EP 327 – Event Mesh Networks Explained: What They Are and How They Enhance Operational Efficiency – Sumeet Puri – Chief Technology Solutions Officer at Solace

by | Jul 3, 2024

"So an event mesh is a data movement fabric, okay, yep, so the network, but it's the layer between the network and the applications. So if we explain it a little bit further, event broker is this technology that allows you to move data in a one to many manner, in a publish to subscribe manner. So for example, if an order is created, and we say that order event is published to the event broker, that order may have to go to the order management system. It may also go to the digital receipt system. It may also go to the loyalty system. So the same order has to go to multiple systems. And this pattern is called publish, subscribe, which is done by an event broker." - Sumeet Puri

In this episode of the Asia Tech Podcast, Sumeet Puri, Chief Technology Solutions Officer at Solace, shared deep insights on the transformative power of real-time data and event-driven architecture. He explained how advancements in network and computing capabilities have enabled industries to transition from traditional data handling to real-time data processing. 


Some of the topics that Sumeet covered in detail:

  • The evolution from traditional data handling to real-time data processing.
  • Real-time data is no longer a luxury reserved for high-frequency trading desks. It has permeated various industries, enhancing customer experience and operational efficiency. 
  • The importance of event-driven architecture in facilitating real-time data movement. This architecture operates on the principle of event meshes, which are networks of event brokers that handle data in a publish-subscribe model. 
  • Real-time data processing democratizes access to information and services. Puri highlights initiatives like Singapore’s SGQR and India’s UPI as examples where real-time data movement has enabled even the smallest merchants to participate in the digital economy.
  • The concept of ‘Event Storming’ workshops is an innovative approach to identifying and leveraging real-time data opportunities. By breaking down business processes into individual events, companies can uncover previously hidden efficiencies and opportunities.
Read the best-effort transcript

Read the best-effort transcript below (This technology is still not as good as they say it is…):

Michael Waitze 0:04
And I’m ready to go. You ready? Yes, let’s do it. Hi. This is Michael waitze, and welcome back to the Asia tech podcast. Sumeet Puri, the Chief Technology Solutions Officer at Solace, is with us today. I am mumbling and stumbling. I love it, Sumeet, thank you so much for coming on the show. How are you doing today?

Sumeet Puri 0:24
Very good. Michael, very nice to be on the show, and thanks for having me. Really good to talk to you.

Michael Waitze 0:30
It is my complete pleasure. Look, before we jump into the main part of this conversation, let’s give the listeners a little bit of your background for some context. Yeah,

Sumeet Puri 0:38
sure, yeah. So I’m Sumit. I’m based in Singapore. I’ve been here for about 23 years. Oh, wow. I’m Chief Technology Solutions officer at Solace. What that means is that I work with various enterprises, B to B, scenarios B to G, scenarios G to be scenarios, basically helping organizations move their data in real time, helping their integration better. And by training, I’m an architect, software engineer. As a software architect, work with companies like Tibco, BA Systems and Singapore Ministry of Defense over the years. Oh, wow, yeah. Sometimes tinker and write some code. Sometimes I write some code.

Michael Waitze 1:23
Do you miss writing code?

Sumeet Puri 1:25
I do, actually, I really do. So I don’t get to write code as a part of my job. There’s just not enough time. But I do some tinkering. So one of the recent things I did was, we’re a family of cricket lovers, and there’s this IPL which is going on. So for my son to get him interested into computer science a little, we wrote a little program which is, you know, screen capture ability of scorecards, which captures the real time score, and if our team has hit a six I was taken a wicket is doing well, the smart lights in my living room basically listen to that and they go crazy. So that’s a little hack that we did again, real time information as it happens, the house needs to respond.

Michael Waitze 2:13
I was gonna say, How old is your son? If you don’t mind me asking,

Sumeet Puri 2:15
he’s 17. Oh,

Michael Waitze 2:17
my God. Okay, that’s great. What a great experience. And I love the fact that you’ve integrated this is like such a great segue your house into the experience of writing the code and keeping yourself involved actually makes it a family affair, which makes it really cool. Can you talk to me a little bit about maybe some TEPCO experience, and why you think real time data, just at scale, actually matters, and then let’s talk to about it, about it in the context of Solace. Yeah,

Sumeet Puri 2:41
definitely. See, we were used to real time in the real world, but not so much in not so much in the digital world, because as computer systems evolved, networks were traditionally weak, right? So why do we need real time data? And I’ll get to networks, right? So when you make a payment, do you want to get the receipt instantly, right? When you order something online, do you want to track that order instantly, in real time? Do you want to see whenever you want? You want to see where that order is, if you’re a business leader, do you want to see the state of your business whenever you want? Or do you want to wait for a weekly, monthly, quarterly report as to how the business is doing? And again, your background? Michael, capital markets, when price sticks us, dollar to Singapore, dollar ticks in New York. Do you want to have that price available to you instantly, in real time? Yeah, so that you know what you’re going to trade. So real time information, no matter what use case, look at you. Look at this cricket score. I want it in real time so that my lights can respond and I can celebrate. So we all want real time experience for better customer experience, better business decisions, more insights, and, you know, even better innovation, and that’s the need for it. But we would challenge historically, and it’s been, been getting incrementally better. But when I was growing up, or when I studied Fortran seven, seven in school, and giving my age a little, basically, in that era, this was much faster than network, right? Networks practically existing, and it just started to happen. So to move information at scale in real time, the underlying physics of the computer system hadn’t evolved sufficiently to make it happen. So computers and applications programs were written in a way that you leverage a lot of disk, right? So you write to disk, you read from disk, and you don’t get to move information real time.

