How regulated financial institutions drive adoption of digital assets

All right. Hello everyone. Let's get started. Welcome on a Monday morning. Very happy to see you here in this breakout session.

Uh let's wait until the, the background music. It's a very rousing music. I think I'd like to talk to some songs, but here we're gonna be exploring how regulated financial institutions are actually adopting digital assets.

And my name is John Liu, the head of product for Amazon Web Services and manage blockchain services. And joining me are Kadu Ma, the CIO from Itau Unibanco and Uman Chandran, the director of architecture from Fidelity Investments.

We'll start with an overview of how the digital assets are being adopted by the financial industry. We'll look at some of the blockchain challenges they faced along the way as they're incorporating this new technology and some of the solutions available to them.

Then we'll look at how AWS helps our customers build simple and secure digital asset platforms. Before turning it over to two customers who were at the forefront of digital asset adoption. Itau Unibanco will share their story on digital custody and Fidelity will outline their asset tokenization platform.

Now, this is a Level 100 session. So I'll be introducing a technology at a very high level. So if you want to learn more about what we talked about today, please find me after the session.

Over the last six years, the financial industry has gone from skeptical of digital assets including cryptocurrencies such as Bitcoin and Ether to one of being increasingly accepting of these digital assets and to see whether this shift in attitude is a lasting trend, it helps to look at how the financial industry has adopted digital assets over the last decade.

If you were to plot the adoption of any disruptive technology over time, you would often see a stepwise function where there are long periods of investment in the infrastructure and understanding that infrastructure and from the outside, it doesn't seem like anything's happening and these periods will follow jumps on adoption and in these jumps in adoption is when you actually see the impact of this new technology, a very timely example of that is generative AI, right?

It seems to have blossomed overnight in the last couple of months. But really this is on the back of decades of improvement in machine learning algorithms, increasing compute power and the increasing availability of data and the digital assets and the public blockchain and private blockchain technology powering these digital assets follows a similar pattern.

From 2012 to 2020, the financial industry was mostly in a state of learning and experimentation. Although there were some milestones like when the CME launched its Bitcoin futures. In 2017, it was really in 2020 to 2022 when the financial industry had its first jump in adoption.

During this time, major financial institutions such as JP Morgan and Morgan Stanley and Goldman Sachs launched cryptocurrency offerings for high net worth individuals and institutional investors. At the same time, trading was also supported by increasing number of hedge funds and family offices for cryptocurrencies and a handful of corporates starting holding cryptocurrencies on their balance sheet for diversification.

At the same time, a rapid rise in interest for non-fungible tokens or NFTs and their ability to digitize brand intellectual properties and create new forms of engagement with the fans brought in media companies and retail companies, entertainment companies into this digital asset space.

As these companies launched digital assets on public blockchains, they started turning towards financial services industry to help them manage these digital assets. However, the billions of dollars flowing into the space in a largely unregulated environment led to unsustainable pockets of risk and unsustainable workloads.

So regulators have been putting in place the necessary guard rails to enable mainstream adoption since the end of 2020-21. Now, even as regulators have been putting in place, the clarity needed, the adoption by financial industry for digital assets continued, the number of regulated institutional custody solutions continued to grow and companies such as Fidelity and Charles Schwab introduced products for retail investors to trade cryptocurrencies.

And now we stand poised for another jump in adoption. BlackRock's filing of its Bitcoin ETF in June of this year led to multiple filings for cryptocurrency ETFs from multiple institutions. The SEC is currently reviewing 12 Bitcoin filings and ETFs is an $8 trillion industry and serves to enable mainstream access to any type of asset.

And the same holds true for cryptocurrency because if you're an institution that wants to get cryptocurrency exposure, you no longer have to invest in specialized blockchain infrastructure. You no longer have to get specialized licenses. And if you're an investor that wants to buy cryptocurrencies, you can buy an ETF as easily as you can do stocks. You don't have to learn concepts like private keys and wallets.

Beyond ETFs, cross border settlements with stablecoins is emerging as another adoption driver. There's already $100-$120 billion of stablecoins today settling billions of dollars daily. And PayPal's launch of its Pax USD in August of this year is set to further push the adoption of stablecoins into mainstream use cases.

