top of page

simplyblock and Kubernetes

Simplyblock provides high-IOPS and low-latency Kubernetes persistent volumes for your demanding database and other stateful workloads.

Writer's pictureChris Engelbert

How Oracle transforms its operation into a cloud business with Gerald Venzl from Oracle (video + interview)

Updated: Jun 12

This interview is part of the Simplyblock's Cloud Commute Podcast, available on Youtube, Spotify, iTunes/Apple Podcasts, Pandora, Samsung Podcasts, and our show site.


In this installment of podcast, we're joined by Gerald Venzl (Twitter/X, Personal Blog), a Product Manager from Oracle Database, who talks about the shift of focus away from on-premise databases towards the cloud. It's a big change for a company like Oracle, but a necessary one. Learn more about the challenges and why Oracle believes multi-cloud is the future.



Chris Engelbert: Welcome back to the next episode of simpleblock's Cloud Commute podcast. Today I have a very special guest, like always. I mean, I never have non-special guests. But today he's very special because he's from a very different background. Gerald, welcome. And maybe you can introduce yourself. Who are you? And how did you happen to be here?


Gerald Venzl: Yeah, thank you very much, Chris. Well, how I really don't know, but I'm Gerald, I'm a database product manager for Oracle Database, working for Oracle a bit over 12 years now in California, originally from Austria. And yeah, kind of had an interesting path that set me into database product management. Essentially, I was a developer who developed a lot of PL/SQL alongside other programming languages, building ERP systems with databases in the background, the Oracle database. And eventually that's how I ended up in product management for Oracle.

The ‘how I'm here’, I think you found me. We had a fun conversation about 5 years ago, as we know, when we met first at a conference, as it so often happens. And you reached out and I think today is all about talking about Cloud Native, databases and everything else we can come up with.


Chris Engelbert: Exactly. Is it 5 years ago that we've seen last time or that we've seen at all?


Gerald Venzl: No, that we've met 5 years ago.


Chris Engelbert: Seriously?


Gerald Venzl: Yeah.


Chris Engelbert: Are you sure it wasn't JavaOne somewhere way before that?


Gerald Venzl: Well, we probably crossed paths, right? But I think it was the conference there where we both had the speaker dinner and got to exchange some, I mean, more than just like, "Hello, I'm so-and-so."


Chris Engelbert: All right, that's fair. Well, you said you're working for Oracle. I think Oracle doesn't really need any introduction. Probably everyone listening in knows what Oracle is. But maybe you said you're a product manager from the database department. So what is that like? I mean, it's special or it's different from the typical audience or from the typical guest that I have. So how is the life of a product manager?


Gerald Venzl: Yeah, so what I particularly like about Oracle product management, or especially in database, obviously different lines of business and then inside Oracle may operate differently. It's a job with a lot of facets. So the typical kind of product management job, the way how it was described to me was, well, you gather customer requirements, you bring it back to development, then it gets implemented, and then you kind of do go-to-market campaigns. So basically, you're responsible for the collateral, what's the message to advocate these new features to customers, to the world, and then that's not so true for Oracle. I think one of the things that really excites me in the database world, it's like this goes back to the late 70s. I mean, other than Larry, not that many people are around from the era anymore. But Oracle back then did a lot of things that were either before its time or when there simply was no other choice or way of doing it, commensurable wisdom, I would say. So one of the nice things in Oracle is that actually the coming up with new features is really a nice collaboration between development and product management.


So development just as much has their own ideas of what we need to do or should be doing, like the PMs, and we really get together and discuss it out. And of course, sometimes, there's features that you may or may not agree with personally or don't see the need for. And often, actually, and much more so, you get quite amazed by what we've come up with. And we have a lot of really smart people in the work. And one thing that, yeah, not to go too much into a rabbit hole, but a couple of things that I really like; believe it or not, database development, it feels a lot like a startup. There's no fixed hierarchies as such, ‘you can only do this. You must only do this or anything like that.’ You can very openly approach the development leads, so even up to the SVP levels. And actually, just as we started now, one of those guys was like, "Hey, let's talk while I'm driving into work." I was like “sorry, I'm busy right now”. So you have that going. And then also, there's a lot of the product management work that has a lot of facets to it. So it's not just ‘define the product’ or anything like that. That is obviously part of it, but also it's evangelizing, as I'm doing right now. I speak to people on a thought leadership front for data management, if you like, or how to organize data and so forth.


