Aman Sardana

Good morning, everyone. Today we are here to talk about hybrid cloud.

First of all, why are we talking about hybrid cloud in the current situation, where most of the companies are building stuff in cloud native, and all the modern type enterprises, they don't have on-prem infrastructure? But now why, what is the significance of hybrid cloud?

There are still enterprises like financial services, retail, healthcare, that run a lot of the infrastructure on-prem. And they want to innovate. They want to make sure that they build cutting-edge solutions, but they can't just simply lift the entire infrastructure and put it in cloud as a journey to take the existing on-prem infrastructure in cloud.

And that takes us through the journey of hybrid cloud. It's all about making sure that there is a well-architected hybrid cloud strategy for companies to make sure that they can modernize legacy infrastructure, and also at the same time enhance their operational resilience.

And that's our topic for today: leading through complexity, hybrid cloud strategy for scalable innovation.

I'm Aman Sardana, application architect at Discover, which is now part of Capital One. I have my co-speaker. I'll let him introduce.

Vijay Kumar Soni

Thanks, Aman. Good morning, everyone. My name is Vijay Kumar Soni. I'm expert application engineer at Discover, now part of Capital One.

As Aman mentioned, we are going to talk about the hybrid cloud strategy for scalable innovation. Hybrid cloud is kind of the thing today for all the large organizations where they are looking to adopt the hybrid cloud. But there are challenges, so we will go into detail.

But before I get started, I will just read the disclaimer we have from my company. It's a shorter one, not as long as big as what we saw for Fidelity.

So the opinions expressed in the presentation are those of the presenters in their individual capacity and not necessarily of the company.

Before we get into detail of hybrid cloud, let me tell you who we are and what we do.

We are Discover Global Network. We are a payment company. We are one of the major networks in the United States among Visa, Mastercard, and American Express.

We constitute three different companies: Discover Network, Diners Club International, and Pulse. It's kind of a unique combination. So the way it works is Discover Network majorly operates in the United States with the integration with other partner networks across the globe.

We have Diners Club International, acquired probably about two decades ago. This runs the international franchise network across multiple countries. And then we have Pulse, another payment network, which mostly runs the debit card network across the United States.

So we have about 30-plus global alliance partners, making sure that we have outreach across the globe.

Now, some of our numbers. We have about $622 billion in the spend as of 2024, 190-plus countries where our card network operates in the partnership of other networks, about 378 million cards which run on our global network, and about 1.8 million global ATM network, where people can withdraw cash.

Now let's talk about some of the critical platforms we have on our network or in the payments domain.

First one is payment transaction authorization. So when you tap a card or swipe a Discover Network card, you see the card works directly and quickly. Now, if you see the response, it comes quickly, which is less than a second. So sub-second response time is essential for most of the customer experience.

Now, the second part is the availability. So we have to maintain about five nines availability, which is about five minutes and 26 seconds maximum downtime we can afford in the company over the year. So that's the number we have to maintain.

Finally, a couple of more platforms which we call critical, which include digital payments, transaction settlement, and fraud monitoring and detection.

Now, why we are talking about innovation, hybrid cloud, and the payments.

We have a history of evolution, and all the time we have worked on the innovation where we have changed the way how we make the payment, end of the day.

We started probably two centuries ago with a Fugio cent, like federal government probably approved cent, the Congress. And that probably costs 12 cents to produce today. So one-cent coin, but 12 cents to produce.

Now, after a few decades, we had the dollar note in 1862, followed by 1950, we had the first payment card, Diners Club. Now Diners Club is owned by Discover. So you can see that we have been part of the game for many decades.

As we made progress in the payments, payment companies, including the Silicon Valley companies, continued in innovation, and we saw that payment card with smart card and the digital payment started coming through.

Where we are today in the last about 15 years, we have seen massive, massive scale in how we make the payment, which is called card-on-file payments, the credit card in the wallet. We have the e-commerce-based payments and another coin, which is Bitcoin.

So we have started from coin, and now we are using coin. So you can imagine that we have gone too far, feels like we didn't go anywhere, but still we have gone too far. Now you see that from the physical form of the payment, we have come to completely digital payments.

Now, all this happening in real time, let's take a look at how our platform works.

So in the payment, there are a few participants: us as a cardmember, then we have the merchants and acquirers, payment network, and card-issuing bank.

So as a cardholder, we only know about our bank. We don't know what happens in between. So I will take you through the journey of what happens.