Michael Waitze 4:43
Yeah. I mean, look, a lot of stuff was actually put in memory as well to make stuff faster too, right? Because the memory was actually faster than the network as well. Can I? Can I ask you this, though? Um, actually, I want to go back and do and talk about it’s like as technological transformation, if you don’t mind. And I want to go really far back when I was a kid, I lived. New Jersey, but my favorite sports teams were in Boston, and I waited every day for the for the newspaper to come in the morning so I could run outside to the front yard and pick up the newspaper to see the scores. Right? I didn’t have a radio in my room, yeah. Okay, so that’s how I got my scores. So the bandwidth was really the time it took. I want to make an equivalency here, right? The bandwidth was essentially the time it took for the newspaper to get the information, print it in Manhattan, and then drive it on a highway to my house and then have some kid drop it off. That was their bandwidth. But over time, and even during my childhood, we figured out how to send things up to satellites and USA Today was born, which meant you could get any score for any team at any time, right? Because USA Today, then didn’t could print in any neighborhood, and that was that was again, increasing bandwidth. What had to change, right? I know what had to change on the trading side, but what had to change on the general business side to make all this real time data possible, so that the disk wasn’t faster than the network, and when did that kind of intersect and change?

Sumeet Puri 6:05
Yeah, it’s a, it’s an evolution, right? It’s not really a, say, an instantaneous change, but a few things came together. So see, disk bandwidth, give or take, mainstream disk bandwidth is a couple of gigs per second, okay? And till about 10 years ago, mainstream data center networks, when we were deploying stuff, there were one gig per second. And we can achieve more by parallelizing, by putting more ports in there, but the wireline speed became 10 gigs per second in mainstream data centers. Again, investment banks were a bit faster. They could afford it. They needed it. But mainstream data center, seven to 10 years. 10 years, I think would be reasonable to say that data center started to become 10 gig, and that this bandwidth is still, you know, is not 10 gigs, 10 gigs per second, mainstream. So as a result, network allows you to move much more data, so much so that with the five gig, five gig radio on your mobile phones, that itself is significant bandwidth. So that’s one trend that happened, right? Another important trend that happened associated with this is the entire evolution in mobile applications and smartphones, the way that you know, the these systems became very, very distributed, that we’ve got like almost a supercomputer in our hands, relative to the computer that’s put man on the moon. So if you look at that in comparison, so you’ve got better compute in your hands, which is completely connected data center networks are much, much better. And then lastly, the human need for evolution to, you know, keep making things better. The need for data to become real time, the impatience. In other words, we are all getting more and more impatient, right? You were probably happy to have that weight, that suspense, for the scores, for the newspaper to come to your house, and now it’s like, the moment I turn my phone on, the score has to be right on. Yeah, I want to

Michael Waitze 8:07
see the highlights right away. I’m not even satisfied with the score, but, yeah, go ahead,

Sumeet Puri 8:12
right? So I feel, I feel very disconnected. I mean, I literally am. But when I’m in airplanes, like what I’m going to do now, right? Because the network is slow, so you combine all of that, right? The need for applications, need for humans to get this data, the underlying change in the network, the compute available to all of us in data centers, but also in the House and in the form so that distributed environments. And then add to that, companies like AWS and then Microsoft and Ali cloud and all of these companies, Google, basically making cloud a reality, making it secure enough with all the other capabilities that you can have your data in other people’s data centers, and we’ve been using it with email. So when you bring all of those factors together, that’s, I think, the reason why we can think of our architecture differently now and move information in real time.

Michael Waitze 9:07
I want to talk about real time transfers in the finance world, because that’s where I understand them the best. It always surprised me when I made a bank transfer that it took two days to clear when I was when I was younger, even when I was in university, but when I got into the finance world, I realized that, like t plus two or t plus three was the thing that the banks actually had agreed upon for multiple reasons. One of them was probably throughput at the time. But I want to ask you this, we started dealing with real time data on trading desks about 20 years ago, and yeah, we had a ton more money to spend on building those systems. Again, did things get less expensive in relative terms over the last 15 or 20 years to make it possible for, you know, just like a fast moving consumer goods company, to get data in real time as well, because we knew we could do it back then, we couldn’t live without real time trading data. It was not possible to do but FMCG companies have been able to live with it. Until maybe a few years ago. Again, did stuff just get so cheap that it just became possible?