And as I said before, one other thing that I really enjoy working in a team is there's actually quite a lot of really smart people in the org that go back to the 90s and some of them even to the 80s. So I got one guy who can explain exactly how you would lay out some bytes on disk for fastest read, etc. Then this is stuff that I never really touched anymore in school. We were already too abstract. It's like, "Yeah, yeah, yeah, whatever. There's some disk and it stores some stuff." But you get still these low level guys and some of them, one of them is like, "Yeah, I helped on the C compiler back then with Kernighan." It's like there was one of the guys but was involved in it. And so anyway, as you know in the industry, people go around quite a bit. And so that has a lot going there.


Chris Engelbert: So from the other perspective, I mean, Oracle is known for big database servers. I think everyone remembers the database clusters. These days, it's mostly like SUN, SPARC, I guess. But there's also the Oracle Cloud and the database in the cloud. So how does that play into each other?


Gerald Venzl: Oh, yeah. Now things have changed drastically. I mean, traditionally starting a database software in the good old 80s where you didn't even have terminal server or whatever, a client server. So the first version is apparently a terminal based or something like that.


It's like, again, I never saw this. But there was a big client server push. And obviously now there's a big push into what's cloud and a lot of cloud means really distributed systems. And so how does it play into each other? So all the database cloud services in Oracle Cloud, all the Oracle database cloud services are owned by us in development as well.


So we have gone into this mode of building cloud services for Oracle database. And of course, that's really nice because that gives us this visibility to the requirements of distributed storage or distributed workloads and that in turn feeds back into the product. So for example, we are still one of the very few relational databases that offers you sharding on a relational model, which is, of course, much harder than a self-contained hierarchical model such as JSON, which you can shard way nicer. But once you actually split up your data across a bunch of different tables and have relations between those, sharding becomes quite more complicated.


And then of course, it's like we have a lot of database know-how. We also got MySQL, they do also their thing with good collaboration going on with them. So we have sort of quite a good, I want to say, brainpower, intellectual power in the company when it comes to accessing data and to writing data. You mentioned SPARC before. There's, of course, a lot of that going on. And quite frankly, I will say even way before cloud, the fact of accessing data that doesn't necessarily sit in a database but analyze it or query it with SQL. It's like you literally go back like 10, 12 years ago and everybody said Hadoop will kill every database and big data is the way forward. And I'm sure there was the same thing going on in the mid-2000s. I was not in the industry yet. So like, yeah, this notion of that you have data sitting somewhere else and you just want to analyze it has been around for a long time, actually much longer than people see now with object store buckets and data lakes and all the good stuff.


Chris Engelbert: So how does that look like for customers? I mean, I can see that smaller customers won't have an issue with the cloud, but I could imagine that banks or insurances or stuff like that may actually have that. What does the typical cloud customer for Oracle look like? I think it may be very different from a lot of other people using Postgres or something.


Gerald Venzl: Yeah. I mean, you kind of mentioned it before. I think there is, ‘are you small or are you large?’ Right. And the SMB, small, medium business customers, the smaller ones, obviously, they're very much attracted by cloud, the fact that they don't have to stand up servers and the data centers themselves to just get their product or their services to their customers. Big guys are much more like ‘consolidation’ and the biggest customers we work with, it's really like their data center costs are massive because they are massive data centers. So they are looking at a more of a cost saving exercise. Okay - if we can lift and shift this all to cloud, not only can I close down my data centers or large portion to it, but of course also most of them are actually re-leveraging their workforce. So people, especially the Ops guys are always very scared of cloud or often very scared of cloud that will take their job away. But actually most customers are just thinking ‘rather than looking after the servers running this good stuff, maybe in 2024, we can leverage your time for something that's more important to the business, more tangible to the business.’ So they're not necessarily looking so much to just get rid of that workforce, but transforming it to take care of other tasks.


A couple of years ago when we did a big push to cloud for Oracle Database and our premier database, Cloud Service Autonomous Database came out, there was quite a big push for the DBAs to transform into something more like a data governance person. So all the data privacy laws have crept in quite heavily in the last 5 to 10 years. I mean, they were always there, but with GDPR and all these sorts of laws, they are quite different in what they are asking from data privacy laws before. And this is getting more and more and more complex, quite frankly. So there was obviously a lot of aspects of, ‘hey, you are the guys who look after these databases storing these terabytes and terabytes of data.’ It's like, ‘now we have these regulatory requirements, where this needs to be stored, how this needs to be accessed, et cetera.’ And I might try to have you figure that out and figure out whether the backup was successfully taken or something like that. So you're looking at that angle.


