Get Started Free
December 7, 2020 | Episode 132

Apache Kafka and Porsche: Fast Cars and Fast Data ft. Sridhar Mamella

  • Transcript
  • Notes

Tim Berglund (00:00):

Are you a car person? If you are, you can probably respect the notion of driving a Porsche and maybe even driving it on the Autobahn, where there are famously no speed limits apart from the normal relativistic ones. And hey, talking about speed, Sridhar Mamella has built out a standardized event streaming framework, organization-wide at Porsche. He's going to tell us all about it and the results he's gotten with it on today's episode of Streaming Audio, a podcast about Kafka, Confluent, and the cloud.

Tim Berglund (00:36):

Hello, and welcome to another episode of Streaming Audio. I am as ever, your host, Tim Berglund, and I'm joined in the virtual studio today from across the Atlantic Ocean by Sridhar Mamella. Sridhar is a platform manager for data streaming at Porsche. Sridhar, welcome to the show.

Sridhar Mamella (00:57):

Hi, thank you for having me.

Tim Berglund (00:59):

Hey, how did you come to work for Porsche? I know that cars are tremendous generators of streaming data, but I don't know that that's common knowledge. When people think of fast cars, they don't think of fast data. How did you get into this work?

Sridhar Mamella (01:17):

Honestly, it was a long journey. I was born in Bombay, did most of my education in London, and now I'm living in Stuttgart, which is like the heart of Porsche. Porsche was born in Stuttgart [inaudible 00:01:32] housing. I moved to Stuttgart for about three and a half years from London. Basically, I have a classic IT guy's education, with a bachelor's in computer engineering and a master's in Big Data, both from London at the University of Greenwich, and which wasn't long ago, I completed my masters in 2015. That's when I had this fancy masters with this fancy buzzword attached to it.

Tim Berglund (02:00):

Sure, [inaudible 00:02:00] Big Data that hasn't been around for too long.

Sridhar Mamella (02:03):

Exactly. I worked for a few years in London. Firstly, as an intern at the London Olympics, and then for projects for the London Underground. In all these positions, I did something with Big Data. As you can imagine, back then, with all the hype around Big Data, everybody wanted to jump on to the Big Data bandwagon. But with the small data.

Tim Berglund (02:26):

Postgres is not cool enough. We needed to do...

Sridhar Mamella (02:33):

Exactly we wanted to do PM, I don't know, get just kidding me Big Data. Then I thought, “Okay, let's do something crazy. Let's go to a new country, where people speak a different language, the culture is different. I lived in India, I've lived in London. Now, let's do something crazier.” I wanted to do it before I could really settle down in life. So [inaudible 00:02:58], “Okay, why not? Let's go to Germany. I am a petrol head. I love cars. So let's move to Germany.” Ever since I've been working in the automobile industry. Firstly, as an external for non-Porsche car manufacturers. I was mainly working as a Big Data Architect, an architect in Hadoop architecture, setting up Spark streaming hive, MapReduce, and all that jazz.

Sridhar Mamella (03:26):

However, I think my interests were always inclined towards streaming. I think it was more inclined towards something which is different than the bare metal Spark streaming. I came across the Hadoop, also I came across the Apache Kafka ecosystem technology, and that's when in the summer of 2019, I got a chance to join Porsche. Since then, I'm helping them build the next big data, or let's say data streaming platform from scratch. That's what motivated me to make the big switch. Fast cars and fast data. The combination was unbeatable.

Tim Berglund (04:11):

It's plus a Greenfield project. So it's not clear to me how one says no to that.

Sridhar Mamella (04:17):

Mm-hmm (affirmative).

Tim Berglund (04:19):

So tell us about that Greenfield project. We're joking about the term Big Data, I think a little bit, and when you say you have a master's in Big Data from 2015, people in the industry, I'll say, in contrast to the Academy, we're still saying Big Data, not ironically in 2015, but it's still impressive for the Academy to have cottoned on to that by then because they normally, with respect to computer science, things at least go a little slower. But, you ticked off all of those usual suspects from the early and mid-2000s. You were involved in those, said you've kind of were inclined toward streaming, you something, so what is it that you're building that Porsche? What is this Greenfield project?

