MadExchange is a high performance trade execution engine for (crypto) exchanges. Through its microservice architecture its highly scalable, allowing to trade on any number of accounts. The Connector is very flexible, which makes integration of additional exchanges easy, fast and robust at the same time. It only needs a configuration file, which describes the endpoints and datatypes of an exchange. Data is de/serialized with the fastest serializer available for dotnet framework, provided by ServiceStack. Reflection is used to map incoming data into the common domain model. The Reflectioninformation is described by just these configfiles. As a common Data Store, the service architecture uses the distributed cache Redis, providing realtime data-access for tradeexecution. For other usecases, the Connector can be configured to store incoming data in databases or other data services, timeseries dbs, etc. The Connector supports http requests and websocket subscriptions. All control data is handled by RabbitMq, within a cqrs pattern. The result is a clean service operation abstraction.
Coinbase, Bybit, Bitmex, FTX, Blade, etc.. exchange integration will follow
Trade execution Todo :)
-Basic Components
SocketManager: *SocketSubscriptions
*RedisIntegration
HttpClientFactories:
*Configuration -> EndPointAccess
*FaultTolerance -> Resend/Reinitialisation / RateLimit
*Mapping to Domain-Models
BusIntegration
- Extension: Tracing, further ExchangeIntegrations ...
Test : Basic Components
Interface => Bybit
private: Data/Operation Access
public: Instrument Queries
*HttpRequests
*Sockets
*RedisIntegration
-Basic Components
Execution Algorithms:
*ChangePosition
*ClosePosition
*OrderPlacing/Cancellation
Registration/Deregistration/Monitoring ConnectorClients
*Request Arbitration
*Socket Support
Test: TradeExecution in Testnet
Testing/DeploymentServiceIntegration
Service can be started locally via dotnet run
(executed in the src/
directory) or ./scripts/dotnet-run.sh
shell script, by default it will be available under https://localhost:5000.
You can also run the application using Docker docker run --name -p 5000:5000 --network MadXchange (include
-d` to run the container in the background).
It is required to have the basic infrastructure up and running first (RabbitMQ, MongoDB and Redis) in the same docker network named MadXchange
. Services can be started using Docker Compose (and this file)
BTC: 1HF7XTPmkPPYzmj7AZr4eNfwVVLEZ1rogn