trading-matching-service

Description

A financial trading matching service. The service accepts buy/sell orders and matches buy and sell orders that have the same price. Any valid order should have the following information that is it a buy/sell order, quantity, and is it a market price or limit price? If orders have the same price, the priority will be determined by their time stamp (FIFO). Pending orders queue up in the system until they are filled or canceled.

Features

Easy to upgrade components for higher availability, scalability... There are three critical components used in this project, which are message queue, priority queue, and order store. Currently, they are implemented in memory for POC only. To make the system more production-ready, you can choose some external solutions and import them to the system easily by implementing the predefined interface.

In-memory match engine The matching algorithm is run at the memory level. The match engine doesn't need to query the disk for the order information.

High request concurrency supported When the server receives a request, this request will be pushed into the message queue. This design means the server would not burst out of system resources and can handle the requests at a certain speed.

Data consistency and data recoverability from server crash promised Since we can choose another more stable service as an alternative for a message queue. For example, Rabbitmq helps us to keep the messages in the queue if they are not ACKed from the server. This support is considered especially for the case that the server is down because of some unexpected disasters like out of electricity or server panic.

How To Run

Run server

make install-swag # run it if you don't have swag command.
make run

Run server in docker

make rund

NOTE: After running the server, you can open your browser with http://127.0.0.1:9000/swagger-ui/ to use swagger ui to fire requests.

How to Test

Run test

make test

NOTE: If you don't want use swagger to test the service, you can add test cases in ${REPO}/test/unittest/matching_test.go.

How To Place An Buy/Sell Order Or Cancel An Order

Create Order Example

curl -X 'POST' \
  'http://localhost:9000/api/v1/orders' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
  "order_kind": ${the_order_kind},
  "price": ${the_price},
  "price_type": ${the_price_type},
  "quantity": ${the quantity}
}'

Cancel Order Example

curl -X 'DELETE' \
  'http://localhost:9000/api/v1/orders/${the_order_id}' \
  -H 'accept: application/json'

Service Architecture

Flowcharts

Flowchart of match engine handling a message

graph TB
Start((start)) --> Pull
Pull[Pull a message from queue] --> MessageKind{What kind is this message?}
MessageKind -->|Create| OrderKind{What kind is this order}
MessageKind -->|Cancel| Cancel[Cancel Handler]
OrderKind --> |Buy| Buy[Buy Handler]
OrderKind --> |Sell| Sell[Sell Handler]
Buy --> Finish((Finish))
Sell --> Finish
Cancel --> Finish

Loading

Flowchart of buy handler handling a buy order

graph TB
Start((start)) --> Cond1{Any order in sell queue? }
Cond1 --> |Yes| TryEnqueue1{Buy order quantity > 0}
Cond1 --> |No| TryEnqueue2{Buy order quantity > 0}
TryEnqueue1 --> IsPriceMatch{Is price matched}
TryEnqueue2 --> |No| Finish((Finish))
TryEnqueue2 --> |Yes| Enqueue[Enqueue order to Buy Queue]
Enqueue --> Finish
IsPriceMatch --> |Yes| HandleTrade[Handle the trade]
IsPriceMatch --> |No| TryEnqueue2
HandleTrade --> Cond1

Loading

Flowchart of sell handler handling a sell order

graph TB
Start((start)) --> Cond1{Any order in buy queue? }
Cond1 --> |Yes| TryEnqueue1{Sell order quantity > 0}
Cond1 --> |No| TryEnqueue2{Sell order quantity > 0}
TryEnqueue1 --> IsPriceMatch{Is price matched}
TryEnqueue2 --> |No| Finish((Finish))
TryEnqueue2 --> |Yes| Enqueue[Enqueue order to Sell Queue]
Enqueue --> Finish
IsPriceMatch --> |Yes| HandleTrade[Handle the trade]
IsPriceMatch --> |No| TryEnqueue2
HandleTrade --> Cond1

Loading