But yeah, so the big guys, then they, I think to some extent also very quickly get concerned of whether data is stored public cloud or not. Oracle was actually, I want to say we were either the first or definitely a forerunner of what we called Cloud@Customer. So basically you can have an Oracle cloud at your site. So you reinstall Oracle cloud in your data center. So for those customers who say, "This data is really, really precious." You always have a spectrum. It's like there's a lot of data you don't care about, a lot of public data that you may or may not store, reference data and so forth, that you have to have for your operations. And then there's actually the really sensitive data, your customer confidential information and so forth. And so there's always a spectrum of stuff that ‘I don't care can move quicker to cloud’ or whatever. And then of course, the highly confidential data or competitive confidential data – ‘I really don't want anybody else to get a hold of this’ or ‘it's not allowed or regulatory.’


Those systems then they look into a similar model where they say ‘well, we like this sort of subscription-based model where we just pay a monthly or yearly fee per use and still all the automation is there. It's like we still don't have to have people looking whether the backup is successful or something. But we want it in our data center. We want to be full control. We want to be able to basically kind of pull out the cable if we have to and the data resides in our data center and you guys can no longer access it. Sort of that sense. I mean, that is obviously very extreme. And so this is what we call Cloud@Customer. You can have an Oracle cloud environment installed in your data center. Our guys will go in there and set everything up like it is in public cloud.


Chris Engelbert: That is interesting. I didn't know that thing existed.


Gerald Venzl: Yeah, it's actually gotten much bigger now. So just to finish up on that, it's like, so now we have these, I mean, even governments is this next level, right? So governments come back and they say, "We're not going to store our data in another country's data center." So this kind of exploded into like even what we call government regions. So, and there's some public references out there where some governments actually have a government region of Oracle cloud in their country.


Chris Engelbert: So it's interesting. I didn't know that that Oracle or Oracle Cloud@Customer existed. Is that probably how AWS handled all the like AWS or what is it called Oracle at AWS or something?


Gerald Venzl: No, so AWS is different. AWS came out with outposts, but that was actually years later and when you do your research, you see that Oracle had this way longer. But now I think every provider has some sort of like ‘Cloud@Customer’ derivative. But now AWS is Oracle databases and what they call RDS, the relational database services. But I think what you're thinking of is the Microsoft Azure partnership that we did.


So there's an Oracle database at Microsoft Azure. And even that has a precursor to it. So a couple of years ago, basically Microsoft and Oracle partnered up and put a fast interconnect between the two clouds so that you kind of don't go out of public net. But you could interconnect them from cloud data center to cloud data center, they were essentially co-located in the same kind of data center buildings. I mean, factories is really what they look like these days. So that’s how you got this fast interconnect, or kind of like buildings next to each other. And that was the beginning of the partnership. And yeah, by now it was a big announcement, you know, Satya Nadella and Larry Ellison were up in Redmond at Microsoft, I want to say was last fall, around September, something like that, but around the time where they had this joint announcement that yeah, you can have now Oracle database in Azure. But you know, the Oracle database happens to still run on Oracle cloud infrastructure. And why this fast connect is exposed via Azure.


Now the important thing is, all the billing, all the provisioning, all the connectivity, everything you do is going through Azure. So you actually don't have to know Oracle cloud, what effect that runs in Oracle cloud, that is all taken care of. And that caters to the customers, we have, you know, lots and lots and lots of customers who have applications that run on a Microsoft stack, rather than pick any Windows based application that are in Azure, it's a natural fit, what happens to have an Oracle database backend. And I think that in general is something that we see in the industry right now that these clouds in the beginning became this massive monolithic islands where you can go into the cloud and they provide you all these services, but it was very hard to actually talk to different services between clouds.


And our founder and CTO Larry Ellison thinks very highly of what he calls multi cloud or what we call multi cloud, you know, it's like you should not have to kind of put all your eggs in a basket. It's literally a kind of the good old story of vendor lock-in again, just in cloud world. So yeah, you should not have to have one cloud provider and that's only it. And even there, we have already seen government regulations that actually say you have to be able to run at least two clouds. So if one cloud provider goes out of business or down or whatever, you cannot completely go out of business either. I mean, it's unlikely, but you know how the government regulations happen, right?


Chris Engelbert: Right. So two very important questions. First, super, super important. How do I get an interconnect to Azure data centers to my home?


 Gerald Venzl: Yeah, that I don't know. They are really expensive. There are some big pipes.


Chris Engelbert: The other one, I mean, sure, that's a partnership between, you said Microsoft and Oracle, so maybe I was off, but are other cloud providers on the roadmap? Are there talks? If you can talk about that.