Sridhar Mamella (05:08):

I'm a platform manager at Porsche. I'm responsible for building a robust, fault-tolerant and stable data streaming platform. I'm solely responsible for all the internal data streaming platform activities. We just build a platform and we call it StreamZilla. StreamZilla is based on the open source Apache Kafka. We use the Confluent Community Edition. So, firstly, thank you Confluent, for open sourcing and much-refined version of Apache Kafka.

Tim Berglund (05:42):

We're incomplete.

Sridhar Mamella (06:51):

Along with working on StreamZilla, the central data streaming platform for Porsche, I'm also responsible for the data streaming strategy at Porsche, which is kind of new. This is just to be the stepping stone to Porsche's path to the Cloud and a Cloud-first strategy. Basically, the end goal is that all product teams worldwide are to be offered a one-stop shop for streaming services, replacing the existing infrastructure construct of many different applications with this signals central solution.

Tim Berglund (07:26):

When you came to Porsche, what sort of legacy environment did you see, if you're replacing many point solutions with one thing, so what kind of things were there?

Sridhar Mamella (07:39):

As I started at Porsche, we did have Kafka in place, but we had about four or five different instances of Kafka running around. Kafka is pretty easy to set up. But it gets the pain points to do on the [inaudible 00:07:53] come up later in the stage when people are in production, and with all the upgrades, and especially when you're working with open source technologies, you need to maintain a lot of source code, you need to take care of a lot of security issues by us internally. That's something that the product teams don't really want to do in the long run. Then that's how the idea developed to have a central platform.

Sridhar Mamella (08:25):

Yeah, I would say that was almost like stage one. I'd say we started… Our goal was to move to early production use cases. We started small, think big and acted fast. Basically, our main goal was to enable the tech teams with increased agility and flexibility with a so-called streaming platform, or say, a stable streaming platform.

Sridhar Mamella (08:55):

I guess this is when many business teams saw benefit, invite a business use cases, such as improved customer experience, reduced costs, and they quickly got a technological advantage, wherein they didn't have to take care of the... or say maintain the technology once it's been implemented.

Sridhar Mamella (09:21):

Basically, for a customer for a stream to the team is an internal project team, with a data streaming use case or a pain point or a bottleneck. That's how we started with the first pilot project.

Tim Berglund (09:36):

What are the sources of streaming data in a car company like Porsche? Because I have to confess, I can think of two things like manufacturing, there's somewhat traditional IoT, because you build things out of metal and there are devices that might throw off data, that you want to understand. Then connected cars, am I on to it, or where does data come from

Sridhar Mamella (10:03):

Yes, you are right. Nowadays, everything's connected. We have been in the industry for 4.0 revolution where in all the machines are connected, we have predictive maintenance, all the data is being produced, it has to be transported to a different medium or a storage to be analyzed and get real time analysis back. And the amount of things that we could do with this data with respect to predictive maintenance of customer experience.

Sridhar Mamella (10:35):

But also at the same time, we also have use cases. With regards to Kafka, for example, with a workshop, for example, a customer has a car somewhere in Australia, and he goes to a workshop and all the data from his car, or his warranty and guarantee data is transferred over from Kafka or using Kafka from our service in Zuffenhausen, here in Germany, or to Australia, or maybe into America. So this is just some [crosstalk 00:11:06].

Tim Berglund (11:08):

[inaudible 00:11:08] even at point of service for authorized Porsche repair centers.

Sridhar Mamella (11:13):


Tim Berglund (11:15):

Okay, that's very interesting. I'm a Toyota household. I'm not a car guy, I drive a pickup, and my wife has a RAV4. Very ordinary consumer level cars that clearly someone who's not trying to impress anyone with his vehicle. Is that typical of a modern car?

Sridhar Mamella (11:31):

Yes, and no. Nowadays, especially with the electric car, I guess the possibility some more to include more sensors onto the car, and especially everybody wants to have more out of the car. The customer nowadays just don't want to drive the car, but they also want to know what's happening in the car. With the current generation where people are on their smartphones the entire time, then we add this app revolution, wherein you could get all the information, no matter where you are in the world.

