> For the complete documentation index, see [llms.txt](https://candora.gitbook.io/whitepaper/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://candora.gitbook.io/whitepaper/candora-grid/pulse.md). # Pulse Market Data Ingestion and Streaming Layer ### Market-State Synchronization, Event Streaming, and Telemetry Infrastructure Candora Pulse is the market-state synchronization and event-streaming layer of Candora Grid. It is responsible for ingesting, sequencing, normalizing, and distributing market events across the exchange while maintaining a consistent representation of market conditions for all participating infrastructure domains. Pulse serves as the authoritative market-state layer of the Grid. Orbit, Oracle, Risk Engine, Vault, Connect, and Cortex rely on Pulse to maintain synchronized visibility into liquidity conditions, order-book evolution, execution activity, pricing telemetry, and broader market behavior. Unlike conventional market-data systems that operate primarily as passive feed relays, Pulse continuously constructs and maintains a unified market-state model derived from multiple event sources. This allows every subsystem within Candora Grid to operate against a consistent interpretation of market conditions while preserving operational independence and clear authority boundaries. At a system level, Pulse functions as the real-time coordination fabric connecting execution, pricing, risk, settlement, analytics, and external access systems. While each subsystem maintains responsibility for its own domain, Pulse provides the common market-state context that allows those systems to remain synchronized as conditions evolve. ### Market-State Architecture Pulse operates as a partitioned event-streaming system in which each trading instrument is assigned to an isolated processing domain. Every domain maintains an independent in-memory state machine responsible for processing events in deterministic sequence order. This architecture allows the platform to scale horizontally while preserving strict ordering guarantees at the instrument level. Congestion, failure, or recovery activity affecting one market therefore remains isolated from unrelated trading pairs. Market-state evolution is represented as a continuous application of ordered events: ``` State(t+1)=Apply(State(t),Event(n)) ``` Given identical event streams, Pulse will always produce identical market-state outcomes. This deterministic behavior forms the foundation for replay, recovery, simulation, and distributed synchronization throughout the broader Grid architecture. Events may originate from multiple infrastructure domains, including execution activity generated by Orbit, valuation updates published by Oracle, collateral and liquidation telemetry supplied by Risk Engine, settlement notifications produced by Vault, external market-data feeds, and system-wide operational-state transitions. Before distribution, all incoming information is normalized into a common internal event model. This allows internal consumers to operate against a consistent event grammar independent of the original event source. ### Order Book and Market-State Management Pulse maintains real-time market representations for all supported instruments, including order-book state, liquidity conditions, execution activity, and derived market metrics. Order books evolve incrementally through deterministic state transitions rather than full-state replacement. This reduces propagation overhead while enabling low-latency synchronization across the exchange environment. As market state evolves, Pulse continuously derives additional telemetry describing the behavior of each market. These derived metrics include liquidity conditions, spread dynamics, volatility behavior, execution-flow characteristics, order-book imbalance, and broader microstructure signals. Rather than functioning solely as a distribution system, Pulse therefore acts as the shared market-perception layer of Candora Grid. Other infrastructure domains consume this telemetry to support their own responsibilities while retaining independent decision authority. Orbit consumes execution and liquidity telemetry to support execution coordination and market-protection mechanisms. Oracle consumes market-state information to support valuation validation. Risk Engine consumes liquidity and volatility telemetry for solvency modeling. Vault consumes market-state information relevant to collateral valuation and reserve monitoring. Cortex consumes market telemetry for analytics and behavioral modeling, while Connect exposes selected market-state information through external interfaces. ### Event Distribution Framework Pulse distributes market information through a low-latency publish-subscribe architecture designed for high-throughput event propagation. Subscribers consume only the market namespaces relevant to their operation, allowing infrastructure domains to scale independently while preserving synchronization guarantees. The distribution layer is optimized around asynchronous propagation, ordered delivery, bounded resource utilization, and fault isolation. Wherever possible, Pulse distributes compact state deltas rather than complete state snapshots. Snapshot publication remains available for recovery, validation, and resynchronization workflows. Because Pulse operates as the central market-state fabric of the Grid, propagation behavior is designed to remain predictable even during periods of elevated volatility. Distribution policies prioritize consistency and determinism over raw throughput whenever those objectives come into conflict. ### Throughput Management and Backpressure Control Market activity can increase dramatically during periods of volatility, liquidation activity, or sudden liquidity shifts. Pulse continuously monitors incoming event volume relative to available processing capacity and adapts propagation behavior when necessary. The objective is not to preserve all telemetry equally, but to preserve the continuity of market-state synchronization. During elevated load conditions, Pulse prioritizes the propagation of execution-critical events, order-book consistency updates, liquidation-related telemetry, settlement synchronization events, and valuation continuity signals. Lower-priority analytical streams may be reduced, compressed, or deferred when necessary to preserve deterministic processing behavior. Unlike conventional feed-throttling systems that react independently to overload conditions, Pulse coordinates degradation behavior with Orbit and Risk Engine through Candora's shared operational-state framework. This allows execution, solvency, and market-state infrastructure to respond coherently to deteriorating operating conditions rather than competing for resources during periods of stress. Pulse may additionally perform deterministic event aggregation under extreme throughput conditions, allowing large volumes of micro-updates to be consolidated without violating sequencing guarantees or altering resulting market state. ### Shared Operational-State Integration Pulse acts as one of the primary telemetry producers within Candora's shared operational-state framework. The system continuously exports measurements describing volatility, liquidity conditions, spread behavior, execution pressure, propagation latency, and market fragmentation. These signals contribute to the computation of the platform-wide operational state used by Orbit, Risk Engine, Connect, and other infrastructure domains. Pulse also consumes the shared operational state as an input into its own propagation policies. As infrastructure stress increases, synchronization behavior may become increasingly conservative to preserve consistency and deterministic operation. This bidirectional relationship allows market-state conditions to contribute to broader infrastructure awareness while ensuring that synchronization policies remain aligned with the current operating environment. ### Replay and Recovery All market-state transitions are persisted as replayable event streams and synchronized with deterministic checkpoints. Recovery follows a structured process: 1. Restore the most recent market-state snapshot. 2. Replay subsequent event streams. 3. Reconstruct market state. 4. Verify consistency. 5. Resume propagation. Market-state reconstruction follows: ``` Mrecovered = Msnapshot + Σ Δmarket_events ``` Because state evolution is deterministic, reconstructed market state remains consistent with original system behavior. This enables exact reconstruction of historical order books, liquidity conditions, execution environments, volatility regimes, and operational-state transitions without requiring approximation of prior market conditions. Pulse therefore serves as both the market-state authority and the primary recovery anchor for synchronized market visibility across Candora Grid. ### System Role Summary Candora Pulse serves as the authoritative market-state layer of Candora Grid. It provides event ingestion, state synchronization, order-book management, telemetry generation, market-data distribution, deterministic replay, and recovery coordination across the exchange infrastructure. All real-time market visibility, liquidity awareness, execution telemetry, and synchronized market-state propagation ultimately originate from Pulse, making it the foundational market-state and coordination layer of the entire Grid architecture.