Sumeet Puri 10:05
Yeah, I think again, it became cheap enough in relative terms, right? So the cost per byte of data, per gigabyte of data, has been dropping, right? If you just look at your phone contract as well, right? Pretty sure your price is not multiplying in the same proportion that the bandwidth or the data packages, right? So we can say that the cost of data transmission has been dropping, no doubt about that. But it’s also, I think, the overall economy, if you see that the world economy has been growing, especially in the developed world, but now also in the developing world, GDP has been growing. So that means the overall wealth generation in the world. Again, it’s unfortunately, not in an equilibrium, it’s not fair, but in absolute terms, that has been growing. And as a result, you know, normal people have phones like 20 years ago, who would have fought on the trading floor, like when the first, you know, those big Nokia bricks and blackberries were launched. They were luxury. Now everybody has it. So it’s it’s a combination of all of that, and on the trading floor. See banks made money by moving numbers. What products do banks have? Right? They basically have the ability to store numbers and have assets and liabilities associated with those numbers, and if they could move that data quickly, they could act on it. They could buy and sell the electronification of the exchanges of the ECN away from, you know, people shouting in the well. Earlier trading was like, how quickly can you react to the voice in the well? And how loud was it? So that’s your bandwidth latent that became again electronified, and now you’re on a network. And we kept on getting faster and faster. How quickly can you get closer to the well, one foot of cable is one nanosecond latency. So, because Co Location, Co Location Exactly. So that’s basically the, I would say the crucible for event driven architecture, yep, and moving data in real time. So

Michael Waitze 12:05
talk to me about how companies now are doing. What did you call it? Event driven Yeah. What did you call it? I want to, I want to get the term right architecture, event driven architecture, and also flow, right? This idea of flow, and how does it move from this decentralized processes, or creating decentralized processes from centralized processes? We talk about digital transformation a lot on this podcast, but this is like digital transformation at scale, real companies actually changing the way they interact with their data, and moving from these really centralized things to a bunch of separate processes that are decentralized. Tell me what that means and how, how companies can actually go through this process. Yeah.

Sumeet Puri 12:45
So I’ll give you a couple of examples from the trading floor, and then I’ll try to relate it back to FMCG companies or retailers trading floors, because that’s where I learned it. That’s where it started, in my opinion. And if you see any of the banks, right? So the banks that we work with, Royal Bank of Canada, to Barclays, Dr bank, China, turret bank, to name a few, each one of them has an FX trading platform, or an equity stated trading platform, and that’s the front office. So they have to react to prices. They have to place again in simple speak. They have to place buy and sell orders. Once the orders are matched, then they get converted into trades. And these trades have to be settled. They have to report be reported to regulators. So all of these interactions, when you place an order to buy something that generates the price, when that order gets converted into a trade, each of these occurrences, these happenings can be called events. So an event, again, the English translation of the word would be an occurrence, something that happens. Yep, right. And if you also look at it a bit more closely, each of these actions is a combination of a noun plus a verb. Order created, order filled, order, acknowledged, trade settled, it’s always an object or an action, yeah, so that’s what a definition of an event could be in simple literal terms, right, not even in Compute terms. So as you get these orders and trades and settlements done, they are all sent to different systems, right? The order management engine, where your order book is is going to be different from your trade settlement engines. Is going to be different from your pricing engines. It’s going to be different from your regulatory engines. And they’re not at the same place, especially in FX trading US Dollars used All right, so it’s moving in the US, and I’m buying it in Singapore or selling it in Singapore. So there is a distribution requirement for these events that are happening. So what you need to make this happen? I mean, the need is real. It’s a distributed, guaranteed delivery, low latency fabric that is needed to move all of this information in real time. And that’s what we call the event mesh as a pattern. Solace is one of the providers of it that is. Pattern. Now, if I take that into another industry, let’s say we work with the Unilever and a whole like Roche, and whole lot of retail companies like Schwartz, all the little stores connect over us. We work with nets, which is Singapore’s national payment transfer system. All of these companies also have events. Like, in the case of Unilever, they talked about this study that when a ship moves in the ocean, that location update, again, noun, work, that’s an event. And if you want to track soap in real time, like, who cared about soap in real time until covid hit? Yeah, exactly, through covid became a life saving commodity. So now you want to track soap in real time. Is it on the Walmart shelf or not, or where is it? So applying the same principle of order created to location update, and by the way, FMCG companies and retailers, they also have orders, yeah?

Michael Waitze 15:56
So I mean, restaurants have orders, right as well. So orders payments

Sumeet Puri 16:00
are also orders. So what if you mobilize these orders in real time, rather than writing it to a store based disk and then updating the inventory in a non real time manner, and people don’t know where the goods are in what store? The same principles of the event mesh going to different kinds of systems like SAP or Oracle or supply chain systems. So that’s the evolution. Can

Michael Waitze 16:23
you talk about what an event event mesh is for people that don’t necessarily understand it?

