Parallel Processing by Design
The market infrastructure is built to take full advantage of OpenGDP’s parallelism capabilities by isolating state on a per-market basis.- How it Works: Each trading pair has its own independent order book. For example, an order on NVDA/USDC touches only the NVDA/USDC book. Orders on different markets, such as TSLA/USDC, are entirely independent.
- The Benefit: Since markets are isolated, transactions for different pairs can be processed in parallel. Conflicts occur only when multiple transactions interact with the same order book in the same block. This design significantly increases throughput and scalability.
Trading Pairs
Markets are defined by trading pairs. Each pair consists of a base asset and a quote asset. In NVDA/USDC, NVIDIA (NVDA) is the base, and USDC is the quote. Features and Trading Lifecycle- Order Types: Supports LIMIT and MARKET orders, as well as advanced types such as IOC (Immediate or cancel), FOK (Fill or kill), and POO (Post only orders).
- Order Placement: When submitted, orders are added to the relevant order book but not matched immediately.
- End-of-Block Matching (Block FBA): At the end of each block, the matching engine processes all outstanding orders in a single batch. Orders are matched atomically against the existing book and each other.
- Fairness: The orders are executed with a FIFO queue, with orders appearing first in the block being executed before other orders that came later within the same market.
- Settlement: After matching, trades are executed and balances updated atomically within the block.
- Access Methods: The exchange can be accessed via the CLI, RPC endpoints, or directly from smart contracts through the exchange precompile.