DigiTwin - Premium Customer Experience with the “Power of Now”

Kevin Keller:

Hello everyone, and welcome to the HunaTek and Aerospike webinar on Subscriber Data Management Innovation with DigiTwin. Enhancing Customer Experience & Delivering Business Value. I'm Kevin Keller with Aerospike and I'll be your moderator for today. First, I want to go over a few administrative items before we get started. First, this webinar will be recorded and can be accessible for download at a later date. Also, if you can ask questions, and it's interactive, and you can engage with others via social media using the icons that are available on the portal. And you can also download the slides at a later time as well and share those with others. Questions that will be addressed. They will be addressed at the end of the webinar session, so please submit those via the portal, and we'll have the HunaTek and Aerospike team members answer those questions.

So let's go ahead and get started. First, let me introduce the Aerospike and HunaTek team members. So, today, we have the presenters. They're from several different regions. And we have Amit Raj. He's the APAC and Middle East business manager at HunaTek. We have Elaine Fletcher. She's the Americas, UK and Europe business manager at HunaTek. And we have Shahed Mazumder, the global director, telecom solutions at Aerospike. Again, I'm Kevin Keller, your moderator today, but also serve as the global director of partner solutions sales at Aerospike. And we're glad that you're all here, and today we're going to cover a few things.

So you can see the agenda on the right-hand side of your slide, and the topics we're going to cover today really include, one, just providing an introduction and an overview of both Aerospike and HunaTek. Then we'll explain DigiTwin itself. What is it, why is it important in today's digital economy? We'll provide an overview of DigiTwin from a technical perspective, and explain how Aerospike is a difference maker in this type of architecture. We'll also provide a real subscriber data management use case, and specifically focused on Customer 360. And this use case was initiated about five years ago. And it's continuing to provide value for tier-1 telco today. So, it's important to note that DigiTwin demonstrates a more modern and new approach in subscriber data management. And this is where organizations are focusing on digital transformation, and really taking advantage of the 5G economy that we have and that's developing today in the markets. The team will describe how HunaTek and Aerospike can also help you and your organization achieve success with DigiTwin.

With that, I want to going to go over a little bit more information about Aerospike itself. Aerospike is a company that's been around for about 11 years, and it's originated in California, it's a Silicon Valley startup. The founder, Srini Srinivasan, he's also the current CTO of the company. The company itself has global operations and capabilities with offices in the US, and the UK, in India and in Israel. And the company has been experiencing exceptional growth. As customers are really realizing the value of real-time data and the growth of data that's just really expanding in the market today. And really trying to determine how do they use this information and getting the accessibility of this information to make real-time business decisions. So that's becoming more and more important for customers as they're trying to compete in the market today. So, with that, Aerospike is the leader in the real-time data platform space.

And just to give you a little bit more about what is Aerospike and why is it the leader. So there's some important elements of Aerospike. That's why leading companies are using Aerospike across many industries across the globe. And the reason for that is pretty simple. Aerospike is the real-time data platform that combines performance at scale. And it does this with high availability and reliability. And it can even offer strong consistency as a feature with performance at scale, which is really unmatched by others in the industry. So in addition to that, it also has unmatched TCO. And so, there are fewer nodes or instances that are really required to run Aerospike, and that's really where it also differentiates itself in the market. So, in doing this, there's some efficiencies that organizations can benefit from. Not only from reduced costs because there's fewer nodes, fewer instances. But also, Aerospike can help organizations achieve resource efficiency metrics that they may be trying to strive for, whether they have green initiatives, et cetera. And Aerospike can help fit into those objectives as well.

This technology has proven that Aerospike has... There was over 200 customers worldwide across multiple different industries, in adtech, e-commerce, high tech, financial, gaming, and even telecom. And they incorporate Aerospike into their mission-critical applications. Whether that's fraud, money transfer, recommendation engines, customer data platforms, and subscriber data management. Now, I'm going to hand it over to Amit to provide a little more information on HunaTek and DigiTwin itself.

Amit Raj:

Thanks, Kevin. I'm going to talk to you about HunaTek, just to give you a brief overview of HunaTek and our relationship with Aerospike over the years. We are based out of India which is where our head office is. And then we have offices in UK, which runs our Europe and worldwide operations. Obviously, as customers, we work primarily with the telcos and banks, and the digital-first companies. They tend to be the most voracious and most exceptional users of data, both at scale and in terms of volume. It's a natural fit for us that we work with those two industries.