Sumeet Puri 16:27
Yeah, so an event mesh is a data movement fabric, okay, yep, so the network, but it’s the layer between the network and the applications. So if we explain it a little bit further, event broker is this technology that allows you to move data in a one to many manner, in a Publish to subscribe manner. So for example, if an order is created, and we say that order event is published to the event broker, that order may have to go to the order management system. It may also go to the digital receipt system. It may also go to the loyalty system. So the same order has to go to multiple systems. And this pattern is called publish, subscribe, which is done by an event broker. Now, if you have multiple event brokers, network with each other, some in Singapore, some in us, I’m in Frankfurt, some in Mumbai, and as you connect them. Now you create a mesh of event brokers, which is called an event mesh. So one big, logical event broker for publish, subscribe, guarantee, delivery, etc.

Michael Waitze 17:30
Okay, I have a question for you, and I think very few people could answer this question, but I’m pretty sure you can. So don’t worry, there’s no math, not that that would bother you, for sure, but I remember when we first started getting digital connectivity to the exchange, right? And that was fun, because it made trading a lot easier. We had data coming in. We could do analysis faster. We could do analysis using computers, as opposed to just kind of our gut feel or writing stuff down, or using a Monroe calculator to calculate bond prices, right? Made it much easier when we had a wire to the exchange that gave us real time data, one of the things that we figured out was it now could be easier for us to take orders. This is where you talk about order, given, order taken from our clients, and acknowledge those orders, if we built an interface ourselves, whether in the fixed income space, in the in the FX space or the equity space, to be able to understand the orders that they were sending us. And I think each individual broker went out and then built themselves a system. I know we built one. She’ll laugh at this. We built a system at Deutsche that we called smtm, and that was with your hand for show me the money. But anyway, it was an individual it was an individual system. And it meant that every one of my clients, yeah, it was that funny, but it meant that every one of our clients then had to build a bespoke connection just to us, so that they could then communicate with the parameters and the inputs that we had for smtm, the financial services industry got together and created something called the FIX Protocol, right the financial what is it? Financial Information exchange protocol, and it meant that everybody then that was on fix could send any order to any person and any broker could receive an order from from any buy side entity. And that essentially changed the whole tenor, because it meant that now it was just as easy to send an order to, you know, a third tier broker, as it was to send it to Goldman Sachs. Now I’m curious if you’re seeing this same evolution happen, because whether Levi’s is ordering something or Zara is ordering something, it’s still they’re ordering fabric needles and stuff like that, or sending an order to a factory. Has anybody built a kind of fixed protocol for those industries as well? And if they haven’t, do you see that happening? Because the data is the same in the real time analysis on all this could be the same as well. Yeah,

Sumeet Puri 19:43
yeah, I’ll try to unwrap this. I think what you’ve said, the way that I understand this is two, three tiers, right? And it’s, it’s very intriguing, and we should follow that, right? So the fixed protocol does transport as well, but it’s essentially. Be a payload standard. You’re gonna see this stuff, and I’m just

Michael Waitze 20:02
gonna get it, and then I don’t care what order you send it to me, and it has this tag on it. When we see that tag, we put it here in our order system, whatever, and we understand what that is, regardless of how you send

Sumeet Puri 20:12
it, yeah. And it’s universally understood. And as a result, companies like quick fix and Cameron fix, which we work with, basically, were able to build fixed engines and anybody could, right? So, but that’s a payload standard, right? But how does that data actually flow between the banks? How does that data flow between the exchange and the bank or any other providers? So in the first place, we are saying standards are great, right? Because it unlocks a lot of innovation. But then standards are also tiered at the pure network level, the standard is TCPIP. You imagine there’s no TCPIP, right? But then at the transport, at the messaging and eventing space, there are standards like GMs historically, or rest or MQ TT or AMQP, and these are the standards for publish, subscribe of that data, one to many, or replay or guarantee delivery, or everything that is to do with that data transport. So three layers of standards, TCP at the network level, GMs, rest, AMQP, MQTT at the messaging level, which we support. All of them right? And then fix. Or in the payment world, Swift, which is pretty much the same. And in in the recent swift implementations, especially the image and as banks are moving from ISO 8583 to 20 or 22 Swift is also it used to only connect to IBM MQ now it connects to any GMs and any MQP broker. So we’ve got customers actually leveraging our backbone as a mesh, rather than just a queuing system with Swift but when it comes to retail, as you mentioned, Levi’s and all of that, I think there will be a standard there. I think right moving in that direction, there was an attempt to do some of that with Blockchain, but there are practical challenges by putting this volume on a blockchain. Yeah, throughput. What I’ve seen, and I come from India originally, so India has this initiative called the ondc, which is exactly that it’s a it’s a standardization of the supply chain, so that, you know, any small, medium enterprise producing anything in the countryside can actually advertise their catalog and their products and their orders and have, like, a open standard based logistics backbone for the entire country.