Pax USD is a regulated stablecoin and is fully backed by cash and cash equivalents and it can be used in web3 native use cases like decentralized lending applications as well as PayPal's network of tens of millions of merchants and hundreds of millions of consumers.

Asset tokenization has also emerged in the last six months as another adoption driver. It's really seen a resurgence since the early years. Over the past five years, the financial industry has learned some of the benefits of asset tokenization such as the ability to settle transactions 24/7 and automate certain intermediary functions.

With these learnings paired with a 5% risk free rate, the asset tokenization this time is focused on liquid capital markets such as Treasuries and money market funds. And this represents a well understood low risk asset to bring traditional finance into the world of digital asset industries.

While at the same time, bringing web3 investments into traditional finance. Now, at each jump in adoption that we've covered the blockchain technology that has been powering digital assets has been more tightly integrated into the core financial infrastructure and institutions that have been building with blockchain have had to consider some key challenges.

First, can the technology keep up with an industry that requires throughput of thousands to tens of thousands of transactions per second with sub second confirmation times?

Second, how can institutions protect their end users from having to hold cryptocurrencies just to pay for blockchain transaction fees known as gas and likewise, how can institutions budget for periods of volatile gas fees so that their critical transactions are not delayed?

Third, how can data privacy be maintained when certain sensitive traits need to be isolated from public blockchains which can be accessed by anyone?

And fourth almost conversely, how can KYC or Know Your Counter Party and AML or Anti Money Laundering requirements be met when anonymized transactions and wallets are part of the technology being used?

Fifth are their key management solutions which balance the benefits of being an individual's own bank while still meeting the security and compliance requirements of financial institutions?

And six in the kind of multi blockchain world that has emerged, how can institutions create an interconnected marketplace where investments and liquidity can flow freely from one blockchain to another?

Now, luckily, blockchain protocols and many tooling providers have been considering these exact same challenges as they look to enable mainstream access to this technology.

When we think about the challenges of scalability, gas management and data privacy layer two or app chains have emerged almost as a three for one solution. They use different types of technologies like zero knowledge, roll ups, optimistic roll ups. And you may hear terms like supernets and subnets, but really they're just balancing different aspects of scalability, privacy and compatibility with the main layer one chain regardless of the technology that they are used.

The principle actually has is very similar to a traditional financial approach where you settle or process transactions locally in an institution and then periodically batch settle these transactions to a larger network.

Now, by taking some of these transactions off the main layer one chain into a layer two chain, these layer two chains are able to achieve transaction throughput in the thousands to tens of thousands of transactions per second with sub second confirmation times and transaction fees that are fractions of a penny.

The important point is they're also built on top of that layer one chain. So they inherit the security and immutability of that layer one chain. A layer two chains also offer levers of customization. So institutions depending on how they implement the layer two technology can customize how much of the gas they want to pass on to the users of the layer two chain.

They may even subsidize the gas fees completely. Institutions can also tailor for different levels of privacy. Certain layer two chains actually commit only the end balances of all the wallets on that layer two chain to the layer one chain, thereby protecting all the transactions that took place in that period. And that's kept private on the layer two chain.

Turning to KYC and AML, there are a variety of solutions available that can be combined and to meet institutional needs. The data on public blockchain is now better understood. So institutions actually have an easier time monitoring transactions and they can create risk scores around these blockchain transactions as well as around the wallets associated with these transactions.

On top of the risk scores, they can build risk policies. In addition, there are a variety of solutions that make it easy to combine off chain, identity and off chain data with wallets. This allows institutions to take their traditional money laundering approaches and anti money laundering finding type of algorithms combine them with on chain forensics to create a complete profile of who and what world they're transacting with.

Finally, advances in decentralized identities is enabling a more efficient form of KYC where the individual owning that identity, which is typically the person themselves can choose to control how much of their sensitive information they want to share who they want to share that information with and for how long they want to share that information.

For at the same time, the institutions get the benefit of receiving the verified information they need to complete their KYC requirements and they don't have to store that sensitive information in their own databases.

There are also key management solutions that balance the benefits of decentralized account control and still meet the security and compliance requirements of financial institutions, multiparty compute protocols and MPC protocols such as threshold signature schemes create the cryptocurrency equivalent of multi signature bank accounts.

This is a situation where the number of signatures and the rules required to control that account can be customized. So by generating the private keys which control the wallets in separate shards and storing these shards separately. A new form of security is brought into the wallets.