Employee-wise, we are growing very fast. We are about 600 people, and have worked on multiple Fortune 10 clients, delivering a variety of use cases. From fraud, real-time monitoring, payments engine, sanctions. And in terms of telcos, you will see us working on revenue assurance, subscriber data management solutions, and in the areas such as analytics and some of the other dashboards kind of things. In terms of our full-service solutions for our customers, we work obviously with Aerospike, but along with Aerospike, we use our full stack and analytics engineering expertise to deliver the solutions, which encompasses Aerospike and provides a complete suite of business capability to a customer, like we did for Customer 360 that Kevin earlier mentioned in the opening remarks.

I will just go to our relationship with Aerospike. With our natural affinity for the data and the DNA that HunaTek had right from our inception, a decade ago, We have been working with various solutions. And about five years back on the back of tier-1 telco opportunity, we set aside time to jointly engineer, build, and deploy, a DigiTwin solution, which was for subscriber data management, allowing customer to do their Customer 360 view for all their lines of business. My colleague, Elaine Fletcher, will talk some more about it in the later amount of data slides. This subscriber data management obviously leveraged Aerospike and its capability right from the get-go to deliver some exceptional business value, and has been running for the last five years in a always on, always available paradigm with our customers. Obviously, we work with Aerospike in all the geographies, APAC, ASEAN, EMEA, and Americas. And later on in the Q&A, we'd be more than happy to take up questions on how we can add value to your journey regarding our solutions. Elaine, off to you please.

Elaine Fletcher:

Hello. So I'd like to talk a little bit about what DigiTwin actually is. So it's a prebuilt, highly configurable subscriber data management platform. And you can use it to drive interactions across all of the channels that you use to communicate with your clients. It allows you to really highly personalize each of the interactions, in fact, right down to the segments of one. So when each client can see a different interaction and you can really script in a very highly configurable way. We don't believe in writing code or you writing the code. We've prebuilt everything and put a very highly configurable layer on top to allow you to do very rapid personalization.

So why did we do this really? We did it to solve two basic problems that we're seeing across most of the customer base that we're working with in the telecoms markets. And that's to solve the dilemma that subscriber-facing digital apps need ultra-low latency, definitely real time. But the majority of the investments, and this is massive levels of investments that the telcos have in OSS and BSS systems, very few of those applications that are mission-critical will actually support the ultra-low latency that is required for the subscriber-facing digital apps. So, when we've solved that problem, the other problem that we've set out to solve is in today's market, it's not just good enough to deliver your direct services to your clients. What you really want to do is allow your business executives to do rapid business innovation, so that they can monetize third-party propositions and take those directly to your clients.

Now, the instance that Amit mentioned a moment ago, where we've been supporting a tier-1 global telco for the last five years, it definitely isn't a stable solution, and what is running now is what we built five years ago, because literally every two weeks, the different business domains within that operator, they will put a new release together that will go out. So new direct services, new-third party services, lots of things get released as complete new services. But also they're constantly innovating exactly how each personalized interaction to each of their different client base actually happens. So, you'd never recognize that system from basically one to the next. There's so much innovation going on there, and that's really allowed them to stay ahead of their market.

So how have we done that? Basically, DigiTwin sits as a layer between the subscriber channels and the enterprise channels on the left-hand side of this picture, and your back office OSS and BSS services. So at the heart of it is Aerospike, where we've built our own in-memory subscriber data and transactions repository, so we have our own sort of conformed logical data model there. And we've wrapped that with a very thick layer of integration, because it's not just good enough to plug things into a data model. There's a lot of heavy lifting required to take the data from the subscriber portal, et cetera, into the various third-party applications and OSS/BSS apps that you have. So I'll talk a little bit more detail in a few minutes about that rich layer of real time and batch integration and transformation. But basically for the moment before we talk about those details, I'd first like to spend a quick minute just sort of examining why. Why would you bother with DigiTwin?

So first of all, the most important thing for us is that DigiTwin improves the customer's experience. So we really do set out to simplify the omnichannel customer journey. And that for us is the most important because their customer experience will reduce churn. The next point really is the focus and precision to empower marketing to allow a personalization of experience. I mentioned earlier the segment of one. So what we've basically got there is a queryable customer database, not only for operationally delivering those personalized conversations, but also the fully queryable customer database allows all of your back office analytics, artificial intelligence, machine learning, et cetera, to run at a massive scale to look at everything that that customer has done as they've interacted with you. And that can then drive the decisions and the automation as to how the next conversation will go with that same customer. So, it's really about sort of closing that loop really between each of the experiences the customer has with you on any channel, and then preparing for the next conversation that you have with that customer.

