Rusty_Josh

Single asset exchange usefull to emulate real order book base trading.

Links

Queue

• https://doc.rust-lang.org/std/collections/vec_deque/struct.VecDeque.html#
• https://doc.rust-lang.org/std/collections/vec_deque/struct.VecDeque.html#method.pop_front

Libs

• https://dev.to/fadygrab/learning-rust-15-how-you-can-organize-you-rust-code-with-modules-2c28

json

• https://github.com/serde-rs/serde?tab=readme-ov-file

Web API

• https://actix.rs/docs/

To Do

• Vec deque
• libs
• New order handling (matching func)
• Check top book refresh on order handling
• Timestamp
• Study web API
• Add partial flag to resting order
• IMPORTANT: The matched info from the limit order are wrong
• A general check of the matches vector
• Test tot order handling
• Event: new market order when rest size is insufficient
• Add market order in IO

Bugs

• the new_order_handling function panics when there is no best bid aor ask at the moment that hread 'main' panicked at src/order_book_mod.rs:226:75: called Option::unwrap()on aNone value
  • Create the code safe if there are no orders on the book (even just on one side)
  • It is better an exception handling or a simple if option?

Ongoing

• Match incoming Order
• Engine return a list of match struct for flexibility
• Testing engine

Ideas

• Add Best BA to the orderbook struct
• Put in the Order Book struct all the functions to avoid complicating the code with Ownership
• Add a general check in the OrderBook to see if there are some problems

Road Map

1. Orderbook functionalities: insert, remove quotes, best b/a
2. book visual for debug
3. Matching engine
4. Market data export
5. Json based messages to be globally compatible
6. Local participant (co-location)
7. remote connection (http).
8. Account manager (unsafe for testing)

Client

Virtual Partecipants

Type of Participants

• Market Maker : Liquidity provvider, spread can be influenced by volatility
• Retail: Source of noice, small orders, bias against news
• Shorters: bias agaist market
• Big inverstors: Block trading during the day
• Profesionals: Complicated trades

Bots

• with python
• create subjects (retail, big institutions, market maker) with bias and peculiarities (volume, long short etc)

Enviroment

• news
• bull or bear market

Database

It will be usefull to use a sql (possible MySQL) database to keep all tidy all the inforamtion not only for the normal operations but also to summarize an entire session to study.

• Create the graph
• Study how to insert it in the program

Server

• To send the market data first it needs to be made in a format transladable to JSON. Check the best AB funtion or make a new one

Clearing House

Income

• Check for wash trading (needs a list of all active orders)
• Check margin

Matching Engine

• fn incoming_orders_processor : Should be used to route Ordes to the Lim_O handling or market_O handling.

• incoming orders

• Check if the new order ID is in the list of processed orders

• Need a vector with struct to keep all matches containing:

  • ask/bid gen info
  • partial fil volume
  • who was the resting order
  • time stamp

• for partial fills ad a letter to the end of ID

/// - resting:
///     - true = bid was the resting order
///     - false = ask was the resting order
pub struct Match{
    id_b:,
    id_a:,
    volume_filled:u32,
    price:u32,
    time_stamp:,
    resting: bool
}

Orders

Types

• Limit ✅
• Market ✅
• Hidden ❌

Modifications

Rules:

ID

The ID is created with in the Clearence house

Format

123456789-55555-11-22

1. Order code
2. Partecipant
3. Modification
4. Partial fill

---

Time stamps

An Order should have multiple timestamps for:

• Arriving at the Exchange
• Been inserted to the Book
• Full or partial fill
• Modifications
• Cancelations

This multiple possibilities require a flexible array, a good one can be a simple Queue that preserves the order of the events. This struct is for internal use ONLY.

pub struct Order {
    pub id: u8,
    pub size: u32,
    pub price: f64,
    pub side: bool,
    pub o_type: u8
    pub time: VecDeque<TimeStamp> // to do
}
pub struct TimeStamp {
    pub event: u8
    pub time: Timestamp
}

Order types

• 0: Market
• 1: Limit

Side

• true -> Bid
• false -> Ask