Michael Waitze 22:26
Interesting, this is very similar to what India’s done in the payment space as well, right? Creating the payment rails. Like, come on, refresh my memory. I can’t remember off top my head. Say it again. UPI, yeah. UPI, right. So the university payments interface, so same thing. And I love the fact that they’re doing this, because what it’s done in India is the same thing you’re talking about for a small merchant. I mean, for small producers, is that it’s given even the tiniest merchant the ability to take a digital payment so that it doesn’t what’s the right word, it doesn’t bifurcate this skill set that a massive entity can have, and just like a mom and pop shop in a small town or a third tier city can have and up is amazing, yeah, yeah,

Sumeet Puri 23:02
and they’ve democratized it, and Singapore has something similar. So we’ve got Singapore SG, QR, and I’ve had the pleasure of working very closely with that, because again, 110,000 terminals and the entire backbone runs on Solace. So it’s a very public and the interesting thing is that, see, this is what standards and eventing as a pattern, real time data movement does. Yeah, from Singapore, we have end to end connectivity with rupee rupay and UPI in India. So actually, I could transfer money from Singapore to any Indian phone number through that mesh. And similarly, with Thailand, with prompt pay, yep, so payments while they were national, there’s already international examples which are live and again, logistics would head in that direction as well. And and you know what intrigues me? Tell me, imagine this logistics backbone. So we deploy our technology a lot in supply chain, right goods movements and shipments, land, sea, air, anything, right? Everything is an event at the end of the day. Imagine taking that world and mixing it to the bank world, while the banks are also real time. And imagine what this could do with trade finance and trade finance without the need for collateral, if the bank knows that this small, medium enterprise is actually supplying sugar to Unilever, who’s making ice cream. The risk profile for this customer, just by having that visibility into the event stream of his orders and his manufacturing opens up great possibilities for trade finance. Yeah,

Michael Waitze 24:38
it’s really interesting, right? Because without the connectivity between those two verticals, that small merchant may look like a gigantic risk, but understanding that supply chain connectivity and the customer and the provider implications, it means that you’re creating financial inclusion at scale, right in a way that didn’t seem possible before you were able. Run this kind of event driven real time data infrastructure, though.

Sumeet Puri 25:04
Yeah, and the impact it has on businesses as well as human life, right? So, yeah, one of the first projects I did in my life was I built the procurement system for Singapore. I was a part of the tendering team, and it was probably the first, if not one of the first, electronic tendering systems. And this is going back 2000 2001 I think. And what we observed after launching that was that, you know, the price of photocopying, of all things, the price of photocopying, the price of pencils at which the government was procuring right, went down 10 times. Had

Michael Waitze 25:39
to, almost, by definition, because, because the market that they were addressing, from a purchasing standpoint, must have gotten broader, which means that everybody was more there was more competition, so that the prices for stuff had to drop. Yeah, sorry, go ahead, yeah,

Sumeet Puri 25:51
and they’re new suppliers. As a result, you’ve got a lot of diversity in the supply chain, and competition, and competition leads to innovation. You know, everybody. Of course, there were some losers, people who had set up shop just down downstairs from the ministries. But democratization and disruption is generally a good thing, agreed. I don’t

Michael Waitze 26:14
want to get back to this idea. You talked about fairness at the beginning, right? You said GDP in the in the world has been growing, although unevenly. I don’t want to spend a lot of time talking about that, but essentially, what you’re talking about is connecting all these systems to each other. We talk about democratization. A lot of the financial system this has to whether anybody wants it to or not, or whether it’s happening purposely or not. Almost by definition, it’s going to happen because of all this connectivity that you’re creating. No absolutely.

Sumeet Puri 26:39
And again. See, connectivity is one thing, but you know when you can tune into data. So when you’re putting those nouns and verbs, so when you say payment, slash, slash initiated or order, slash created right. The power of an event driven approach, the power of event meshes are distributing it globally, if you want, or wherever the brokers are, in a one to many manner, without any performance implications, practically speaking. So that means newer participants or analytics engines or AI capabilities can just connect to it, tune in and do interesting things with it. And when you do that, the insights that you discover, the actions that you will take on the back of it, will bring efficiencies never seen before, because you have access to real time data whenever you want.

Michael Waitze 27:26
Yeah, I spent a lot of time talking about income arbitrage and salary arbitrage, and I think that over time, the things that you’re discussing are to change the salary arbitrage as well. And I do think that salary is going to refer to the mean, if nothing else, we should see salaries in the developing world come up higher to match, not to match, but to get closer to those in the developed world and vice versa, the salaries in the developed world drop because that arbitrage is going to have to close, and it doesn’t close in a unilateral way. So that’s interesting to me as well. No,

Sumeet Puri 27:56
just my I’m sensitive to the human side of it, right purchasing power parity angle. So, I mean, we want a fair world, but I hope, I hope, all of that eventually leads to us being all citizens of one big global village. And this, you know, the equality starts getting better, yeah, but we head in that direction, right? I mean, when I was growing up, I never would have imagined, why, in my mind that we’re sitting in Singapore someday, working for a global company, right, working for a Canadian company, in a very diverse environment. So it’s already happening, a product of that. I’m experiencing that.