So it's important to be able to do that. We have to obviously have extremely high performance. So for one example, the client that Amit mentioned earlier on, that's now been running for five years with vastly reduced latency. So where typically 50,000 transactions per second is what that client has been running now in full production for five years. Now, if we look outside of that client and wider across the whole of the client base that we've been working with together with Aerospike, we're typically seeing 200% to 400% performance increases for the same workload. So, really that depends on what the workload is going to be, but what we've generally found is the heavier and more complex the interactions in that workload, then the more towards the 400% of improvement we are.

Legacy offloading as well is an interesting approach, because we're typically seeing a cost saving there of around 10k a day in UK pounds because I'm English. Now, what some of our clients have done is they wanted to use DigiTwin to improve customer experience, but where did they get the budget from? Well, legacy offload savings has actually provided the budget for several of those people to be able to actually purchase, deploy, and operate DigiTwin. At the heart of our design decisions when we produced DigiTwin is rapid integration. We don't want anything to be hardwired, because we know that as soon as anything is coded in a hardwired fashion, your business is constrained, we've put you into a straight jacket. And that's no good for your business. You've got to be able to move really quickly to compete in the market you're in. So, where we've put a lot of investments and work behind producing a whole big suite of configurable APIs for reads and writes, so that means that the integration happens in days versus months of coding.

Shahed will tell you a little bit more later, but just to give you a perspective on this, the global tier-1 telco that we've mentioned that does 50,000 TPS, et cetera, that we've mentioned several times, they actually went live from start to finish first time after three months. And most of that three months was actually going through their boot to live, the gated process, et cetera, that the operator would have. So, you can see this is not typical of these kinds of SDM platforms, which would typically take 12, 18 months, maybe even 24 months. We did that in three months, and it's been in successful production ever since then. Yeah. Also, what's important is not just the day one getting you live in days. It also means that because we have a repeatable configurable solution, not a whole load of code writing, we've got a whole suite of reusable components. So, that means rapid business integration releases every two weeks, et cetera, to bring new partnerships to market, change conversations with clients, et cetera, basically just to stay ahead of the competition.

So, the next slide. There we go. So I'm going to talk a little bit more about how we do this. So, I guess everyone on this call listening to me now is saying, "Well, there's a lot of promises you've made there. You said a lot of big things, but we live in the real world and we know integration requires significant heavy lifting, lots of transforms, mapping, and also all of these conversations with every single customer needs absolute transparency. So, you've got to have a full audit log, you've got to be able to prove back to the regulator exactly what happened. When your customer care staff have a customer on the phone with a query about their billing or their whatever, that customer care agent needs to be able to have a look back and see what that client has been doing, and all of the different interactions that have happened with that client, and do that very quickly and easily."

So, what you'll see on this diagram, basically at the heart of DigiTwin is our conformed telco data model. So, you'll see us running across multiple clusters there, and there's obviously full load balancing between all of those activities on those clusters. Now, on the left-hand side of the diagram, you'll see the full frameworks there in blue. Those are what we're using to ingest data from absolutely any sources that you want to work with, whether they're your sources, third-party sources, et cetera. Now, you'll see there that we can read data in and we can write data. So we can take data in a real-time fashion, we'll validate that data, we'll transform it automatically in a configured way into the DigiTwin conformed telco data model. We'll do all of the data mappings, et cetera, and all of the interfaces into Aerospike, and we'll have a full logging and reporting facility against that.

We also realized that a lot of the data you'll want to put into DigiTwin is actually day-to-day data where we'll always have to reconcile automatically between the same data and how that looked yesterday. We need to reconcile that, so we'll unload the data first, we'll compare it with the data from that same feed yesterday. And we can resolve differences in a way that you've instructed us, if you've given us some business rules for resolving, et cetera, and comparing, and all that reconciliation. Then we can automatically do that resolution, and we can log everything and report everything, so that people can see what we've done. Migration framework, that is also important, because a lot of the data sets that you will be given by third parties and indeed internal ones as well, will come in bulk form. So, from databases and from files. So, again, configurable business rules here for validation, transform, mapping, et cetera, and full logging and reporting on everything that we're doing.

And over on the right-hand side now, what you're seeing on those two APIs are the online usage. So here, we've got some rest services, so you can do primary searches, secondary searchers. We'll do the subscriber checks, et cetera. And so, all of the things that we do, as you've probably guessed, you can figure up what we do, how we do it, et cetera. And not just for the straight through happy path, but all of the exception handling as well, everything is configured. The online write framework as well, it allows us to load the attributes, validate everything, pause everything, and again, subscriber checks against all of that. So, what you can see on the diagram is basically the data going in and out of the DigiTwin database.

