Courtesy of DeepSeek
This diagram represents a simplified, high-level architecture. Real-world systems are far more complex and have additional layers for risk management and monitoring.
+-----------------------------------------------------------------------+
| MARKET MAKING BOT |
| |
| +-------------------+ |
| | Configuration & | |
| | Strategy Engine | |
| +-------------------+ |
| | |
| v |
| +-------------------+ +-------------------+ |
| | Pricing & Quote | <--> | Risk & Inventory | |
| | Engine | | Management | |
| +-------------------+ +-------------------+ |
| | ^ | |
| v | v |
| +-------------------+ +-------------------+ |
| | Order Management | | Hedging Engine | |
| | System (OMS) | | (Optional) | |
| +-------------------+ +-------------------+ |
| | | |
+---------|------------------------------|------------------------------+
| |
v v
+---------|------------------------------|------------------------------+
| EXCHANGE CONNECTION LAYER |
| |
| +-------------------+ +-------------------+ |
| | Market Data | | Order Execution | |
| | Feed Handler | | Handler | |
| +-------------------+ +-------------------+ |
| | ^ | ^ |
+---------|--------------------------------|----------------------------+
| |
v v
+---------|--------------------------------|----------------------------+
| EXCHANGE(S) |
| |
| +-----------------------------+ |
| | Order Books, Trades, etc. | |
| +-----------------------------+ |
+-----------------------------------------------------------------------+
Let's break down what each component does:
-
Configuration & Strategy Engine This is the brain that defines how the bot will market make.
Inputs: Human-defined parameters (e.g., spread width, order size, maximum inventory, target symbols).
Function: Calculates the core logic for bid/ask prices and sizes based on the chosen strategy (e.g., fixed spread, inventory-sensitive, volatility-adjusted).
-
Pricing & Quote Engine Continuously calculates the exact bid and ask prices to post. It uses the market's mid-price and applies the defined spread and skew (e.g., if inventory is long, lower the bid price to discourage buying).
Inputs: Live market data from the exchange and parameters from the Strategy Engine.
Outputs: The desired quotes (price, size, side) to be sent to the market.
-
Risk & Inventory Management Monitors the bot's exposure. It tracks how much of an asset the bot holds and calculates the associated risk. If the inventory exceeds a predefined limit, it signals the Pricing Engine to adjust quotes to reduce the position (e.g., sell more aggressively if too long).
Inputs: Current positions (inventory), PnL, and filled orders from the OMS.
-
Order Management System (OMS) Manages the lifecycle of all orders. It sends new orders, cancels old ones, and updates existing orders as market conditions change. It ensures the bot's intentions are accurately reflected in the order book.
Inputs: Quotes from the Pricing Engine.
-
Hedging Engine (Optional but Critical for FX/CFDs) In markets where it's possible (e.g., using futures or a different venue), this engine immediately executes an offsetting trade to "hedge" the inventory risk and lock in the spread. For example, if a bot sells BTC on the spot market, it might buy a BTC perpetual futures contract to become delta-neutral.
Inputs: Filled orders from the OMS (e.g., the bot just sold 1 BTC, so it now has a short BTC position).
-
Exchange Connection Layer Market Data Feed Handler:
Function: Establishes a low-latency connection to the exchange's data feed (often via WebSocket). It receives real-time updates for order books, recent trades, and ticker information, then parses and normalizes this data for the Pricing Engine.
Order Execution Handler:
Function: Establishes a secure connection (using API keys) to send orders, cancel orders, and check the status of orders. It receives confirmations (fills, cancellations, errors) from the exchange and relays them back to the OMS and Risk Management modules.
-
Exchange(s) The external venue where the trading occurs. It provides the market data and executes the orders sent by the bot.
Market Data In: The exchange sends live market data (e.g., the latest order book snapshot) to the Market Data Feed Handler.
Price Calculation: The handler passes this data to the Pricing & Quote Engine. The engine, guided by the Strategy Engine and current risk from Risk & Inventory Management, calculates new bid/ask quotes.
Order Submission: The new quotes are sent to the OMS, which decides to send, cancel, or update orders. The Order Execution Handler transmits these commands to the exchange.
Fill & Feedback: The exchange executes a trade (a "fill") against the bot's order. The fill confirmation is sent back through the Order Execution Handler to the OMS.
Risk Update: The OMS notifies the Risk & Inventory Management module of the fill, updating the bot's inventory and PnL.
Hedging (Optional): The Risk module may trigger the Hedging Engine to execute an offsetting trade on another market to neutralize the newly acquired risk.
Loop: This entire process runs in a continuous loop, often millions of times per day, with the bot constantly adjusting its quotes in response to the market.
Before the library can be compiled on Linux, make sure the following dependencies have been installed:
# For apt
sudo apt install build-essential pkg-config libssl-dev
# For yum
sudo yum groupinstall "Development Tools"
sudo yum install openssl-devel