light I'd say a data Privacy Law. I think so, you know, I mean, you know, and I don't know if you remember that about 10 years ago. People, you know suddenly people have deployed, you know researchers found 50,000 databases wide open on the internet, right and a lot of data could be leaked that way.

Right and we know now today this is not okay. So the gdpr says they can find you up to up to you know, 20 million euros of you, you know, or up to four percent of your entire revenue and your Revenue, right? And California is coming with their own privacy laws, right?

They've you know, they I think they're coming up with a new privacy low next year, right, which is very close to the gdpr. Right? So even in the US you have different states, you know coming up with privacy knows and so so and and the problem is, you know, if you don't treat your data carefully if it gets me somewhere, you know, this will this will cost you a lot.

Right? So from that perspective, you know and the prices they need to understand which data they have they need to have you know the you know this sort of you know, they have they have to have audited all the types of data. They need to know if it's critical if it's non-critical and they need to have Control, especially the critical data because and you know, we've already compliance like for example the gdpr you you can get fined and that will cost the money.

No, but you can cause quite a lot of money in the old days at least there was a database administrator who kind of took care of all this stuff but has that function changed as we move to this new era. Are we seeing? Yeah, we hear a lot about data apps we hear about Chief data officers.

I guess. I'm wondering who's managing the data? Who's mine in the store these days.

Yeah. I mean we still have dba's right but if you know, if you look at how databases are being deployed today and the majority of new databases are not running on the oracles, right or the Secret Service like the bulk of the new data goes into open source databases, right and they are very often. You don't have DVDs right full-time TV is dedicated for the inside and you could do that 20 years ago you would have A big DBS for Oracle for your SQL server, but today and you know, we have mostly devops people and managing this databases.

And the other thing is look at the application architecture right applications. Typically, they are shipping with free four different types of databases so you can have specialized dbas for each database time. So so, you know, we we have the emergence of like, you know, like like, you know, the devops crowd way up.

They also need to be able to manage their databases and that's why dbass is really nice because you don't need to get into the needy gritty of operating the you know, the actual database all that is automated right and but suddenly when you take it on Prem there is no there's no real debas on Prem. Right? So so that that is a bit of a problem which is why Enterprises they want the d best experience the convenience the automation, right?

So what we are seeing is you know with with the Sovereign debas concept, right and you have infrastructure to help you automate these databases so that all these devops who are having to juggle with multiple database Technologies. They are allowed, you know, they are capable suddenly to manage databases because you know because they have the tool set to do that. Are we drowning in the data that we're starting to collect the same like we're more data than ever.

There's more data types than ever. Can the databases handle it or do we need a lot of different types of databases or can I aggregate all that data in one Central database? What's the strategy?

Yeah, I think you know, you know you we see the the emergence of federation Technologies because trying to put everything in one box doesn't work. Right and the other thing is We've seen the polyglot persistence movement right where you where you actually have specialized data into specialized databases, right? And there's been a bit of a backlash where some of the vendors they've started doing multi model.

And you know databases. So for example, a relational database will be able to do you know key value data, they will be able to do documents stores. They you know, they can do multiple things in the same database so that you you know, so that so that you don't have to have multiple databases because it's more complex more database types.

But what we see is that it doesn't really work right the databases, although they can manage different types of data. They don't do it very well, you know many Enterprises today. If you look at surveys they are still running typically in you know, and Oracle in SQL Server.

They are running my sequel. They're running mongodb. They're doing ready with some you know, in memory data, you know might be doing elastic.

So so there's there's a whole variety of databases which are being used today. And and I think you know the the issue is how how do they get a grip on all these different data stores because now Also have the multi-cloud you have the hybrid experience where they have both data in different clouds and on-prem. So I think you know and federating access across these different sources and having a single while having some kind of easy control plane to keep track of all these databases and to operate them is something that we're you know, where we see the Enterprise is going and there hasn't you know, we have seen a lot of companies building their own databases for that right and everyone control planes and their own homemade debas right to try to operate databases in multiple locations and you know, multiple types of databases because that type of capability hasn't been on the market.

So several lines does what exactly are you providing the debas or do you have a dedicated database or do you let me choose which type of database it's gonna be? Yeah, so we've been we've been around for you know for about 10 years and we've been working on database automation from day one and we automate the most popular databases on the market. We don't do Oracle.

But we do SQL serve my sequel postgres mongodb redis elastic and so on right and we you know, we're kind of staying away from Oracle because typically people using Oracle they already have their own systems for that. Right? So so and so we automate databases and and you know what, you know, we've been doing we've been doing it for 10 years and Sovereign debas.

The name is is really what we see, you know people calling you today because you know people we've seen that and we've all these compliance issues. We've all these, you know, multi-cloud complexity and then database is running running on Prem. So so you have these hybrid environments we see the sovereign debas.

A as a proper name right for the for the for the concepts that we've been building for 10 years. So database automation is is the key thing that we've been building on helping Enterprises automate their databases and you know, wherever you know, wherever they want to automate them. So ask question.

Do you think that AI will help us manage data in the future because automation is part of the equation and everybody's talking about AI these days but how far can we go? I think you know definitely we we have a bit of AI ourselves right in our software. I and I think you know when you when you talk about, you know, managing large amounts of data a lot of things need to be automated right?

So so for example, we use machine learning algorithms to to do failover, for example, right because it database can fail in many different ways and we need to fix it. Whatever right? We need to replace the human.

Right instead of having a DBA who logs in understand what's going on? Look at the logs, you know, and actually, you know perform procedures. We automate all that so that you don't have to do it.

Right? No human will need to do that. So we automate a lot of things and I think I think yeah, I mean we see AI has a role in there.

Right and there are other things that AI is doing for example, automatic tuning of databases. There are some startups focusing on that right and you know oracle has done a lot of research even in IBM has done a lot of research in the last decades on on you know Auto tuning, you know, your workloads and and I think you know that that's one one area that probably needs to hit the open source database market right to having these and you know artificial intelligence ways of you know, automating performance improvements and tuning and but I think yeah, I think you know, we've we've done a lot around And you know keeping things up and running and doing failover, you know automating failover. But but yeah, there are there are still areas to be done so that you know at the end of the day you want to end up with the machine doing all the automation 100% of the database automation so that you don't need any human involvement.

All right. Well, you heard it here first guys never lose controlling your data and there's tools to make that happen. Meaning thanks for being on the show.

Yeah. Thanks, Mike. All right back to you guys in the studio.