What you'll see on the little bubble at the bottom of that, isn't the actual data going through. It's basically your own staff working with the Grafana instrumentation and the EFK stack, to actually look at the audit logs, look at the TPS load. So basically to operate the beast. Yep. So, the full audit log that can be queried. For example, customer care services and AI. So, just to recap, basically, you'll see here for any type of data, be it online, batch, day-to-day reconciliation, anything, we'll do full user validation, business validation, mapping to the conformed data model, all transforms, all reconciliation, et cetera. So, all of this heavy lifting that you're probably quite rightly in your head thinking, "That isn't easy. That can't be done in real time," it absolutely isn't easy, but it absolutely can be done in real time. And we would really like the opportunity to prove that to you.

And my next slide tells you how we can prove that really. So it's a call to action. We're very proud of what we built here and we really want to be able to prove it to you. So we would love the first step for you to just book a demonstration of DigiTwin, have a look at it. If you like what you see, there's a very simple second step, and that's basically a PoC. So, if you can get sample data to us, then we know that within 24 hours, we can deliver a PoC to you.

So the hardest part of that really is about agreeing with the customer use cases and the success criteria that you'd like to see on that PoC. And if you're anything like the rest of our clients, you'll be thinking back, "Well, there are some of my things that are easy today, and there are some of the things that I really do stay awake at night about because they're not the trivial ones." So you'll give us the latter ones, so give us the hardest and most challenging use cases, and the most stringent success criteria, and watch how we do that.

So what we'll do is we'll then configure up DigiTwin, load your customer data, and then demonstrate the functionality and performance of DigiTwin to you. Because like everybody else, I'm the same, seeing is believing. I don't believe people's words until I've seen things. So what I'd like to do now is introduce Shahed, so that he can explain more about what's going on in the heart of DigiTwin. And that heart is Aerospike.

Shahed Mazumder:

Thank you, Elaine. So hi, everyone. So you have heard from Elaine how Aerospike sits at the heart of the DigiTwin solution. But I thought I would take a step back and spend a few minutes to explain a little bit more about the Aerospike Real-time Data Platform itself. So from the left, you can see that we can actually take the data feed pretty much from any sources. From the telco context, from TV, video, even from the internet sites, social or web usage, or even in connected devices ecosystem. We offer a set of pipelines, the connectors to get the data feed into the data platform itself. So, you see under that box, we have listed up the connectors that we offer like Kafka, event stream processing, ESP, or Pulsar and JMS as well.

The next block shows about the data models that we support. We started as a key value source. Of course, we support that. But over time, we have added capability and use cases for DocumentDB or graph, or even time series use cases as well. Next to it, you see a box for storage. So this is again at the center of the data platform. So in terms of how we support or how we house the index and the data, so we actually offer different combinations. From the left, you see in memory, that means in this case both the index and the data will be on DRAM. So typically we see these use cases being deployed at the edge where you need really, really ultra low latency. Next to it is our hybrid memory, which means the index will be housed inside the DRAM, the memory, but the actual data will be in SSG, the flash drives. That's done through HMA, our patented technology.

HMA stands for hybrid memory architecture. More about that later. But then we also actually offer combinations, the next two combinations, where both the index and the data can be housed in persistent memory, PMem. And also in all flash. Everything in SSD, which gives you massive scale, but the cost goes down significantly, because you're not using DRAM or PMem, you're using all flash. So then about our deployment models. So the block above that storage block shows you where and how we get deployed on the edge in particular.

Here, you see that we support different deployment models including bare metal, but also different clouds. We support multi-cloud as well. So for AWS, Google Cloud, Microsoft Azure, we support all of those. Private cloud as well, but in the data center and on-prem, of course we support those too. And we come in terms of our client side, we support applications from, say, Java, PHP, Python, all of those. In terms of use case, of course I mentioned earlier we started with edge data system deployment, but over time we have seen use cases for system of record, the core data based deployment use cases, but also reporting and other use cases too.

Next to it, under the processing box, you see other connectors that you offer, Trino, Presto and Spark, which can be used for doing the post-processing. Things like your AI ML processing that you can use from the data in the SOR, system of record, but then take the data to your AI ML engine using the spark connector for instance, that we offer.

So then let's look at what problem do we solve here. So on the left, we see some issues with traditional first generation NoSQL systems. Oftentimes we hear about concerns like lack of strong consistency or high server count, and scalability issues. Whenever we encounter some customers or prospects who are using Cassandra customer for instance, we hear a lot about high server count on the footprint issue. In terms of scalability, we have heard a lot from our customers or prospects, talking about their experience with Redis, for instance, in terms of scalability. Moving on, the next block in the middle shows the relational systems. Over there, the issues we have encountered are low throughput, high latency, and again, the scalability issues. And those are common concerns, including the distributed relational systems that we see in recent times. Last but not least, to the far right, there are the back office systems. Things like IBM DB2 or Oracle. As we all know, well known issues there are the cost of course, and then also the internet scale. So scale is across the board one issue we see coming up again and again.