Sridhar Mamella (12:05):

For example, you're on a business trip in China and your car's parked in Palo Alto, in California, you can still log into your app and check your car's tire pressure. That's ridiculously awesome. I guess, we're getting into this generation, or say, this revolution, sorry. That's where we are getting to the face, therein senses are always going to be increasing.

Tim Berglund (12:32):

That's fascinating. That does mean a lot more Kafka topics and a lot more Kafka partitions, kind of globally in the world, probably. I guess that's a good thing if you love Kafka. I'm funny, I'm just thinking about this, because the extreme end of connectedness and app integration.

Tim Berglund (12:52):

Like I've got a friend who drives a Tesla, that sort of thing where it just feels like an extension of the app. My guess is that's kind of where you're going to, as a pickup truck driver, just as an aside, as a pickup truck driver, it turns out that tech like this comes slower into those, I drive a 2018. So it's pretty late model pickup and other kinds of Sedans in that same model year, had features and integrations, and reporting that pickups don't. So I guess that's the trade off you make, you get all this utility and four-wheel drive and a big bed, you can haul things in, but you don't get to check your tire pressure remotely. I have to say, Sridhar, right now, I'm a little sad that you've said that I kind of want to be able to do that. I don't know what I'm going to do with the information, like am I going to make an exclusion?

Tim Berglund (14:42):

Right. Okay, well, goals for me. This is a helpful conversation. Thank you. I think this may inform future car purchases of mine. You have mentioned StreamZilla, now this is I assume the unified streaming platform. You came in and you saw lots of small things being built, which, by the way, on the cusp, or as a technology wave is beginning to break, that's what happens, you get people who are technology leaders, whether they have a leadership position in the formal organizational hierarchy or not, but they're like, “Oh, hey, streaming is a thing.” Maybe they saw it in 2016, or something like that, because it just super pression and forward looking, you know those people.

Tim Berglund (15:29):

So you get those little teams in a large company, building systems like that, deploying their own stuff, getting it done, advocating for technology change in their part of the organization. Before there is appropriate consolidation around standard solutions in that new space, you get lots of weird stuff.

Tim Berglund (15:50):

Like, I don't know what it was, like, I was in high school in the late ‘80s. I don't know what it would have been like to deploy a project on a relational database, then. That was the equivalent. It was probably that same sort of messiness. Then in the early ‘90s, you'd be like, “Well, okay, let's all make this Oracle.”

Tim Berglund (16:09):

So, you saw that diversity, which is, you probably coming in, you're like, “Oh, this is a disaster, I need to fix it.” From where I sit, I see that as a sign of an innovative technology organization that drives decision making down to the lowest level possible and then think there's all kinds of good about that. But clearly, you need to get it reined in, and that's your job. So tell us about StreamZilla? Tell us the story, like how did that start? You've hinted at it, but walk us through it.

Sridhar Mamella (16:44):

We started off of the existing Kafka users, we had around three or four Kafka productive teams, which were live in production, but with their own little services. We got together with them. Then we had to sit down, we had to basically understand what their requirements are, what their pain points are, what are the bottlenecks that they're facing, and basically trying to keep the startup mindset, we had to understand the customer firstly, try to build a service and at the same time, keeping in mind the user experience. Sridhar Mamella (17:26):### We bought some professional services from Confluent, we arranged a lot of trainings, the first step was to train and educate customers or product teams, and also us internally, and trying to understand what we want to set up. I guess the next stage was trying to free up data which are stuck in legacy systems or external data stores and trying to brainstorm and trying to understand new streams of data, which were previously not possible to process with legacy systems. Well, I'm talking a lot about legacy systems here. But here, one need to understand that as a very traditional luxury car manufacturer, we focus a lot on quality. There's a German saying that says, “Never change your working system.” Sridhar Mamella (18:18):### That was one of the main pain points are trying to get people on board with a new system that's being built when something's already working stable. We worked a lot with open source technologies. Firstly, we started with open source on Prem Kafka clusters in the pilot project. We had also set ourselves a goal of publishing all our source code as well.