Gerald Venzl: Yeah. I mean, I'm too far away to know what exactly is happening. I do know for a fact that we get the question from customers as well all the time. And, you know, against common belief, I want to say, it's not so much us that isn't willing to play ball. It's more than the other cloud vendors. So, we are definitely interested in exposing our services, especially Oracle database services on other clouds and we actively pursue that. But yeah, it basically needs a big corporate partnership. There’s many people that look at that and want to have a say in that. But I hope that in some time we reach a point where all of these clouds perhaps become interconnected, or at least it's easier to exchange information. I mean, even this ingress/egress thing is already ridiculous, I find. So this was another thing that Oracle did from the very early days. It's like we didn't charge for egress, right? ‘If data goes out of your cloud, well, we don't charge you for it.’ And now you see other cloud vendors dropping their egress prices, either constantly going lower or dropping them altogether. But you know, customer demand will push it eventually, right?


 Chris Engelbert: Right. I think I think that is true. I mean, for a lot of bigger companies, it becomes very important to not be just on a single cloud provider, but to be just failure safe, fault tolerant, whatever you want to call it. And that means sometimes you actually have to go to separate clouds, but keeping state or synchronizing state between those clouds is, as you said, very, very expensive, or it gets very expensive very fast. Let's say it that way.

So because we're pretty much running out of time already, is there any secret on the roadmap you really want to share?


Gerald Venzl: Regarding cloud or in general? I mean, one thing that I should say, is like Oracle database, you know, a lot of people may say, ‘it's like, this is old, this is legacy, what can I do with it, etc.’ So that's all not true, right? We just kind of announced our vector support and got quite heavily involved with that lately. So that's new and exciting. And you will soon see new version of Oracle database, we announced this already at Cloud World, that has this vector support in it. So we're definitely top-notch there.


And the ‘how do I get started with Oracle database,’ this is also something that often people haven't looked for a long time anymore. So these days, you can get an Oracle database via Docker image, or you have also this new database variation called Oracle Database Free. So you can literally just Google ‘Oracle Database Free’, it's like a successor of the good old Express edition for those people who happen to have heard of that. But too many people didn't know that Oracle Database, there was a free variant of that. And so that's why we literally put it in the name, ‘Oracle Database Free.’

So that's your self-contained, free to use Oracle Database, you know, it has certain storage restrictions, basically, and then you kind of go too big as a database. And but the big item doesn't come with commercial support. So you can think a little bit of like in the open source world of Community Edition and Enterprise Edition. So you know, it's like, Oracle Database Free is the free thing that doesn't come with support, it's essentially restricts itself to a certain size. And it's really meant for you to tinker around, develop, run small apps on, etc. But yeah, just Google that or go to Oracle.com/database/free. You will find it there. And just give Oracle Database a go. I think you will find that we have kept up with the times. As mentioned before, you know, one of the very few relational databases that can shard on a relational model, not only on JSON or whatever. So certainly a lot of good things in there.


Chris Engelbert: Right. So, last question, what do you think is like the next big thing or the next cool thing, or even maybe it's already here?


Gerald Venzl: I mean, I'm looking at the whole AI thing that's obviously pushing heavily. And I’m like old enough to have seen some hype cycles, you know, kind of completely facepalm. And I'm still young enough to be very excited. So somewhere on the fence there to be like, AI could be the next big thing, or it could just, you know, kind of once everybody realizes…


Chris Engelbert: The next not-big-thing.


Gerald Venzl: Exactly. I think right now there's nothing else on the horizon. I mean, maybe there's always the always something coming. But I think everybody's so laser-focused on AI right now that we probably don't even care to look anywhere else. So we'll see how that goes. But yeah, I think there's something to it. We shall see.


Chris Engelbert: That's fair. I think that is probably true as well. I mean, I asked a question to everyone, and I always would have a hard time answering myself. So I'm asking all the people to get some good answer if somebody asks me that someday.


Gerald Venzl: Yes. Smart, actually.


Chris Engelbert: I know, I know. That's what I that's what I tried to be. I wanted to say that I am, but I'm not sure I'm actually smart. All right. That was a pleasure. It was nice. Thank you very much for being here. I hope to see you somewhere at a conference soon again.


Gerald Venzl: Yeah, thanks for having me. It was really fun.


Chris Engelbert: Oh no, my pleasure. And for the audience, hear you next week or you hear me next week. Next episode, next week. See you. Thanks.

Comments


bottom of page