A set number of these shards must be brought together to sign a transaction which of course controls that wallet. Wallets that use MPC type of technology remain secure even if some of the key shards are compromised.

In addition, should some of the key shards be lost? The account itself can still be recovered provided the minimum number of parties consent to the account recovery. Finally, the end investor can choose to hold some of these shards themselves so that they still control their assets even while they store their assets at trusted financial institutions.

Rounding out our list of challenges, interoperability. And on this front layer zero protocols have emerged as a combination of open banking APIs, standard messaging protocols and clearing houses all in one.

Now, these layer zero protocols can be a full blockchain protocol which other blockchains can build upon or other applications can build upon or they can be more specialized cross chain bridges. And unlike the first generation of cross chain bridges, which focused on seamless token transfers and relied on centralized escrow models which made them more susceptible to hacking the new generation of layer zero bridges and cross chain bridges actually use a decentralized network of oracles or observers on the different chains and relays which pass any arbitrary information between blockchains.

This can be data of course token transfers and it can be generalized contract execution. What this means is a smart contract on one blockchain can now react to some state change or some event that took place on another blockchain.

However, taking all these solutions and making them available with enterprise grade scalability, availability and operational excellence required by financial institutions still requires undifferentiated work from financial institutions.

Take for example, connecting to the nodes which are these servers that are required to interact with the blockchain. Financial institutions have to create highly resilient and available fleets of nodes across different regions.

In addition, they have to make sure these nodes are kept up to date with the latest patches because if they're not, there's a chance they might be interacting with the wrong version of a chain, putting their digital assets at risk, they must do this for every single blockchain they support, which greatly increases the operational load of managing blockchains.

At the same time, getting access to the data on public blockchain may seem very easy because it's always available. However, delivering this data in a way that downstream applications and analytics can use them is actually quite difficult.

Not only do institutions have to maintain those resilient fleet of nodes I mentioned earlier, they often have to run heavier nodes which are archive nodes, they have access to the full history of the blockchain.

They have to perform complex data transformations often across multiple sources of information on the blockchain including smart contract logs to get what they need. They also have to process and deal with the eventual finality situation of blockchains, which means that the recently processed information including balances and transactions might change.

And they have to build this logic to always deal with the nonstop streaming of information that keeps coming from the blockchain itself.

So customers to use AWS to help them with this undifferentiated lift a variety of customers from the leading institutional enterprises to leading innovative web3 protocols. Today use AWS for core compute and storage services for application services and for building access to the data services as well.

They also have access to the latest machine learning algorithms and generative AI models. Builders of digital asset platforms also use AWS security services such as AWS KMS and Nitro Enclaves to help them manage the private keys, which are very important obviously in protecting the digital assets.

Furthermore, customers who are working with blockchain use Amazon Managed Blockchain to lighten their blockchain development. Amazon Managed Blockchain provides highly scalable service infrastructure to blockchain nodes and it delivers the commonly requested data with sub second latency.

Amazon Managed Blockchain is made up of two services: Amazon Managed Blockchain Access, which provides that fully managed node and instant connectivity to the blockchain APIs, and Amazon Managed Blockchain Query, which provides the commonly requested information such as balances and transaction history without any of that complex ETL framework or work that I mentioned earlier and delivers all this data with sub-second latency.

Amazon Managed Blockchain supports both Bitcoin and Ethereum today and we're constantly talking to customers to see how we can better help our customers with their work. So I'm very thrilled to announce today, the public preview of Amazon Managed Blockchain Access for Polygon.

Polygon is a Layer 2 high throughput EVM compatible chain, which means it's compatible with Ethereum, has a diverse ecosystem of Web 3 native applications and is used by some of the leading brands such as Nike, Starbucks and Fidelity Investments to actually bring their Web 3 workloads to reality.

Some customers are already using Amazon Managed Blockchain Access Polygon today to lighten their development load. Take for example, Oasis Pro. Now, Oasis Pro provides end to end solutions that bring Web 2 to Web 3 for traditional financial institutions. And part of their solution includes an asset tokenization platform that is FINRA registered and can support both private and public tokenized securities. They were used by JPMorgan and Apollo as part of the Monetary Authority of Singapore's Project Guardian. This POC that was built was set to demonstrate the benefits of tokenization on interoperable blockchains and it used both private and public blockchains.