Tim Berglund (18:47):

StreamZilla, is that available as an open source project [inaudible 00:18:52]

Sridhar Mamella (18:52):

No. We do believe in the fact that when open source flows in, open source must flow out. Obviously, I don't think it would be possible to open source the entire StreamZilla solution, but we do have a few open source components that we would want to open. I'll say make available in the open source market, but we are right now in the process of getting the required... what do you call it? The required goals for internal Porshe.

Tim Berglund (19:26):

Open source revolves on-

Sridhar Mamella (19:28):

Approvals. Exactly. Thank you.

Tim Berglund (19:29):

I think what you're telling this American is that there are lawyers in Germany also.

Sridhar Mamella (19:36):

Exactly. And the German company even more.

Tim Berglund (19:39):

Go on. I'm sorry. I have some more questions about StreamZilla, but I'll let you finish.

Sridhar Mamella (19:43):

No, that's it. I guess, once we have all these approvals, the idea is to release a part of the code on to GitHub, and contribute something to the open source community at Porsche.

Tim Berglund (19:54):

Wonderful. Part of what's really useful about this kind of conversation, is that, what you're doing with StreamZilla, Sridhar, I keep wanting to say FileZilla. If I do, I'm going to give instructions for it not to be edited out, because I think everybody should just hear that mistake. It comes so close to happening so many times, I just want to hang a lantern on it and say, "This is just a thing." And I might do it.

Tim Berglund (20:27):

Anyway, StreamZilla, when you're building a thing like StreamZilla, one of the things that's useful about this kind of conversation is that, other organizations of your size and larger, do this kind of thing. It's not unusual to try to create standards around how event streaming is going to work. I don't know if this is a thing you know, or can tell me, these are separate questions. But how many developers would be at Porsche using a system like this? Sridhar Mamella (21:00):### It varies a lot. We started off pretty small with about just three product teams for whom we built a centralized streaming solution. As for now, I guess we have about 20 product teams who are our most productive life. We're looking at extending and going global with the solution to the Porsche subsidiaries all across the world, and also, Porsche China and Porsche America. I guess the user basis increasing day in and day out. We do have-

Tim Berglund (21:37):

You could get to a point where you had 1000 developers easily [inaudible 00:21:41].

Sridhar Mamella (21:41):

[inaudible 00:21:41] be too many, I guess, for a company like Porsche, which is not an IT company. But I would say definitely, a thousand people that have something to do with Kafka. They don't work directly with Kafka, but definitely are relying on the system Kafka to work, so that they could do their job, or do the analytics or Big Data?

Tim Berglund (22:05):

Yeah. But this is consistent with what I see, what I see is when you get into the 10 to the two, 10 to the three kinds of developer accounts, hundreds, or small thousands, wherever that you are there, it starts to become valuable to provide the Porsche version of Kafka. Tell us from an API perspective like, "I'm a developer working on a product team. Suppose I know Kafka? Maybe I'm particularly forward looking, and I'm one of the people who built some early stuff. Or I've just hacked on my own, or previous job, whatever. I know Kafka API's, and basic semantics." You are you're providing… StreamZilla is a number of things, but I'm guessing at least a rapper. There's now an API that a Porsche developer sees, what does that look like? What kind of abstractions have you chosen to provide? What are you trying to hide? What decisions does StreamZilla want to make? How did you work through those trade-offs

Sridhar Mamella (23:08):

In starting off, we had pretty clear set goals on what we wanted to achieve and what we wanted to provide as a service. The first three or four teams that we told we started off, they had production level Kafka, and they were already in production, and they knew what they wanted. So basically, they had to deliver what they wanted, and they had different use cases running across the globe, with regards to warranty guarantee worldwide support and the online portal for Porshe owners.

Sridhar Mamella (23:52):

Basically, it was trying to fulfill these needs that they had, and secondly, trying to show them what else could be done. Because basically, when these teams productive started off with Kafka, it was just to solve one single problem, say, one single problem from that project perspective.