Michael Waitze 28:32
Yeah, so am. I mean, there was no way in the world that I would ever consider that I was gonna be living in Bangkok in my life, some little kid from California and Connecticut, so in a way, we’ve kind of crossed paths, but I think to the same result, I want to ask you this, if you have any specific like customer success stories from the stuff that you’re talking about, maybe maybe improving it, their operational efficiency and even their customer satisfaction?

Sumeet Puri 28:56
Yeah, so customer satisfaction, we’ve got a bunch of them, right? So I think I’m not touched on telco as a domain, so maybe I’ll talk about that a little. So we’ve engaged with telcos in Southeast Asia, South Asia, India, because, you know, those are massive populations, and as a result, the loads on the telco systems are quite huge. Yeah, when I say load, you know, just a simple action of topping up your prepaid phone. And these are predominantly prepaid markets. We’re also doing stuff on the post paid side related to customer experience. But when you say you want to top up your phone with, you know, 20,000 rupiah or 500 rupees. We do this with Airtel. We do this with Jio in India. We do this with excellexiata in Indonesia, in the problem statement many years ago, when we were when we started working with them, was that, you know, this going to the dealer, and topping up that balance to seeing the benefits would take many minutes, sometimes half an hour, sometimes longer. Yeah. So. Uh, you’ve done the recharge, and the minutes are not getting added, or the voice the SMS is not getting added. So how do you solve that problem from a customer experience standpoint, right? And there was a public talk that the CIO of Excel axiata gave a few years ago where she was talking about how their net promoter scores were suffering because of this and as they became event driven, and, of course, with Solace. But that act of the top up is also an event when you say top ups flash started or initiated, right and then, because it’s one too many, because it’s published, subscribe and you can tune in using, even for the technical audience wildcards, the voice system would tune into the packages that have voice the data systems would have data, and similarly, the SMS system would have tuned into their appropriate package, and they would add the benefits as quickly as they could. So we parallelize the entire processing, and as a result, we are able to add these benefits faster. So think Airtel CIO made a statement that as a result of this event driven approach, 97.5% of the top ups happened in or the benefits were extended in less than five seconds, and at India scale. So we are talking 2020, 30,000 top ups happening per second. But then it doesn’t stop there, right? So that’s on one hand, the customer experience. Add agility to it. Now, the telco goes and signs up with Netflix or Amazon Prime, and they start to offer that service, right? Nothing changes in their integration, because guess what? The order event is already on the mesh. You just bring a little Netflix consumer, and that Netflix consumer tunes into the packages that have been added to Netflix, probably even in a different location, and maybe it’s a REST endpoint, and that’s it. You just start unlocking your services. So the time to market, the innovation, the ability to launch a product faster, basically comes down to weeks, sometimes less than that.

Michael Waitze 32:02
So Netflix is a really good example of this, but they’re a digital first company, right? So for them to integrate event driven stuff into their business, they had to do it just to exist, right? Yeah, but tell me this, how hard is it, or what are the challenges for a company that’s not digital first, but knows that it has to because I never even thought about this idea of the top up being an event driven thing. And I remember walking away from my telco. I pay on a monthly basis. I’m prepaid, but I still pay every month, right? And I remember walking away I missed my I missed my payment because I was traveling for like six weeks, and I came back and I was like, Oh, I forgot to pay my phone bill. It’s okay. I just go to the store and I pay it, but they’re like, give it about 15 minutes and you’ll be back online. It’s like, 15 minutes. You get my money right away, and I’m waiting 15 minutes now. It was faster than that, but they must be running again, some sort of back end event driven system too, that as soon as somebody puts that payment into the system, I get my coverage back. Yeah, yeah.

Sumeet Puri 33:01
Or if it took 15 minutes, probably is a micro batch. Micro batch, yeah. So interesting.

Michael Waitze 33:07
So, yeah, so Netflix, again, is a good example of a company that was digital first. But how do companies that aren’t digital first go through this process?