Another customer that is using Amazon Managed Blockchain Access Polygon is Mystic Moose, which is the independent game developer, Mighty Mojo Melee. It's a very, very good name I have to say. And Mighty Mojo Melee is an auto chess battler which has deep gameplay with captivating visuals and it blends the Web 3 components seamlessly into the actual gameplay itself. Certain aspects such as storing the heroes, they're actually stored as an NFT on public blockchains. And Mighty Mojo Melee announced a collaboration with Amazon Prime Gaming this summer where Prime members could actually get access to NFT characters.

Rounding out our list of customer highlights is Magic. Magic is a wallet-as-a-service provider that allows businesses to seamlessly onboard their users onto Web 3 by instantly creating non-custodial wallets. Magic uses emails and social logins so that the creation of these wallets behave just like the experience today when you sign up for any type of Web 2 type of application. And they are one of the examples I gave earlier of a tool that institutions can use to link off-chain identity and data to wallets with their technology. Magic has already created 25 million wallets.

We're very excited to see what else our customers are going to do with Amazon Managed Blockchain Access for Polygon. And on that note, I'd like to turn it over to Caio Ma who is going to share Itaú's digital custody solution called Thinking Crypto.

Thank you John. I'm Caio Ma, responsible for Global Markets and Treasury Wealth Management services and Institutional Trading at Itaú Unibanco. For those who don't know, Itaú Unibanco is a full service bank in Latin America, in Brazil. We have almost 100 years of history. We're the largest financial institution in Latin America and we're also the most valuable brand in Brazil. We have more than 70 million customers and we're not only strong in the retail market but also on the wholesale market.

Itaú has the largest private asset manager in Brazil and also is a top 5 investment bank in the region. Moreover, it is also providing several services to institutional investors. And in that scene, the digital asset landscape is evolving pretty fast. In Latin America, we according to researchers, 10% of all business transactions will be using blockchain infrastructure by 2027.

The regulatory environment also in the region is evolving. The CVM, which is the equivalent to SEC, created and published a new regulatory framework and institutional investors will be able to invest and allocate crypto assets in their portfolio. And also the Brazilian Central Bank launched an initiative to create the new infrastructure and to create the Brazilian digital coin, the CBDC that will be used next year.

And we've been studying at Itaú Unibanco the blockchain infrastructure for quite a long time now. We started back in 2016 with the Center of Excellence. The idea behind it was to experiment and see what's the technology, what the technology can help us in the entire ecosystem. Then we created in 2021 the Digital Asset team that focused on crypto markets. The initial idea was to have a tokenization service, custody service and crypto distribution.

Then in 2022 the government created the regulatory framework so that the entire ecosystem could evolve. And in 2022 again, we launched a new service, the crypto custody service. And by 2023, this year, the government selected Itaú Unibanco as one of the institutions to help develop the new ecosystem and to help with the tech in a joint technical effort to develop the CBDC.

And the Resolution CMN 175 will allow next year all the funds to allocate in crypto assets. So we see that as a huge opportunity as a bank with a good history and a good reputation. We believe that providing a secure and easy access solution to crypto asset custody is key to evolve in this market. We also want to strengthen our leadership and position in the market as the leading choice for digital asset custody and expand our portfolio of services.

The idea behind it is to make sure that all our institutional customers and clients can use the crypto assets in an easy way to abstract all the complexity that John mentioned here.

What do we mean by a crypto custody platform? First, it's a combination of several different layers. The first one being the security layer, the security layer is where all the safe transaction propagation happens, where the transaction signing happens and where the key management happens. On top of that, we have the compliance layer so we can apply all the AML rules and policies. We do the screening to make sure the risk is acceptable and we do the onboarding of the crypto to our custody.

On top of that, we have the governance layer. The governance layer is where the controls are tailored to each client. So we can control the operational limits, we can do the account and sub-account management and we can also do the portfolio management and the risk management. Then we have the external audits and where everything gets encapsulated so we can provide as a service to our clients.

We have a wide spectrum of clients. We provide services to retail customers to have their custody, and we also provide this to institutional investors.