Sridhar Mamella (24:11):

One of the trade-off was trying to educate these teams again, or say, trying to show them the opportunities how they could use Kafka differently, or in some different use cases. That also meant trying to include them in the development right from the start, rather than building a product, or service and just pushing it onto the people telling them, “Hey, this is what you have to use.” But trying to build it together with them, and trying to build a useful product that they could use. Let it be with control over their own topics, or having topic level controls, so topic level ownerships or defining your own ACLs and your own certificates. We have a lot of security, which we need to take care of as a German car company. Everybody has really high importance on such issues. That's how it started.

Tim Berglund (25:20):

Due to consumer protection regulation and privacy laws, that's a lot of work you have to do there. There is a certain amount of that security stuff that you want to make a standard part of what you present to product teams. But it sounded like you said, you also let those abstractions leak. Like you would give control to the product team to be able to see levels or see details as fine as Kafka ACLs. Did I hear you right?

Sridhar Mamella (25:54):

Yes, and no. As in hold, the entire StreamZilla solution has the security clearance from the Porsche's security team. We did have a very intense penetration testing and all sorts of security clearance documents that we had to fill out and get all the high level clearances that we needed for this platform to function. That defines StreamZilla as security, or say, a platform that is… How do you call it in English? I speak a lot of German these days. In German is called [foreign language 00:26:39]. A system that's been given free.

Tim Berglund (26:47):

Production system that's been released, that's been published?

Sridhar Mamella (26:52):

No. A system that had… No, sorry, exactly. A system that has a security clearance at Porsche's level and at Porsche's standards. Once a system had these clearances, they could go ahead and operate with production data that is a must for every system. Once we have this, then based on the guidelines that we have defined in our security documents, we could go ahead and give the teams the functionalities that they've requested for all that we have promised in our security documents.

Tim Berglund (27:30):

Tell us about day two of StreamZilla. So you gather requirements, you build a thing, you have made the point that you have to deliver feature parity to existing teams, will go back to the German phrase that says, “Don't replace a working system.” You can't give them less than what they have, because free people won't do that. So there you go, you build it, you deliver that parity, and it's running on day two, what's it like? How is it in production? The first little bit?

Sridhar Mamella (28:05):

Yeah, and in the stand we did have, I would say in the stand in the first couple of weeks, we did have a couple of hiccups. With regards to production data, some brokers were down or we were still fine-tuning or monitoring systems, and bits and bobs like that.

Sridhar Mamella (28:22):

But I guess, as a whole now being live for about a year and a half. It's running pretty stable. There's not much that we need to tinker with. Kafka in itself is pretty fault tolerant. But for us, I would say the tool looks different rather than trying to maintain what's already running, rather trying to improve what's already running. That's where the company-wide strategy comes in place. That's carving Porsche's path into the Cloud.

Sridhar Mamella (28:58):

We have Cloud-first strategy at Porsche now, which means connecting the internal data streaming platform to the Cloud hand in hand with Porsche's so called, Cloud-first strategy. This aims to bring all services and infrastructures into The Cloud within the next five to 10 years. This means a paradigm shift for us as a technology, or as a traditionally established company, whose main focus over the past years was not IT, but was on producing great sports cars. We always try to convey the same message to every user is who's driving a car. Also, if there's IT running in the background, he should not be affected by IT.

Sridhar Mamella (29:50):

There are tons of people in this world who still love to listen to the engine sound when they're driving 911 or 718 Boxster. That's the feeling that we don't want to take away from these people. That's where we need to keep in mind that, “Yes, we are looking into the Cloud, we are looking into more software.” When software is playing a major role.

Tim Berglund (30:18):

Yeah. This gets us back to the old Marc Andreessen saw about software leading the world. And the Jay Kreps, a strong form of software leading the world, I'll link to a blog post in the show notes if you want to read more about those. But it's true. The phrase is a cliche, but like, most cliches, there's a significant amount of truth in it. Porsche finds itself becoming kind of a software company a little bit, maybe not so little.

