WO2019071278A1

WO2019071278A1 - Distributed crypto-currency smart order router with cost calculator

Info

Publication number: WO2019071278A1
Application number: PCT/US2018/064115
Authority: WO
Inventors: David Marc WEISBERGER; Ian Joseph WEISBERGER
Original assignee: Coinroutes Inc.
Priority date: 2017-10-08
Filing date: 2018-12-05
Publication date: 2019-04-11
Also published as: WO2019071278A9; EP3692486A1; EP3692486A4; EP3692491A1; EP3692491A4; WO2019071277A1

Abstract

A distributed cryptocurrency smart order router (DCCSOR) is disclosed, which allows clients to keep control of their own private and exchange keys to their wallets and accounts, but can execute orders across multiple exchanges simultaneously.

Description

DISTRIBUTED CRYPTO-CURRENCY

SMART ORDER ROUTER WITH COST CALCULATOR

BACKGROUND

[0001] The current structure of the Crypto-Currency market is fragmented, poorly coordinated, and difficult for clients to navigate. Many of the exchanges which trade crypto-currency products require clients to use the exchange "wallets" and maintain control over client's ability to initiate trades on their order books. Investors, therefore, trade on one exchange at a time, despite multiple exchanges quoting the same product pairs , often at very different prices. As a result, clients can lack the ability to execute at the best available price, since quite often that would require trading, in whole or in part, on different exchanges simultaneously.

BRIEF SUMMARY OF THE INVENTION

[0002] The invention is the creation of a distributed cryptocurrency smart order router (DCCSOR), which allows clients to keep control of their own private and exchange keys to their wallets and accounts, but can execute orders across multiple exchanges simultaneously. The DCCSOR software is facilitated by a peering connection to a decision engine, which subscribes to an integrated data feed of exchange bids and offers to determine the optimal package of orders to send. This software is paired with a user interface that displays potential routing via a cost calculator function. The cost calculator displays both the optimal order package the router would send based on the order parameters at the time the calculator is invoked, while also displaying the predicted savings from using the DCCSOR vis a vis the execution capabilities of a single exchange. The router is context sensitive to each client's wallets and exchange relationships configured within the client' s own network or private cloud to select the exchanges eligible to send orders to. The router also incorporates logic to determine when it would be optimal to transfer funds or cryptocurrencies between customer wallets, decentralized exchanges, and centralized exchanges to facilitate the transaction. The router utilizes an

¹ Crypto-Currency is defined herein as BitCoin, Ethereum, LiteCoin or alternative coins that are based on blockchain technology.

² Crypto-Currency products consist of a primary coin and a reference currency or coin. Examples of the different types are BitCoin-USD and BitCoin-Ethereum. integrated data-structure which creates a custom view of the available bids or offers across all included exchanges to each eligible exchange based on that data, order parameters and statistical analysis.

Figures Depicting System Flow and Screen Captures of Working System

[0003] Figure 1. Figure 1 shows the system flow including the deployment of components in both client and systems own virtual private cloud and the process flow detailing client control over their wallet and exchange keys.

[0004] Figure 2. Figure 2 shows the selection box of the working user interface where the user selected a BUY order for 500 BTC denominated in USD compared to GDAX as a single exchange.

[0005] Figure 3. Figure 3 shows the results from the working user interface where the user selected a BUY order for 500 BTC denominated in USD, compared GDAX as a single exchange.

[0006] Figure 4. Figure 4 shows the selection box of the working user interface where the user selected a SELL order for 500 ETH denominated in USD, compared to Gemini as a single exchange.

[0007] Figure 5. Figure 5 shows the results from the working user interface where the user selected a SELL order for 500 ETH denominated in USD, compared to Gemini as a single exchange.

[0008] Figure 6. Figure 6 shows the selection box of the working user interface where the user selected a BUY order for 1000 ETH denominated in BTC, compared to GDAX as a single exchange.

[0009] Figure 7. Figure 7 shows the results from the working user interface where the user selected a BUY order for 1000 ETH denominated in BTC, compared to GDAX as a single exchange.