So then let's look at what exactly we are doing in Aerospike data platform to specifically alleviate and target those issues. So going from the top left, I mentioned earlier about our hybrid memory architecture where in a default setup we use DRAM for the index, and SSD for the data itself. That SSD actually is optimized to work as high performance manner as the DRAM, so that you don't compromise on the performance side, but you get a shrinkage of footprint, and also of course on the cost side as well. So that's essentially the first bullet says, DRAM performance with persistence. We have actually shown a very high node density. So here you see 50 terabyte per node. Recently we did a benchmarking study with AWS and Intel where we achieved one petabyte level of data with only 20 nodes. So that translates into 50 terabyte per node.

The next one underneath that is the strong consistency. So this is about transactional robustness, similar to what you see in relational databases. Obviously we get deployed in a lot of use cases including financial services industry or even in telco context. Also, sometimes strong consistency is a key requirement, so we support those requirements. The third one underneath that is the intelligent clients. One of the reasons why we are so fast and low latency as a platform, is because we always have just one hop, a single hop to the data. Because the smart client knows where exactly in which node, in which location the data is housed.

And so that's essentially first class participant in cluster. And in combination we support that. Again, as I said, the result of that is very, very low latency for the reads in this case. Moving on to the right side, you see the dynamic cluster management. This is an umbrella term we use for a few things. Those are listed underneath. Dynamic rebalancing, self-healing, rolling upgrades, elasticity, and self sharding as well. The last but not least, we wanted to mention cross data center replication, XDR feature that is a very well-loved and well utilized feature of Aerospike data platform. This is actually used for cross cluster replication, hot standby, high availability, and dual redundancy like disaster recovery type of use cases. This is again, a replication solution can be used within clusters across sites and zones.

So let's switch gear here. And instead of we saying what's good about the Aerospike data platform, let's look at from the other side. What our customers are talking about Aerospike. Why they choose us, what they like about us. So on this slide I would like to introduce customer input from APEC region. All of them are tier one telcos. And the first one here is a customer based out of APEC again, as I mentioned. The use case over there is Customer 360 and also close to 20 other use cases. So when we had this conversation with them recently, they mentioned a few things why they like Aerospike.

So if we go from the top, you see stability and reliability of the platform. That came up as number one for them. Then they mentioned about lowest latency versus the peers that they compared us against. The third one, they mentioned about XDR, across data center replication that I was introducing in the previous slide. They also talked about favorable product roadmap that Aerospike offers, and the transparency which we share with them. They lauded our accurate sizing, good document support, and last but not least, they talked about great relationships at all levels. The quotation actually came underneath from actual user of the data platform. He mentioned that Aerospike gives them zero operational headache. And they also specifically talked about the XDR, how much they love this feature.

Moving on, the second reference, this is also a telco in APAC region. The use case over there is the backend database for their nationwide narrow band IOT network, for the HSS. They mentioned high data availability as their number one. Beyond five nines, actually. They also talked about low latency with consistency, they like our distributed architecture. They also talked about XDR feature for the geo-redundant deployment that they use us for. And then last but not least in this case, they mentioned about the report generating capabilities of Aerospike, which comes up in minutes, versus end of day, which they used to use before Aerospike was implemented in the solution. And the quotation underneath came from that angle. Actually they mentioned that when they tested with three to 4 million subs, it took only 30 seconds to generate the report, versus the much longer time in order of magnitude that they were used to, before Aerospike was introduced.

Then the last one here is another tier one telco in APEC. The use case over there is, we are the engine for their real time microservices. And there are a few others, but this one is the main one. And what they talked about as the most important factor is scalability. Then they talked about the high volume for the campaign management use case that they use us under that microservices platform. They mentioned that we were able to support 80 million records in one campaign versus 15 million, which was done by the previous solution. They also talked about the support for concurrent sessions with only six nodes and 15k per node. We supported 90k concurrent sessions, versus they found that RDBMS relational database system maxed out at 5K sessions.

Interestingly, they mentioned another thing, the last bullet here, that we are offering WhatsApp group chat for support, which is super fast and super convenient for them, for their engineers to use. And you actually see actually a very, very encouraging comment from the director in charge of that streaming data platform. He mentioned that, "As we transform us..." us means that, telco, "... into a digital service provider, Aerospike is like our central nervous system." So very core to their capability in the transformation from CSP to GSP.