Tim Berglund (30:52):

But your Cloud migration is interesting because we talked a little bit, your German company. That's home base, but you're also a global company. I don't know how many Porsche entities there are, and how many countries you formally operate, and to whose laws you must comply, but it's a big number. That's legally complex. So moving to the cloud, sounds like there will be interesting data sovereignty rules. How does that work? So you want to be 100% of the Cloud? Will there be a hybrid thing? Do you just have to have Cloud providers with regions in certain countries? Can you give any insight into that?

Sridhar Mamella (31:38):

Well, it's a topic that changes every day. Every country has its own guidelines, for example, China is a very hot topic these days, at the same with America. In the past, okay, just some numbers, in 2019, Porsche sold about 280,000 cars, and we put about 5000 Porsche Taycans on the road in China, America, and Europe.

Sridhar Mamella (32:07):

And 80% of those cars were in China and in America. When we're looking at going into the Cloud as a company-wide strategy, then we need to take things into account that the service level agreements that we have, or the services that we're offering, need to function exactly the same, irrespective of it being in America or in China, or in Europe, for example. That is a tricky question, or tricky... not tricky question, tricky problem to solve.

Tim Berglund (32:49):

Quite so. Quite so. It is, to some degree, like you said, the application functionality needs to be the same.

Sridhar Mamella (32:58):


Tim Berglund (32:59):

People don't normally drive their car from Germany to... or from China to the United States, that would be a little weird. You'd have to get a boat involved at some point. Although with higher-end cars, like Porsche, maybe that happens more than it does say, Toyota Tacomas. But definitely possible to drive a car from, say, Ireland, to Germany through a number of other jurisdictions, with completely different data privacy, and data sovereignty regulations. As you guys build that out, and I know, it's not done yet. Those aren't all answers that you've got right now. But it occurs to me that that's going to be super interesting. Dealing with that legal mess, you're going to spend time on the phone with those lawyers. I think you should make friends with them, and you should treat them nicely. Those are big problems to solve. Tell me about Taycan? What's Taycan?

Sridhar Mamella (34:01):

A lot of people firstly ask me, what does the name Taycan mean? Do you know, Tim? What does the Taycan mean?

Tim Berglund (34:08):

I don't and I don't want to Google it. I feel like it would be dishonest. I feel like I just need to say I don't know.

Sridhar Mamella (34:13):

I could see you type.

Tim Berglund (34:15):

The nice thing is, this is an interview and I'm asking the questions. So what does Taycan mean?

Sridhar Mamella (34:21):

Oh, you want to put it that way? Okay. Well, Taycan, it's the new Porsche electric car. Apparently it's a name that fulfills every phonetic, legal, creative, strategic, and model-specific requirement. It's basically composed of two terms. I guess these if I'm not mistaken, these two terms have a Turkic origin. I guess the word can be roughly translated as, 'soul of a young spirited horse'. That's exactly what Porsche is; lively, impetus, vigorous, light-footed on long stretches without tiring. And yes, obviously free-spirited.

Tim Berglund (35:09):

Okay. And this is an electric car?

Sridhar Mamella (35:11):

It is an electric car, and the advertising slogan is sold electrified. I guess for me as a petrol guy I guess it fits perfectly with the souls and the future of the brand, which is the horse on the Porsche crest, the expression of its soul, and the way it's going to a new era of the sports car.

Tim Berglund (35:37):

Yeah, well, folks, I did just Google Tycan after Sridhar told me what it meant. The good news is, I can put a link in the show notes to what one of those looks like. It's pretty cool. So yeah, you may Google yourself T-Y-C-A-N. But it's a nice looking car. I don't feel… like I'm not, as you say petrolhead. I've never been a car guy. I don't feel like I'd be mad, though. I don't feel like driving one of those as a rather beautiful looking performance vehicle. I don't feel like that would just make my life get worse, somehow. I can also say for a fact Mrs. Berglund would not mind me dropping one of those.

Sridhar Mamella (36:22):

That's the thing to think.

Tim Berglund (36:26):

Looking for the downsides, and they're not in evidence. So yeah, that's a nice looking vehicle. Anyway, out of cars.

Sridhar Mamella (36:35):

It could do 180 miles per hour.