DETAILED DESCRIPTION AND BEST MODE OF IMPLEMENTATION

[0010] The Distributed Crypto-Currency Smart Order Router (DCCSOR) is a system for clients to use to transact in cryptocurrencies at optimal prices. All cryptocurrency products are represented by a pair of currencies. In each product pair, the first is a cryptocurrency that is the primary product being traded, and the second can be either a crypto or fiat currency, which is also referred to as the pricing currency. For the sake of clarity, two product examples to illustrate both types of product-pairs are BTC-USD and BTC-ETH. In the example of BTC-USD, Bitcoin is the primary product and US Dollars is the pricing currency. In the example of BTC-ETH, Bitcoin is also the primary product but Ethereum is the pricing currency.

[0011] Design: The smart order router is designed to operate within the clients Virtual Private Cloud (VPC) and, therefore, allows its users to maintain control of their wallets and exchange API keys to interact with the exchanges that the DCCSOR sends orders to. This unique method of interaction means that the only information that needs to be transferred to the Virtual Private Cloud where the market data, decision engine and centralized UI services reside are the order parameters and Metadata. Metadata includes the information relevant to the selected cryptocurrency pair that is needed to determine what exchanges are eligible for the client as well as the position information at each exchange that determines the buying or selling quantity limits at each. The SOR will interact with each exchange on the client's behalf as if the client were exclusively trading on that exchange. The exchanges require their users to enter their Exchange API keys to trade with their platform and the instance of the DCCSOR within the client VPC will do so, but those keys never get transferred anywhere other than the exchanges.

[0012] Architecture: The architecture of the DCCSOR is a distributed Smart Order Router (SOR) connected to a Decision Engine which connects to a Market Data Service. The SOR is deployed within the client virtual private cloud, with access to the client wallets and exchange accounts as depicted in figure 1, while the Decision Engine and Market Data Service are centrally housed within their own virtual private cloud, along with a UI framework that provides access to the cost calculator as well as displays of calculated cryptocurrency market data.

[0013] Distributed Smart Order Router: The SOR performs four functions:

collection of user input and information, packaging of metadata for delivery to the decision engine, creation and sending of order instructions to exchanges, and collection of execution information for reporting back to the user or creation of subsequent metadata for a next round orders.

[0014] Collection of user input and information. For each transaction, the SOR accepts user input for the product-pair, a buy-sell indicator, quantity, limit price, and optionally a routing strategy. The buy-sell indicator and the quantity are interpreted by the SOR, in all cases, relative to the primary product. The limit price refers to the maximum (if a buy) or minimum (if a sell) price, in terms of the pricing currency, for the execution of the transaction. This information is collectively referred to in the rest of this document as the order parameters. In addition to user input of the order parameters, the Distributed SOR Client (DSC) must also have access to the exchange keys, and the Distributed SOR Decision Engine (DSDE) must have access to account balance information, at each exchange, of both the primacy and pricing currency in the cryptocurrency pair being traded.

[0015] Packaging of metadata for delivery to the decision engine. The order parameters are included along with account balance and exchange eligibility information in the Metadata that is pushed to the virtual private cloud of the centralized service to be used by the decision engine. While the Distributed SOR Client (DSC) could be enabled by the customer to utilize the keys to query each exchange directly for the account balances, the client could also provide those balances and exchange eligibility directly to a central version of the SOR that runs in the same private cloud as the decision engine. This is an important part of the design, because the client keys are only used by the Distributed SOR Client to transact with the exchanges and are not passed outside of the client's virtual private cloud when the DSC is operated in the distributed mode.

[0016] Creation and sending of order instructions to exchanges. The Distributed SOR Client receives a package of order instructions from the Decision Engine, including quantity, limit price and order type information for each exchange in a common format, or, in certain cases, instructions to transfer cryptocurrencies from one wallet to another. The DSC is responsible for formatting the specific orders to transfer instructions into the APIs of each exchange and transmitting them, while local plugins store all API keys locally and can communicate with a local wallet for p2p execution.