When you go online, or you go to any merchant's store, you tap, swipe, or enter the card information, or you have your card already on your merchant and just start making checkout. So as soon as you click the button for the payment or swipe the card, transaction goes through another bank called an acquirer.

As the transaction makes progress to acquirer, acquirer determines which network or which bank this transaction needs to be routed. Finally, the network comes into play. Because if you can see that there are millions and millions of merchants across the globe, they cannot directly talk to bank. It's very difficult or complex integration.

So we have acquirer on one side, and then the issuer on one side. And in the middle, that's where the payment network sits, and that's where you can see Visa, Mastercard, American Express; in China, you have UnionPay; Japan, JCB; and Discover in the United States.

Now, as the transaction goes through the network, we do some enrichment, probably fraud checks and whatnot. And then finally, we route the transaction to the issuing bank. If issuing bank decides to approve the transaction, as a customer, you walk away with the goods or services.

So in this whole system, if you see, the real-time payments are complex, although it looks simple. Like, you tap the card or swipe the card and your bank just approves. But in the middle there are a lot of entities. And maintaining a sub-second response time actually requires a lot of effort.

As we see that now we've turned into real-time payment, there are no more coins which we have to use. Our wallet is probably empty, and then we are just using our phone to make the payment most of the time.

Now, with that, I will ask Aman to take us on the hybrid cloud journey.

Aman Sardana

Yeah, thanks, Vijay. What a journey it has been from coin to cloud. A lot of innovation happened, and hybrid cloud plays a very important role in that innovation.

So now let's step back and, just for a moment, spend time on understanding what hybrid cloud is. Probably all know what hybrid cloud is. So hybrid cloud is basically where companies deploy their infrastructure across both on-prem as well as in the public cloud.

Now, public cloud, they can have a single cloud provider or multiple cloud providers, doesn't matter as long as it's spread across on-prem and one of the cloud providers they have. And then there is an integration that happens between on-prem and public cloud to provide that seamless communication between disparate infrastructure.

Now, there was a study done by Gartner in 2023, which highlights the importance of cloud. So back in 2023, the way the cloud adoption was being perceived was more of a technology disruptor. It was more of having a tech objective of making sure that companies can move their infrastructure to cloud. It was more about getting into that bleeding-edge innovation.

But now what will happen in the next few years is companies are looking at transforming business model itself.

One example, as we all know, is Netflix. They completely transformed their business model by moving completely into public cloud, AWS, from being a DVD rental company to an online streaming company. And they did so much innovation in the last 10, 15 years that they have a completely new business model altogether. So they are probably way ahead.

But that's how the industry, most of the industries like financial services, payments, and other healthcare industries, would catch up in the next few years.

Now, what are the benefits of cloud? It comes with a lot of challenges that we'll look at, but there have to be some benefits, right? Why companies are investing in hybrid cloud.

Some of the key benefits that we can highlight here are risk management. Definitely, having presence on both on-prem as well as cloud provides a company to have a balanced risk approach, where if something happens in the cloud for critical infrastructure, they can always come back and operate their system on-prem. So there's always that benefit of being in hybrid.

Cost optimization: OpEx versus CapEx, that's a main differentiator here. Sometimes you might want to have a right balance between how much capital expenditure you have versus operational expenditure you have in your environment.

Global reach: companies have a benefit of going global because of hybrid cloud. Sometimes companies have to expand their business into different regions. So instead of procuring their own infrastructure, they can simply just provision new services in a public cloud and expand their business across the globe.

Innovation: as we know, there's a lot of innovation happening. We heard so much about AI, GenAI, today morning, and most of that innovation is happening in cloud. If companies are operating their critical infrastructure on-prem, it'll be really hard for companies to integrate with the innovation that's happening in the cloud.

So there's a necessity that companies have to go into a hybrid mode in order to make sure that they can reap the full benefits of innovations happening in AI and other advancements.

Scale: as we know, cloud comes with unlimited capacity, perceived unlimited capacity. So if you have to scale for bursty traffic, that becomes quite easy with hybrid cloud.

Finally, compliance and data sovereignty. There are mandates that are coming up in different regions and different geographies where you have to keep the data local in the geography. And data localization mandates make it absolutely must for financial service and payments to go hybrid so that they can keep the data local in the region where the mandate is.