Tim Berglund (36:38):

Oh, okay. I would do that. I might get in trouble here in these United States. But that sounds like something worth exploring. How about from an operational perspective? I meant to get to that. We talked about API's and the developer-facing aspects of StreamZilla, but what did you learn operating this, or what have you learned so far?

Sridhar Mamella (37:01):

Mm-hmm (affirmative). See, the two sides to this learning curve or learning coin, as I would put it. One is the DevOps perspective, and the other would be the business perspective. I'd say from a DevOps perspective, the central learning of the transition process has definitely been that it's simple. You just need to talk to people and just need to ask them what they need.

Sridhar Mamella (37:27):

Often, no one outside the team knows in detail what they need to work on to do their best. That's in every relationship. Often when you speak to your spouse, you just need to communicate and that's just the key to success.

Sridhar Mamella (37:43):

The second learning would be a however, it's just as important to remain in constant exchange, and not to lead the team to learn when you have a finished product. So always trying to be with the teams trying to build on top whilst keeping them involved with your product, and for us, it means absolute added value. We build a platform in such a way that every requirement has been considered, or at least looked into all is on the pipeline, or in the pipeline.

Sridhar Mamella (38:17):

I would say the second learning would be from a business perspective. The switch to StreamZiller was not only relevant for the DevOps guys, but also had a significant positive effect on the business area. For example, a lot of ideas are always generated here in the business area when we just talk to people and these ideas, obviously, from the business perspective need to be implemented quickly and need to be always implemented now, because there's no tomorrow.

Sridhar Mamella (38:50):

But thanks to the central Kafka data exchange platform, the teams know that they have precisely one contact person for all the concerns, who is always available for them if they want to talk if they want to discuss the use cases. This is how we could implement ideas and use cases pretty quickly. This knowledge from the central platform stays in one place rather than being diverted into small teams. If this person switches from one team to another team, or if there's a person who has switch into a different company, all we want to avoid the brain loss that we have, a brain dump, not brain dump, brain loss, right?

Tim Berglund (39:37):

Brain drain.

Sridhar Mamella (39:38):

Brain drain, yeah, thank you. That's the main advantage for use cases, I would say from a business perspective to speak to the people where all the expertise lies, and one contact person.

Tim Berglund (39:53):

Yeah, when you get really an organization that executes a very mature adoption of event streaming, they really drive it to an organization-wide sort of thing. There have been previous podcast episodes on what we call our streaming maturity model. This is level five, where the business is an event-driven business. And all of the IT stack in the business is more or less event-driven. So when you start to get towards that, you do find, like you just said, again, and this feels like an event streaming evangelist talking point, but you actually see where data that would have been in a silo in somebody's database in some department that nobody else gets to see, when you've got a StreamZilla kind of system, that data becomes more accessible to other parts of the organization. You still got governance, and we talked for a while about the complexities of that sort of thing as a global business. That doesn't go away. But the nature of the technology is to make the data available. And so there's value the business gets to unlock, that would have been hard to realize otherwise.

Sridhar Mamella (41:08):


Tim Berglund (41:11):

We're recording, this is late August 2020. So we're, I like to say in the middle of the pandemic, as far as I know, we could still be in the beginning, I think we're in the middle, I think maybe we're-

Sridhar Mamella (41:27):

[crosstalk 00:41:27] in the middle.

Tim Berglund (41:29):

[inaudible 00:41:29] towards the end of the middle. I must be careful with that. But I'd like to believe it's the middle because that means there's an end. It's not too long from now. But it's been typical for me to be a person who travels a lot. I've actually never been to Stuttgart, which I view as a personal failure. But when life gets back to something approximating the way life was before, however, we do things differently, and then circumstances have evolved. I would love to be able to visit. I don't know if you know you have test cars that are available for demos, or anything like that, I had to put you on the spot like this. But that seems like that would just be agreeable to me. Again, I don't think my life would be worse if I got to drive one of these things like a Taycan or something?

Sridhar Mamella (42:17):

Well, we could definitely get you into one of the cool cars. No promises, but we could set you up at something when you come down here. We could go for a spin on the Autobahn, and you could probably do a... you know that we don't have speed limits here on [inaudible 00:42:34] Autobahn.