[0017] Collection of execution information. The Distributed SOR Client receives the execution reports back from the exchanges along with any cancelation or other messages. The DSC reports each execution to both the client system and the User Interface for display. If the order submitted by the client was not completely filled, after all the reports from the exchanges are received, the subsequent action is determined by the routing strategy and the client limit price. Depending on the routing strategy, the DSC will either leave orders unfilled on the exchanges, or cancel those orders. If orders are left unfilled on the exchanges, the Distributed SOR Client will wait until either they are filled or the client instructs the DSC to take further action. If the DSC cancels those orders, and the routing strategy warrants it, the DSC will send new metadata to the decision engine to continue trying to execute the original order.

[0018] The Distributed SOR Client software will be open-source and can prompt the user to accept an execution before proceeding. This protects the user against

compromised data feeds.

[0019] Market Data Service: The market data service is composed of a set of real time feed handlers and APIs that feed a real-time database that captures bids and offers at every exchange in time sequence. The service can create, on demand, a composite order book with all of the bids and offers across the requested exchanges. This allows the decision engine to get a view of the order book limited by the exchanges for which the client is eligible to trade. Lastly, the market data service can create historical datasets for the creation of statistical analytics that can be used by the decision engine and other user interface functions.

[0020] Decision Engine: The decision engine houses all the routing strategy logic in the system including the interpretation of client instructions and account information to frame its data acquisition, and the calculation of what orders to send to each exchange to achieve the optimal price. For sell orders, it also evaluates potential cryptocurrency transfers in situations where the client assets are not housed at the exchange with the best price.

[0021] Interpretation of client instructions. The decision engine utilizes the product- pair, buy-sell indicator and limit price to properly select the market data required to make its determination. It will request the full depth of bids to the limit price if the client is selling and the full depth of offers to the limit price if the client is buying. If the client does not have a limit price, then the decision engine will request the full depth of book on the appropriate side.

[0022] Interpretation of account information. The decision engine will use the client account information to request that the market data service return only the data for exchanges where the client has accounts. It also uses the account information, in the case of buy orders, to determine the client's buying power on each exchange, and, in the case of sell orders, the amount of the primary cryptocurrency held at each exchange that is eligible for sale.

[0023] Calculation of orders to send. The decision engine will, subject to the constraints imposed by the client account information, calculate the orders to be sent to each exchange to achieve the optimal price based on the filtered, composite order book returned by the market data service. The decision engine will aggregate the quantity per exchange it should send to each, by adding the quantity of each order in the composite book on each exchange in price order to a running total for each. For buy orders, the calculation starts at the lowest offer price and concludes when the total quantity aggregated across exchanges equals the client order size.

[0024] For sell orders, if the client has enabled the feature, the decision engine will determine both an optimal set of orders for all the exchanges the client is eligible to trade on, as well as a set of orders subject to the current constraints of the account holdings. The decision engine will then calculate, based on a statistical formula that analyzes the volatility of the product compared to the relative price advantage of trading on exchanges where the client has insufficient inventory, to determine if it would be advantageous to initiate a transfer of the product to that exchange before passing the sell orders to the SOR. This feature is based on a model that incorporates a predicted volatility calculation to assess the risk of waiting for the time that the engine predicts it will take to transfer the product.

[0025] Cost Calculator: The cost calculator software utilizes the same logic as the decision engine to determine the optimal set of orders, and then displays the results. It shows the potential aggregate execution price, cost savings, aggregate executions per exchange and individual orders the SOR would send, compared to the aggregate execution price and individual orders that would be sent to the selected individual exchange.

[0026] It also includes an interface that allows the client to choose which exchanges to include when routing, if they wish to include exchanges for any reason.

Claims

CLAIMS What is claimed is:

1. A distributed cryptocurrency smart order router comprising:

a connection to one or more data feeds from a plurality of cryptocurrency exchanges, wherein the one or more data feeds include exchange bid and offer data for a given product pair from each respective exchange of the plurality of exchanges; and a user interface that includes information about an optimal order package for trading the given product pair across the plurality of exchanges that is based on a specific client's balances and buying power on each respective exchange of the plurality of exchanges.