Now, what this leads to, definitely there's a lot of complexity. Companies start with on-prem. They have mainframe infrastructure, still virtual machines, containers, private cloud. So companies always started doing some modernization being on-prem by containerizing their applications, having their own private cloud.

But again, they couldn't fully innovate with the infrastructure they have on-prem, and they started moving to public cloud, where either they take the containers to public cloud or they reap the benefits of the innovation happening in public cloud, where they either go serverless, where they don't have to manage the infrastructure anymore, or they go with managed services, which are very specialized and provided by cloud providers.

For example, they don't have to build their own call center application, or cloud providers are building out-of-box call center applications that they can leverage. Or they can go with still virtual machine if they have critical workload that they want to lift and shift, and then having an interconnect through a co-location site that provides a connectivity path between on-prem and public cloud.

So as you see here, it's a lot of complexity, especially with the hybrid cloud. It's one connected ecosystem that has to be built, and there has to be a proper DevSecOps to make sure that you have a consistent release process across both on-prem and hybrid cloud.

Now let's spend some time on the critical key challenges that companies face when they run in hybrid cloud.

And what we have done is, based on our analysis, we have split it into two parts here.

The category one is mostly centered around infra and app services. So this has to do with the complexity that arises in the infrastructure tier and the services that run in the infrastructure.

For example, one of the complexity that arises in hybrid cloud is mostly around network, right? You have disparate networks. You have your own on-prem infrastructure, cloud. You have IP address space across both on-prem and cloud.

So there are challenges where your IP addresses conflict with each other, and then you introduce network latency as well. Most of the critical infrastructure companies operate on-prem. They are mostly co-located, but now there are different roadmaps that each of the product areas have, right? Some services are moving to cloud. Some are staying on-prem.

Now you need network latency, and that really impacts the response time that end customer sees. So as Vijay explained, some of the critical infrastructure requires under-one-second response time, and you cannot afford to have a lot of back and forth between on-prem and cloud because that will impact the customer experience.

Service discovery, cluster management are a few other challenges that companies face when they move to hybrid cloud.

Security is a big one. Depending on the industry, it could be a big challenge. For example, in payments and financial services, it's a big challenge because you have to make sure that your sensitive data is secure and you are complying with all the regulations. In case of payments, there's a PCI compliance that has to be met. And that becomes a challenge in hybrid cloud.

The second category that we have here is DevSecOps. This is to do with how you make sure that you have robust, resilient pipelines to release the software into production.

We mostly think of resiliency as something that is more runtime: how you make your processes resilient in production. But there are things that have to happen as you build your software to make sure that you can release the software in a timely manner in the cloud as well as on-prem.

So there are integration constraints, DevSecOps, DevOps effort that has to be considered. How do you make sure that your deployments are safe and you can always do a disaster recovery when running in a hybrid mode? How do you make sure that your platforms are highly available? How do you monitor your infrastructure?

Lastly, how do you make sure that you have the right business cost allocated to hybrid cloud? So that brings us a good thought process around what key challenges are and how do you mitigate those challenges.

Now, those are some of the technical challenges that we looked at, but as leaders in our industry, there are some leadership challenges that we would like to call out that's mostly centered around some of the important considerations around hybrid cloud adoption.

We have been talking about innovation, right? Today morning also, there was so much emphasis on innovation that is coming up with AI and cloud technologies. But one thing that leaders are challenged with is the upskilling. Do they have the right skills in the organization to innovate with the new technologies, and how to upskill their resources to start leveraging the new technologies while they still have to support their legacy infrastructure on-prem?

How do they make sure that the deployments are secure and compliant? That's mostly a leadership discussion that happens quite often.

What is the cost optimization and return on investment as companies move to hybrid cloud?

And then finally, as leaders, there's always this challenge of making sure that the system is resilient. You cannot afford to have a critical infrastructure going down in production. That's a no-go, right? That should never happen, especially with five nines. It's a continuously available system.

So how do you make sure that with hybrid cloud, which raises a lot of challenges, your infrastructure is still resilient?

And ultimately, it leads to the statement here that hybrid cloud is not a technology choice, it's a leadership decision. It's not just that you have technology goals, but also you need to make sure how those technology goals meet the business goals. And that discussion happens at the leadership level.

Now, we talked about the challenges, benefits, but then what's the blueprint of having a robust strategy around hybrid cloud so that there is a successful outcome?

It all starts with having a vision and establishing goals as to what you want to achieve from business perspective by going into the hybrid cloud.