Tim Berglund (42:36):

[crosstalk 00:42:36] I've experienced that with my colleague, Kai Vayner, a few years ago. Let's just say the vehicle was not a Porsche. We were going fast by American standards, but he had Ford and it was, I don't know, like 120, 130 miles an hour.

Tim Berglund (42:57):

Folks, just you know, that Sridhar and I did talk about that question. I did not just do the tackiest thing in the world and put him on the spot in a podcast asking him to get me in a car. Yes, that's the thing we want to do. I would love to be as you guys and talk to you about whatever is current and whenever I can get there.

Sridhar Mamella (43:14):

Yep, let's do that.

Tim Berglund (43:16):

My guest today has been Sridhar Mamella. Sridhar, thanks for being a part of Streaming Audio.

Sridhar Mamella (43:20):

Thanks, Tim, for having me.

Tim Berglund (43:22):

Hey, you know what you get for listening to the end? Some free Confluent Cloud. Use the promo code 60PDCAST to get an additional $60 of free Confluent Cloud usage. Be sure to activate it by December 21st, 2021 and use it within 90 days after activation. And any unused promo value on the expiration date will be forfeit, and there are a limited number of codes available, so don't miss out.

Tim Berglund (43:50):

Anyway, as always, I hope this podcast was helpful to you. If you want to discuss it or ask a question, you can always reach out to me @tlberglund on Twitter. That's T-L-B-E-R-G-L-U-N-D. Or, you can leave a comment on a YouTube video or reach out in our community Slack. There's a Slack sign up link in the show notes if you'd like to join. And while you're at it, please subscribe to our YouTube channel and to this podcast wherever fine podcasts are sold. And it you subscribe through Apple Podcasts, be sure to leave us a review there. That helps other people discover us, which we think is a good thing. Thanks for your support, and we'll see you next time.

We have all heard of Porsche, but did you know that Porsche utilizes event streaming with Apache Kafka®?  Today, Sridhar Mamella (Platform Manager, Data Streaming Platforms, Porsche) discusses how Kafka’s event streaming technology powers Porsche through Streamzilla.

With the modern Porsche car having 150–200 sensors, Sridhar dives into what Streamzilla is and how it functions with Kafka on premises and in the cloud. 

He reveals how the first months of event streaming in production went, Porsche’s path to the cloud, Streamzilla's learnings from a developer and a business perspective, and plans for parts of Streamzilla to go open source.

Stick around through the end as Sridhar talks through cloud migration, cloud-first strategy, and Porsche’s event streaming use cases. This Streaming Audio is all about speed—fast cars and fast data, an episode you won't want to miss!

Continue Listening

Episode 133December 16, 2020 | 47 min

Choreographing the Saga Pattern in Microservices ft. Chris Richardson

Chris Richardson, creator of the original Cloud Foundry, maintainer of, and author of “Microservices Patterns: With Examples in Java,” discovered cloud computing in 2006 during a Google talk about APIs for provisioning. Chris explains what choreographed sagas are, reasons to leverage them, and how to measure their efficacy.

Episode 134December 21, 2020 | 10 min

Apache Kafka 2.7 - Overview of Latest Features, Updates, and KIPs

Apache Kafka® 2.7 is here! Here are the key Kafka Improvement Proposals (KIPs) and updates in this release. Find out what’s new with the Kafka broker, producer, and consumer, and what’s new with Kafka Streams in today’s episode.

Episode 135December 22, 2020 | 46 min

Mastering DevOps with Apache Kafka, Kubernetes, and Confluent Cloud ft. Rick Spurgeon and Allison Walther

How do you use Apache Kafka®, Confluent Platform, and Confluent Cloud for DevOps? Integration Architects Rick Spurgeon and Allison Walther share how, including a custom tool they’ve developed for this very purpose.

Got questions?

If there's something you want to know about Apache Kafka, Confluent or event streaming, please send us an email with your question and we'll hope to answer it on the next episode of Ask Confluent.

Email Us

Never miss an episode!

Confluent Cloud is a fully managed Apache Kafka service available on all three major clouds. Try it for free today.

Try it for free