Sumeet Puri 33:15
Yeah, that’s a it’s a really interesting point, right? That, okay, I really want it. But how do I get started? We can want a company, or I’ve got API first as an approach, which which served us really well, and now it’s the time for the next wave, which is event driven, or I’m service oriented, or I’m just siloed, right? So the way that we work with customers, so we have a methodology called Event storming. Others use it as well, but we do these event one workshops, where we make you think of your use cases, where you have customer experience challenges or agility or InsurTech challenges, right? And we break it up into events. And as you mentioned, top up is an event when you open a door. That’s an event when you close the door, when you lock a door. Again, nouns and verbs. And in Singapore, we connect all the busses, and when the door of the bus opens, that’s an event, and it’s a pretty important one. We connect the entire airspace for the US, for Canada, for Singapore, for Europe. When the wheels go up in a plane, it’s a pretty significant event. Now the plane is going on cruise, and you know, you can do interesting things with it. So many enterprises are doing this in an event driven manner. And many of these industries are very legacy. So what we do is we opportunistically look for projects that are in flight, and through that for systems that were never defined or never made event driven. There are adapters. So the event mesh, the platform that we have, has connectors, or we call them micro integrations to sit at the edge of the mesh, and they will make the legacy event driven by either like, you know, making a fast micro batching architecture, or even evented characters. But then your newer systems that you’re building, or as you’re upgrading, makes them event native, so that you’re not you’re not. Just hung with adapters. You’re you’re progressing to a much more seamless data movement as you go.

Michael Waitze 35:09
I, when I was trading stocks, I used to run through Tokyo with one of my buddies, where we when we were both on the trading desk, and we would play this little game of, if we saw a sign for a company that was listed, we would try to guess it stock code, or try to remember its stock code, which we could know. Like, 7203, Toyota, 7267, is Honda, I think. And I haven’t looked at this in 15 years, right? Maybe 6758, is Sony. Like, I can’t remember for sure, but it’s like, once you am I wrong? I think I’m right, actually. But like, once you see it, you can’t unsee it. 6991, anyway, it doesn’t matter. But my point is that once you see it, you can’t unsee it. Do you have the same experience like when you’re like, driving around Singapore, walking through Singapore with your family? You’re like, when that door opens, that light goes on in the back end. I did kind of thing. Do you know

Sumeet Puri 35:54
what I mean? You have talked to my kids about it. They make fun of me. I’m a passionate person that must be running on Solace,

Michael Waitze 36:02
they’re like, Dad, come on.

Sumeet Puri 36:06
But yeah, especially like out here, we get to see it right? The port is connected. Then, you know, it’s a fully autonomous port in Singapore now, and that port, again, everything that moves on the port as an event. So for example, if a truck moves, if a truck get doesn’t get its events every 15 seconds, it has to stop because it doesn’t know whether to do, yeah, yeah, what to do, right? And that’s a very interesting proposition, because you cannot have that event go from the port finger to the cloud and back, because it’s too slow. If you really wanted the truck to stop, or, you know, the crane containers are crossing, you want to take action. The speed of light matters. So in that case, these event brokers are sitting on the port finger, so you’re doing very low latency, publish, subscribe between the various machines, but then they’re also going up to the data centers or sometimes to the cloud, because, again, one too many. The feed can go to many other systems. So your ot world is integrated with your IT world. So the patterns are, again, similar, the use cases are very varied. So

Michael Waitze 37:15
can I ask you this, when you go through this event storming, which I love, this word, by the way, event storming, whoever came up with it, they should get some kind of some kind of marketing prize. Yeah, but when you go through this event storming, you’re obviously doing it with companies that you’re not necessarily working with. Yet, has it, it must be the case that when you’re going through it and you’re actually writing, because I did this on the trading desk, right again, when fix was implemented, we could do so many more things, because we didn’t have to think about confirming the stock codes, the buy or sell or the share amounts, right? Those were already done for us. So what other things could we do? Right? Those events got taken care of on their own. So when you’re going through this event storm and you have clients, realize I never thought about that as an event, but if I knew that already, then I can do this, or give these three other choices, and they have these little like micro epiphanies as well. That maybe changes the way they deliver their product and even price their product. Do you see that

Sumeet Puri 38:04
happening too all the time? And I mean, it’s such pleasure to be in that workshop, right? Because we learn a lot. Yeah. So like this happened, for example, with Unilever, again, they wrote a beautiful blog about it, and recently, Gartner has done a very nice case study on them for supply chain, where, while they started with location events, right? So you want to track soap as an example. And that system became from non real time to real time. The they then use that in a very interesting way when the Suez Canal crisis happened. So remember when the air ever given got stuck in a canal and disrupted the global supply chain. So I think Unilever, if I remember correctly, in one and a half hours, we’re able to deal with the situation, to see which ships are going alternative paths, which are going to be left to wait. And the beauty of the events is they did not just look at the location event, or as a result, the price impact of that they also had emissions event on the mesh. Interesting. What is going to be the impact of rerouting the ships in a particular way on the carbon footprint? Because shipping is the sixth worst culprit on for emissions, if you were to compare it with countries. So that’s again, we never thought of that. But in that discussion with the head of supply chain back then, they were saying that, Oh, if location is an event isn’t port source. Change an event is it port destination. Change an event isn’t carrier booking. An event is an emission that event. And there are 52 such events, or 50 something events. I can’t remember the specifics on the mesh, just for tracking supply chain.