Then you might want to take a look at what your current infrastructure is. What are the critical applications? What are the non-critical applications? And what type of workloads you have. Do you have APIs, batch workloads, event processing? So you do assessment of your workload.

Then selecting the cloud provider is a very important and difficult choice that companies have to make. There are a lot of cloud providers: Amazon, Google, Microsoft. Which one do you pick? Do you pick one? Do you pick multiple? What are the criteria for cloud provider selection? That's an important consideration.

Then designing for the hybrid cloud, making sure that you have the right reference architecture in place. You don't want to just let your product teams start building in hybrid cloud. You need to make sure that there are right patterns and reference architecture documented.

There are different approaches that teams can take. They can either take the current application and do a lift and shift of the current application, or they might want to refactor the current application and maybe make it more cloud native, or you might want to just completely rewrite the application as you move the application to cloud.

So those decisions have to be made by the product teams, but then they have to be given some guidance through some reference architecture that has to be built upfront.

Then finally, teams start moving the infrastructure to cloud. Then you get into that running mode, right? You want to make sure that your data is secure and compliant, making sure that you are doing continuous testing and optimization of your infrastructure, and then finally monitoring and doing cost management.

In our experience, what happens is, as you let your developers provision anything in the cloud, they will provision more and more stuff without understanding the cost that comes with it. So making sure that there's a proper cost management structure around your hybrid cloud is really important.

With that, I'll hand it over back to Vijay to take us to the remaining portion of the slides.

Vijay Kumar Soni

Thanks, Aman. Thanks, Aman. A lot to get those five minutes and 26 seconds.

All right. So we talked a lot about payments and hybrid cloud and the strategies.

Now at Discover, we started our journey quite a while ago, and over the last few years we started our migration journey. And one of the major platforms I talked about in the very beginning of the presentation was the tokenization platform.

So we delivered our tokenization platform in the hybrid cloud mode. Now, as you see, you use digital wallet, you go to merchant and do the cloud payments and whatnot. All these are using the tokens, and then tokens are growing. And having that scale is almost difficult to develop and manage on the on-prem.

So now the compute and the storage requirement is growing. That's where the hybrid cloud came to rescue, and we deployed our tokenization platform in hybrid mode, where some part of the sensitive data stays on-prem and remaining on cloud.

Now, with that, in 2023, our first tokenization platform was live. And I think so far today, we have seen tremendous growth in the platform.

Now, the next one we did with partnership with AWS, where we had our payment network transaction settlement processing speed up by 66%. So this is the page from AWS website itself where we had presented our benefits of being on hybrid cloud.

Finally, our OpenShift platform. We worked with IBM a few years ago and developed our OpenShift container platform. If you are in hybrid cloud, unlikely you'll probably shy away from containers. So we have OpenShift container platform, where we have both critical and non-critical workload, which helps us scale on demand as well as deliver our value to our product.

Now, these are a few things our leaders have mentioned.

Now going to the key takeaways, that's our last slide.

What we say is that hybrid cloud is still relevant to many industries. Some new industries probably might just go to the cloud because they don't have any legacy and they don't have any challenge of maintaining the data set and whatnot.

Especially in payment and financial services, we might still have challenges because the banks have existed for many decades. And what we say is that when you move to hybrid cloud, you have the complexity and you cannot just bypass it. You cannot just move to cloud and shut off your on-prem right away. You will have to define a strategy and finish the journey from on-prem to cloud.

If a business wants to completely move away from on-prem, in that case you have complexity and you have to lead through it.

Another key takeaway, we think, is that if companies want to innovate, then it's not the technical architecture infrastructure. The business also has the largest stake in this decision making, into how they want to purpose the hybrid cloud setup so that innovation can continue happening.

Finally, the team structure. So you have the infrastructure, you have business strategy, everything, but people are the most important assets.

So people need to be skilled to make sure that they follow the governance policies and understand that the hybrid cloud is not just, there are less controls unless you would well define it. So the people who are working on the new products or R&D or whatever, just make sure that you have enough governance and controls around it to innovate.

So the bottom line is, if you lead through the complexity of hybrid cloud, you create an engine for innovation and growth. So if any organization wants to adopt hybrid cloud, then there is complexity, but those are not impossible to deal with. It's just that you need the right strategy.

With that, we thank you here. And if there are any questions, we are open to answer. Thank you.

Thank you, everyone, for being here.