Skip to content
LIAW_2024_06 // NET:01

A Hardened Core on a 100G Backbone

Why the ingest path runs on dedicated 100-gigabit links and what that buys you when a venue floods the socket.

Latency budgets are spent in microseconds, and the network spends them fastest. LIAW runs ingest on dedicated 100-gigabit links so that a venue burst — the kind you see at a funding flip or a liquidation cascade — never queues behind unrelated traffic. The point is not raw throughput; it is isolation.

Each venue socket gets its own ingress lane with independent back-pressure. If one source floods, we shed its excess at the edge and stamp the gap into the ledger rather than letting it stall the core. A degraded source becomes a visible freshness signal, not a silent outage. Consumers see exactly which venue went quiet and when.

FIG.01: INGEST_BACKBONE_TOPOLOGY
  1. VENUE FEEDSINGRESSWSS · FIX · REST
  2. // CORECORE_ENGINENORMALIZEDEDUPE · STAMP · TAG
  3. PROVENANCELEDGER_SYMMETRYHASH · FRESHNESS
FIG.01: INGEST_BACKBONE_TOPOLOGY · INGRESS → CORE_ENGINE → LEDGER_SYMMETRY

The hardened core treats every link as untrusted and every event as provenance-bearing. That discipline is what lets us promise sub-millisecond normalization under load instead of only on a quiet afternoon.

§ · CONTINUE THE EXPLORATION

The ledger is always open.

Read the source-aware provenance trail, or pull the full dispatch for offline review.