And now I'll talk more about the security layer. We have some basic design principles. The first one being that we wanted to have a blockchain infrastructure that is easy to maintain, that is scalable, we can update the clients seamlessly and we have the load balancing. The ecosystem is evolving and we wanted to have high availability, high resilience and to make sure that the infrastructure will hold.

Then data encryption for sure we needed encryption in transit, encryption at rest and also processing using confidential computing and the flows that I will show you later, you see that the key never gets exposed at any time.

And the key management, the key management is not trivial. So we had to use different techniques to have the proper key generation, key storage and key usage.

So going basically through three flows to get a better understanding of how we put every component together.

First, the client creates a wallet using Nitro Enclaves. The Nitro Enclave is connected to KMS using cryptographic attestation to make sure that everything happens in that environment. The wallet gets created, the key gets created and the key gets encrypted. Just after the creation, the wallet service stores the address using the wallet key store, which is a special environment that we have to process and make sure that we have the security needed for that piece of the architecture. And then the transaction in the wallet is registered in the listener service.

We have two components. The first one is the key management and the second one is the blockchain networks and we wanted to segregate the responsibilities to make sure that everything is together and there is no way to access the Nitro Enclaves environment.

Then the monitoring transactions, it's where the listener plugs into the blockchain networks using Amazon Managed Blockchain. You start listening to every node, we listen to the block headers, filter the addresses of the wallets to see if any transactions happened in that wallet and create a copy on the local off-chain databases so we can do all the monitoring needed and we can understand what's happening on the chain. Again, using the key management service here helped us to manage all the connections and segregate the responsibilities accordingly.

And then the transaction signing for the withdraws, here the client requests the withdraw, the wallet services request estimate fee, gets the wallet address, gets also the private key decrypted, private key, sign the transaction and send it to the nodes. The movement services propagate the transaction to the blockchain so everything can update in the chain. So the secure usage of the key is essential.

The Nitro Enclave allowed us to achieve that with the cryptographic attestation and the connection with the KMS using the wallet key storage service.

So Amazon Managed Blockchain and Nitro Enclave helped us first manage the blockchain nodes, get the usage metrics. The usage metrics are especially good on the compliance layer and on the governance layer so we can put up the controls that we needed.

We have all the infrastructure provisioning using CloudFormation. We have simple connection using RPC and the convenience to escalate. Like I mentioned before, the ecosystem in Latin America is evolving fast so having the ability to escalate as we need is really crucial to our solution. We can access the data in real time including all the historical data. And by using Nitro Enclaves we can reduce the attack surface area and also isolating the computing environments.

We process all the sensitive data in a way that the key is never exposed. And looking forward, we will try to use the Amazon Managed Blockchain Query to verify the balance of the Bitcoin wallets. We are also expanding to new tokens and digital coins and we'll have the scalability needed for the Brazilian digital currency. And we are also expanding the use of the platform to other solutions and other products that the bank provides, not only using it on the blockchain digital asset landscape but also trying where other solutions can leverage all the capabilities that we built here.

And with that, I'll hand back to John.

Thank you very much. I'd now like to invite the Director of Architecture from Fidelity to share their asset tokenization platform and learnings.

Thank you John. I'm glad to be here to talk about our asset tokenization platform and our learnings in general.

For people who don't know about Fidelity Investments, which is mostly unlikely, Fidelity Investments is a privately held company. It was founded in 1946. We have been in the industry for more than 77 years and we serve more than 43 million customers in United States and we serve more than 3,600 financial advisory firms day in day out. We have more than 70,000 employees globally to serve our customers and more than 23,000 employers who manage their employee funds through us. We manage more than $2.5 billion...

I would rather call it investment to kind of develop cutting edge technologies to serve our customers - 2.5 billion SB (big B). We started our first cloud application way back in 2016 and we have more than 6500 applications currently running on cloud as we speak.

So myself again, Omar Kanan Chandran, I work as a Director of Architecture in a team called F Digital Asset Management. And I'm specialized in building platforms on F infrastructure, mainly on the emerging technologies.

If you look at the typical FUN flow, basically, if the investor wants to buy funds from us, they come to Investor Gateway or a brokerage account with their money - which is nothing but Fidelity.com - and there are multiple parties involved in order to get the transaction completed in order to get the funds. They have to go through all this process and finally they get the funds.