Michael Waitze 39:46
I want to say this, and then I’ll let you go. I’ve done 1000s of recordings for the age of tech podcast and my other podcast on this platform. And I would say at the most, like seven of the recordings that I’ve done have fundamentally changed. Change the way I’ve thought about looking at businesses. I’m not kidding. This is definitely one of them, if not in the top three, because the more I think about this, the more I think it impacts everything that we do. You’re right. Every little step you take is an event. Like when I go downstairs in my condo, there’s a monitor up above that that unlocks the door without me doing anything. That’s an event. So that means that some system knows that someone’s coming in actually going out, because it knows where I’m located as well on the way in, when I swipe my card. That’s an event because it also knows now the door’s opening the other way because I’m coming in. Yeah, it’s just gonna fundamentally change the way any business thinks about the way they do business when they unlock everything into these little pieces. Again, I told you I did this on the trading desk. I was the one who actually sat down when we implemented fix and said, all of these things are going on. What of them can we eliminate from having to think about and then during the time when we don’t have to think about that, what new things can we think about to make our business more efficient and more profitable? And I’ll tell you this once we once just just, you can understand impact, or so the listeners can understand impact. We took a business at Goldman Sachs, and by putting technology onto it, a lot of it event driven, and took it from an $8 million year business to $135 million year business. It matters.

Sumeet Puri 41:15
Impressive. Yeah, and see you’re absolutely right, if you change your thinking, the impacts could be pretty dramatic, right? So I always encourage people to think eventually, and there’s this good reason for that, right? See, we all, especially people who grew up as programmers, we almost have to unlearn the way that we were thinking for a very specific reason. For the longest time we were taught, and we still are taught because that’s the right thing to do. Function calls right? We will talk modules, Fortran seven, seven. We will talk go to statements like it’s all about, how do you break your core into smaller pieces? And the modularization from function calls became remote procedure calls or objects, and then that evolved into APIs. But what’s common in all of that is, they are point to point, one to one interactions, right? So we’ve it’s ingrained through school, through most of us, unless you experience this on the trading floor, and we didn’t invent it. I mean, there were, there were giants who’ve innovated topic routing and publish, subscribe or subject based addressing. We just do it pretty well. I would, I would argue we’re the best. I mean, that’s why I work here. But, but that whole thought process of point to point interactions has taken us to API first, but now, with all the physics changing and systems becoming distributed, we can think in more natural terms, like, when it rains, it’s an event, it’s not an API call. So if you think of it differently, what can you do if you listen to that event, and then, technically, the mesh moves those events, and then also what I didn’t talk about in like as we reimagine our integration, and that’s our theme, you have to catalog those events, and we have the ability, from our perspective, in the platform where, okay, the the event exists, nobody knows. So like you have an API catalog, how do you and look for events? How do you go and then use those events? Right? So that’s the entire ecosystm, events, move events, catalog. Events, unlock innovation.

Michael Waitze 43:20
That is the perfect way to end. Sumeet Puri, the Chief Technology Solutions officer at Solace, I think I said it better this time, but there are a lot of S’s there. Thank you so much for coming on the show. I really appreciate your time today.

Sumeet Puri 43:32
My pleasure. Michael, thanks for having me, and I really enjoy the conversation. Hopefully see you in Bangkok sometime or in Singapore, absolutely.


Latest Episodes:

EP 326 – Never Forget Who the Enemy Is – Gene Yu – Founder and CEO at Blackpanda

EP 326 – Never Forget Who the Enemy Is – Gene Yu – Founder and CEO at Blackpanda

“Resistance training is, you know, they can take away everything in terms of like, your comfort, your warmth, your food, you know, etc, make you disoriented, but they can’t take away your will to resist, right? And they can’t take away your mind, right? That’s the thing that you can always control in terms of resistance. And so I emphasized that to her what they taught us over and over, which was, never forget who the enemy is.” – Gene Yu

In this episode of the Asia Tech Podcast, featuring ⁠Gene Yu⁠, a Founder and CEO of ⁠Blackpanda⁠, we explored innovative approaches to cybersecurity and the vital role of crisis response. Gene’s unique background as a West Point graduate and a former Green Beret significantly shaped BlackPanda’s approach.

read more
EP 325 – Building a Marketplace and Democratizing M&A for SMEs – Marcus Yeung – Match.Asia and SEAbridge Partners

EP 325 – Building a Marketplace and Democratizing M&A for SMEs – Marcus Yeung – Match.Asia and SEAbridge Partners

So it’s actually very interesting how technology is taking a lot of the grunt work out of investment banking, which allows the real investment bankers to focus on the important stuff, which is, you know, the negotiations and the closing of the deals and more personal stuff. There is a good company in Japan, which has a good precedent called M&A Research Institute, which is a listed. And within it’s a listed company, it’s a similar business, and we’re following basically the same path as they have. And they closed 150 deals last year. And they made $60 million of revenue. And it’s a very profitable business model. – Marcus Yeung

read more