So let's actually do a double click on Customer 360 and then see what's the connection between that and our topic today, the DigiTwin, and how we are essentially supporting the twin reference use case, which is Customer 360 in this case. The next two slides that I'll be showing you, actually the content is drawn from public presentations made by the tier one telco I was talking about earlier, who is our customer for this deployment. So they introduce Customer 360 as their digital brain for realtime personalization. So what do they mean by that? Essentially, let's look at the challenges and the goals that they talk about. So under the challenges, essentially this is unifying the data from various business services. I mean this telco, multi-service, multi-access, telco, I mean of course they use or offer wireless services, but they also have wired line offerings and other services. But the struggle with having different data sources for different business lines.

So that is the starting point. Then they talked about upgrading to a highly accessible and responsible architecture, for scale, for speed, and for agility. And why? That comes in the third bullet here, because they want to develop a unique footprint for the customers and combine different attributes like geographic information, and demographic information, which as you can imagine will come from sources like HSS. Or other parameters like usage of the apps, like Netflix, like Amazon, like Uber. All those things. And all of those are important in the context of delivering better customized experience and services. So essentially that translates into the goals that you see on the right side here. Better understanding of individual customers and their needs, which we talk about segment of one. Another way to project that. So essentially what it means is single view of the customer by combining all these sources that you talked about on the left side. Their voice service and planning information, their demographic data, messaging usage, their emails, their app and social media usage, and things like their purchasing behavior, like as I mentioned earlier on Netflix or Uber, or other third party apps.

The overarching goal here is to increase the customer engagement, to enhance the stickiness, which of course will reduce the churn, but also to drive up ARPU, because with the interaction of 5G and new CapEx, new investments being made on the network side, of course you have to look for new revenue through new business lines. I would like to actually repeat the comment that we heard from the global CIO of that customer. This is a public comment again, she mentioned, "We said we are going to take all of these 350 million, regardless of which line of business they belong to, and create a single repository which will actually power what we call the digital brain of the company. It would also give us a complete 360 degree view of the customer." So I think this is a very nice summary of what exactly they're trying to achieve here and did achieve, through our work, through the work by HunaTek and Aerospike together.

So this slide actually is a more kind of a vivid representation of what we talked about in the previous slide. On the left side of the slide, you see what the telco is talking about. I mean keeping the customer at the center, they are combining different lines of service like their wireless usage, their satellite usage, all of them. On the left side for left bubble, you see, I mean who are included in the family package. And then interestingly, they also track the circle of influence. Where is all the influencers? Where are all the influencers that an individual interact with? Then possible other sources as well, like from B2B side, and also who is paying the bills and all those kinds of information. So with all of those combination happening and giving you a Customer 360 view, what happens? One example is on the right side here.

So that Telco has a My Telco app as with any CSP, any operator out in the world. But what's interesting is in that My Telco app, everything is hyper-personalized and customized. Nothing is coded. Every single piece of text, image, or offer, is completely unique and updated in realtime. So that means if customer one opens up his or her My Telco app, versus customer two opens up his or her My Telco app, they will see different landing pages, to start with, because that's fed by that customer's personalized information that the telco has by combining all the data feeds we showed, I talked about, so far.

So this is going to be my last slide. So just connecting the dots of Customer 360 and DigiTwin, because we mentioned earlier that Customer 360 is the reference use case of DigiTwin, because DigiTwin conceptually is not a new thing. We have been doing it for five years, we have been hearing that from our friends at HunaTek earlier in this presentation. So it has a proven track record. What we're doing this time, is essentially making it generally available, and productizing it for the greater market. So just to recap what we do and did for the Customer 360, on the left side, we combined triggers from real time, all the postpaid, prepaid broadband, DTH, different lines of services and Shelly, but also from the third party sources as I mentioned earlier, giving example of Ubers and Netflix usage, all those things. But then the data lakes underneath that, all the BSS information, but also from the network side, the OSS information.

So at the heart, again, the green box is the DigiTwin solution, which is serving as the digital brain for Customer 360. You have seen earlier from Elaine's presentation that two clusters of Aerospike sitting in the middle in two geographic sites, XDR cross data center replication working between them. But then on both sides, for in and out, we have APIs and connectors, and also the integration and transform layers. With that, the telco is able to do an array of things. Some of those are listed here. Campaign management, omnichannel, next based action, digital advertising, and personalization. At the far right, we have listed up the Aerospike difference. Again, I'm not going to repeat all of those. You have come across all those things. Just to highlight, more than 100 billion records per day, five nine reliability, very low latency, average seven millisec, and 90% reduction in TCO. So that's what we offer and we enable at the heart of the DigiTwin solution. And we'd like you to have a look on our solution. With that, I would like to hand it over to Kevin.