Seems simple when you look at this entire flow, but typically it takes more than two days in order to settle and get your funds in your custody. And in case of any issues, errors happens, the reconciliation takes hours to fix it. Why does it happen? Because it uses the traditional protocols like ACH and file transfers and other well known established protocols.

But there's always hope to implement this whole process. That's when we started looking at, can we leverage blockchain and tokenize our funds? So what is the token? It's a digital representation of your assets, underlying assets on a blockchain and you take the ownership of that right? And any financial instrument can be tokenized and make it available for the investors to come and buy and sell on an exchange.

And how does it gonna go with the existing flow? Assuming that the transfer agent, back office and the custodian are all on the same chain, which will make this whole process simplified. And we are talking about reconciliation could be very simple. And also with the sorry, with the with the whole central bank currencies or digital currencies, even though settlements could be could be much faster, we could we could do it within, within the real time or near real time.

So the numbers don't lie, right? If you look at this particular website who is very specialized in tracking digital assets on chain, there's more than 120 billion tokenized cash right now available on chains and 4 billion is being lent out as private credit loans. And in terms of US Treasuries, more than 700 million have been tokenized on various chains as we speak. It is almost nearing 800 million within a month. It increased by 100 million. So there's a lot of scope for the tokenization and certainly from the Fidelity perspective, we are seeing new way of generating revenues for our customers and with a lot of efficiency and operational improvements, we could actually make this available for customers who are outside the United States.

And with the reduced settlement times, we can have good customer engagement as well as the transparency that it brings because everything is available online or on chain.

Let's look at the capability stack that we built for this platform. The first foremost is the blockchain network and obviously we used both private and public networks in order to keep our tokens on the chain. So how do you really interact with it? Obviously, you have to use any of the node providers. And in our case, we use ConsenSys Quorum to manage these underlying nodes.

The next is the smart contract development. I'll talk about smart contracts a little later. But in order for your tokens to be available on chain, you have to develop these tokens, right, using a specific way. And we have elaborately used a lot of frameworks to develop that. And once it's developed, we have our own pipelines to actually deploy the smart contracts onto the chains.

And once you do it, obviously, you have to build some of the DeFi capabilities in order to make sure that these tokens can enhance its own life cycle. And we will talk about those capabilities in the later slides.

Finally, the integrations where the UI or the API is to interact with all of these layers. So that's very high level of how we build the capability stack.

As I was covering smart contracts, I think John kind of covered up on the NFTs. So any asset that you want to make it available on chain, you have to write a contract and that contract lives on the particular blockchain. So that's what we call a smart contract.

And there are two ways that you can do these tokens. One is a non-fungible token, which is unique by itself. The other one is a fungible token. In our case, actually, we use a fungible token to tokenize our assets.

So if you look at the smart contract development as like any other applications that you write, the tokenization also has to use a programming language. The smart contract needs to be written in a programming language. In our case, we actually used Solidity as a programming language for the Ethereum blockchain.

And there are other languages available for developers to write in, but we chose Solidity to write and we use a lot of leverage, a lot of open source projects. For example, we use OpenZeppelin to write these smart contracts. And also we used Hardhat to deploy these smart contracts onto the chains.

And when it comes to token standards, this is a set of guidelines actually provided by the community in order to write these smart contracts. We have used ERC-20 as a standard to write this and there are other standards which are available and it depends on the companies to kind of pick and choose the right standard for their, their tokens and the design patterns part of it.

By default, when you deploy the smart contracts on the network, these are all immutable by nature, right? Which means every time you deploy a contract, it gives a new address on the chain. So if the upstream system needs to kind of interact with it, you cannot go and modify this every time.

So we used a pattern called UUPS which abstracts the way that the smart contracts are actually integrated in the upstream systems so that it is easier for us to kind of continue to add new business functionalities into our smart contracts.

And as like any other applications, you have to have a good set of unit tests cases to make sure that your functionality is never broken when you're adding new functionalities. So we do have unit tests written in frameworks like Waffle and Chai and make sure that we have a good code coverage of more than 95%.

And the last is the code analysis. Since these smart contracts that you're writing will be always available on chain, we need to make sure that the security aspects of the code are very well maintained. So we use open source frameworks like MythX to do code analysis and make sure that there are no vulnerabilities available on the contract before we deploy it on chain.

