BookUpdateBus

BookUpdateBus is a compile-time alias of the generic EventBus specialised for order-book updates (BookUpdateEvent).
Mode is chosen via the USE_SYNC_MARKET_BUS macro:

#ifdef USE_SYNC_MARKET_BUS
using BookUpdateBus = EventBus<
    pool::Handle<BookUpdateEvent>,
    SyncPolicy<pool::Handle<BookUpdateEvent>>>;
#else
using BookUpdateBus = EventBus<
    pool::Handle<BookUpdateEvent>,
    AsyncPolicy<pool::Handle<BookUpdateEvent>>>;
#endif

using BookUpdateBusRef =
    EventBusRef<pool::Handle<BookUpdateEvent>, BookUpdateBus::Queue>;

Purpose

Deliver book-update events from exchange connectors to multiple consumers (order books, strategies, metrics, loggers) with minimal latency and without extra dynamic allocations inside the bus itself.

Responsibilities

Aspect	Details
Fan-out	Writes each BookUpdateEvent handle into a dedicated SPSC queue per subscriber.
Policy	SyncPolicy → deterministic, tick-barrier coordination for simulation/tests. AsyncPolicy → lock-free delivery for live trading.
Reference type	BookUpdateBusRef lets producers publish without owning the full bus object.

---

Internal Behaviour

• Queues: One ring buffer per subscriber eliminates contention between consumers.
• Sync mode: After publishing a tick, the bus blocks until every subscriber signals completion via TickBarrier.
• Async mode: Publisher writes and returns immediately; subscribers consume at their own pace.
• Memory: The bus itself does not allocate after construction.
  Event objects (not the bus) may come from a pool when they carry containers or other heap-backed data; plain POD events are created directly.

Typical Use Cases

Producer	Consumers
BybitExchangeConnector	NLevelOrderBook, trading strategies, ExecutionTrackerComponent
Historical tick-replayer	Back-test harness, benchmarking tools

Notes

• Thread-safe under the guarantees of the chosen policy (lock-free vs. barrier-sync).
• Define/undefine USE_SYNC_MARKET_BUS before including the header to switch modes.
• One bus instance per market-data feed is recommended to avoid cross-core cache traffic.