Kevin Keller:

Thanks Shahed, and Elaine and Amit. I think this has been very interesting to really see how the power of Aerospike as a real-time data platform for a modern subscriber data data management solution can really be implemented, and sort of the inner workings of what that consists of. And so it's very impressive for HunaTek, was able to put all these pieces together and develop and deploy such a significant solution to show business value and benefit to their customers. And the team is ready now to really share the solution with other service providers or organizations. So that really means what's next. So what's next for all of you out there? And just to sum it up, I think the first step is, one, is make sure you're contacting and reaching out to either HunaTek or Aerospike, and we'll get you to the right resources. But with that, we'll have just basically a consultation. And that can lead to a demonstration of the current solution, DigiTwin, and see how that aligns to your specific business and your use case.

And we can discuss how that fits into not only maybe a Customer 360, but where it could fit into your broader strategy around possibly 5G, what you're trying to accomplish there, as well as just your broader subscriber data management strategy in general. So it's also important to point out that this is pretty easy to set up a proof of concept or a proof of value to really demonstrate what the capabilities are of DigiTwin. And so that's something that can be established fairly easy, low barrier. And we can use your specific data that you want to test out maybe actual real data, or we could use just some conceptual data, and model it as it needs to be for that type of PoC.

But again, a low barrier, it's easy to actually get that tested out, and we think you'll definitely see a good benefit. So in that, that can also be to call out, as we mentioned, this solution can be deployed in many different types of forms. So whether we're going to deploy it in all cloud, whether we want to deploy it on-prem. so there's lots of options that can be considered and has been proven out. So with that, I just want to say thank you all for attending this webinar today. And now we're just going to transition over to Q&A. And we'll just give it a few minutes to gather our questions here and then we'll address those.

Okay. So we have a few questions that have come in. Let's see here. Okay, the first one that I'll ask, this is from Jordan, and he asks, "We want to be able to launch a new feature for customer facing apps regularly, to stay ahead of the market. Can your clients do a new release every two months?" So that would be a question that HunaTek can answer. Elaine or remit?

Elaine Fletcher:

I'll do that one. Yeah, you right, Jordan. It's not just about getting live in the first place. It's about being able to change and stay ahead of the market every year that you're in production. So by all means, it would be your choice If you want to do a new release every two months. There's absolutely no reason why you wouldn't do that. And I know it sounds unthinkable at the moment, but hear me when I say that. One of our most innovative clients actually makes a new release every single two weeks, not two months. It is highly agile, and to be honest, the solution is as agile as you want to use it. That particular client, they're not a small simple client, they've got 450 million customers across 18 countries. So it's quite a challenge to deliver something new to all of those clients every two weeks, but they've done it, and they've not just done it once, they've done it regularly every two weeks for five years now. So they really do push their innovation to the limit.

So how do we do that? We've got a reference customer data model. It's got a set of 360 attributes already designed. And as I mentioned in the presentation, we come with a full suite of APIs in our framework that allow us to ingest your data and also to query everything. So all the queries that we have in that framework and that data model, they're all fully optimized for performance. So we know a lot about the technology. We are very, very familiar with Aerospike, and getting the best out of Aerospike. So you wouldn't need to develop that deep, deep expertise, because that expertise is already there. We've been running with it for five years now, we've learned a lot along the way. And all of that knowledge and experience is embedded in that framework and in that data model. And that comes out of the box for you. So we will get you live with that on day one, and then after that you innovate as often as quickly as you want to innovate. It's business rules driven, tell us what you want to do and go ahead.

Kevin Keller:

Yeah, great. Yeah, so definitely much sooner and quicker than two months. So that's a week's, today's timeframe, is pretty impressive. So we have another question that's from Emma, and so she asks, "What sort of end use cases do you think will make telcos seriously considered DigiTwin?" Shahed, do you want to maybe take that?

Shahed Mazumder:

Yeah, let me take that. Very good question, Emma. So let me give the short answer first. So it'll be relevant for any new applications or existing applications that the telco wants to introduce or already have in place, which needs very fast access to data, subscriber profile data for in and out. So that's where digital would be an ideal fit. Now let me try to answer that in a slightly longer format. So taking a step back, so when you say DigiTwin, digital twin of what? So it is the digital twin of HSS or other subscriber profile data store, which can be inside an app server like any data source, like billing charging, or any other app. So because what we're doing here is essentially making a mirror database by pulling in that subscriber profile attributes and giving a single point of query from the application side where they can query the data in a very fast manner.