So I was kind of touching upon the DeFi capabilities. So some of these are very, very critical if you want to really make sure your asset tokens are available on chain without any issues, right?

The first is the business events. So any of the events for blockchain is a 24/7 kind of an ecosystem, right? You need to make sure that you have the control to make that tokens available without any issues from the operational perspective, as well as the business continuity perspective.

So we need to listen for the events and make sure that when you receive the event, you perform the action on the chain without any issue. So I'll give you an example, right? For example, if you want to pause all the operations on chain, this is not your system where you can go and just stop it so that people will not be able to do it. Since it's a decentralized network, you have to put appropriate controls in place so that if people wanted to perform these operations on chain, they'll not be able to do.

So we have introduced all of that in our business events and operator controls so that the audit part of the whole equation is taken care of. And of course, the right people with the right access can do the particular operations that will take that is taken care by the role management.

And we do have off chain on chain, off chain integration since Fidelity uses its own KYC process, we don't want to kind of make that available on chain. So we have an off chain API which we built a layer to kind of interact with it to make sure that every user who is doing the transaction on the system is verified through that particular API as a concept.

With that, you will be able to achieve that. The last two is the compliance rules. Since most of this token which is written has to be compliant with the rules set by the regulatory body, we need to make sure that those rules are actually available on the smart contracts.

So we have written a lot of rules off chain as well as on chain. So we'll make sure that every transaction which just goes through has to be kind of approved by these rules.

And last is the transferability since these tokens can be actually transacted peer to peer. Like if you take a MetaMask wallet, people who have, who wanted to have the non-custodial of these tokens would be able to still transfer the tokens from one wallet to the other wallet, obviously with all the KYC done on the wallet side.

So how did we really develop this entire platform? So as I was indicating we are on AWS ecosystem, we use a lot of AWS functionalities. So it was easy for us to write a smart contract.

So once we wrote the smart contract, we actually leverage AWS Managed Blockchain to deploy it into the public test network. So we started with the test network. But actually, if you look at the actual design, which is very high level, not simplified one, we use actually Hyperledger Fabric as a private network to test our smart contract before we put on any public network.

So if you look at the flow, we write the smart contract, use the deploy pipeline to store it. And once it's stored, we make sure that we use the deploy service to deploy that into both private and public network through AWS Managed Blockchain.

We put it on a public network through AWS Managed Ethereum cluster to put it on our private network called Fabric. So any business operations when we, when a user actually wants to transact a specific number of tokens, the request goes to DeFi service which internally kind of makes sure that all the compliance rules are met.

So once it verifies and validates, it gives the controls to the execution service. The execution service actually uses the gas optimization to make sure we pay less gas for that particular transaction. And once it's done, it will submit the transaction to blockchain which will get invoked.

And we get the transaction details back from the blockchain. So those details actually get stored into a metadata database. In this case, it's stored into a DynamoDB. And also we have a listener called event service which listens for all the events which is emitted by our smart contracts.

And we store that information back to the metadata store for the transaction. So this is a very, very high level architecture of our tokenization design.

What are our learnings with this platform? So since it's a very new technology, there's a lot of learning and experimentation that happens throughout the life cycle of this platform.

Fidelity gives 4 hours every week for associates to kind of come and try new technologies and not only just learning, they wanted to kind of apply that learning into an experiment. So this was started as an experiment and we were able to successfully demonstrate the capabilities of the smart contract end to end with that.

And obviously, we would like to share our learnings whatever that we did as part of this whole experimentation with the global audience. Obviously, I've traveled all the way 29,000 miles to kind of share this information with you all.

And also we do share that information within Fidelity itself. There are a lot of teams who are interested in understanding how we really written the smart contracts, what are the challenges that we faced, all of that. We were sharing internally and that was well received.

And we, we, as I indicated, we use a lot of open source frameworks. The thing is we just don't stop using it. We also try to contribute back to the community. In this case, we are contributing back to Hardhat for one of the issues that we faced within that particular framework.

So as I was indicating Fidelity always kind of gives customer first approach and we actually looked at this platform as customer first angle so that we know what exactly the customers want and we develop the framework based on that.

And we have immense talent and we work as one unified team to make sure that whatever we wanted to really deliver, we deliver as it is.

And thank you so much for listening.