So these days we hear a lot about to MAEC, Multi-Access Edge Compute and MAEC is all about low latency. So we talk about low latency of network, low latency of computation, low latency of data processing. And that's where DigiTwin, our solution, will be relevant. Because we are cutting an order of magnitude in the access time for the data processing side. So think of MAEC use cases such as online gaming or IOT applications, or image or video analytics. All of those things are like data heavy edge use cases which need very fast access to database. So DigiTwin will be an ideal fit for those kind of applications. Sitting in front of all the existing data stores within the network.

Kevin Keller:

Yeah, I totally agree. I mean, I think there's a nice fit. There's many different use cases specifically for a telco. Yeah, so thanks Shahed, for clarifying all those for us. Let's see, we have time for a few more questions. Let's see. Let's go with John. So he asks, "All my staff uses SQL. How would we go about designing a data model for Aerospike with this DigiTwin solution?" And I don't know if somebody maybe from HunaTek maybe wants to answer that, or from Aerospike. But I don't know, Elaine, I think you were talking about this earlier.

Elaine Fletcher:

Yeah. Yeah, I thought about that. Weren't we? Just a few minutes, actually.

Kevin Keller:

Yeah.

Elaine Fletcher:

So for the most part, our clients don't actually want to roll their sleeves up and build it. That's why they're buying a solution, a platform, as opposed to just building a bespoke project. So what most of our clients do is they take our reference model as the start. So we have that full logical data model there, then we have all the mappings and everything. So what we need you to do, your job, is basically to give us the business rules that you want us to apply, to validate your clients, to map things into the data model, to do the validations, do the reconciliations.

So it's basically getting us underneath the skin of your business so that we can configure up DigiTwin to be unique to you, and to basically represent exactly how you want your business to run. It doesn't mean that you've got to get deeply technical and have acquired a lot of skills in the actual data modeling side. Sure, you can if you want to. So far our clients haven't, but I mean there's nothing to stop you if you want to do that, but I guess my message is you don't need to, and nobody else has so far. So yeah, don't worry about that.

Amit Raj:

Yeah. And Aerospike makes it extremely easy. I mean, our methodology starts with identifying the read patterns that your applications, your third party users of data, need to access the data at speed and velocity. As Shahed was telling earlier in the architecture of Aerospike, the tough problems of storing data at scale and querying it has been solved by Aerospike. So you don't have to do the normal low level stuff around partitioning and so many other things, thinking about how data is going to get replicated, et cetera. That problem is solved.

So we study with the access patterns, the usage patterns. And thereafter, few tricks in the toolkit about a secondary indexes or having kind of mapper tables. These are very straight patterns, especially for a telco universe. So we would be able to tell you that this is the common key for mission based lookups or account number, account number, account name, based lookups, or device ID based lookups, MZ number. So we have a set of five, 10, standard kits that comes... Standard keys, that comes with the starter telco edition in the data model. And using that as a starting point in our experience, has always proven to be a good point from where to start rolling out to initial releases. And then keep expanding it as you onboard new use cases on the platform.

Kevin Keller:

Yeah, I think it's great. And having that base on that foundation to really help operators to get started, I think that's definitely very beneficial. Yeah, so thanks, Elaine and Amit, for that. We have time maybe for one more question. And let's see, it looks like, so Nicole asks, "Can you talk a bit about your references with telcos, why should telco's trust you, and what have you done so far in the telco industry?" And maybe Shahed if you want to speak to that from Aerospike's standpoint. If we have time, we'll... We might not have time to get to HunaTek, but yeah, why don't you go ahead Shahed, and see if you can answer that fairly briefly.

Shahed Mazumder:

Sure. Happy to answer. So in my presentation I mentioned about three tier one Telco. The first one using us for Customer 360 and 20 different use cases. The second one, we are the backend HSS database for their IoT nationwide IoT network. For the third one, we provide the digital service platform. We provide the database for that. There are other references. For example, another tier one telco uses us as a platform for their online charging solution. So that's on the operator side. On the OEM side, on Nokia and Amdocs uses us for five different products as the embedded database underneath those products. And through those two OEMs we had deployed in close to 90 different operators worldwide. So we have a kind of robust deployment reference from the telco perspective.

Kevin Keller:

Yeah, I mean that's great. So there's definitely lots of references, lots of established companies, using Aerospike and involved with the DigiTwin solution. So we're at time. So I want to thank everybody again for the presenters, for presenting today. So Amit, Elaine, and Shahed, thank you. As well as thank everybody out there in the audience, we appreciate you listening into this webinar. And we look forward to speaking with you in the future. Thank you.

Shahed Mazumder:

Thank you.

Elaine Fletcher:

Thank you.

Amit Raj:

Thank you, everyone.

Elaine Fletcher:

Bye.