Sub-Microsecond Matching. Zero Compromise Execution.

Ultency processes every order in under 800 nanoseconds using a deterministic price-time priority algorithm — delivering matched fills with zero requotes, zero asymmetric slippage, and zero order rejection events. Where legacy matching architectures introduce latency variance and execution uncertainty, Ultency replaces them with measurable, repeatable, nanosecond-precise order processing that gives every RTX5 trader a structural execution advantage.

<0.8μs
Order Matching Speed
4.2M
Orders Per Second Peak
99.999%
System Uptime SLA
0.00%
Order Rejection Rate
0.01 pips
Average Matched Slippage
12,400+
Simultaneous Order Books
The Engine

The Execution Core Powering Every RTX5 Trade

Ultency is a proprietary high-frequency order matching system built specifically for the RTX5 platform — processing incoming buy and sell orders against available liquidity using a deterministic price-time priority algorithm. It eliminates execution uncertainty, minimises latency variance, and ensures every order receives the fairest possible fill at the fastest achievable speed. Ultency is not an add-on or a third-party integration — it is the architectural foundation on which every RTX5 execution event is built.

Deterministic Execution Model

Ultency processes every order through a single-threaded event loop that guarantees identical execution behaviour regardless of system load — eliminating the non-determinism introduced by multi-threaded matching architectures.

Zero Discrimination Order Processing

Every order enters the same matching queue and is processed by the same algorithm with the same latency — no order receives preferential treatment based on account size, broker relationship, or trading frequency.

Microsecond Matching Speed

Ultency completes the full order-to-fill cycle in under 800 nanoseconds at median — achieved through lock-free in-memory order book management and kernel-bypass networking that eliminates every unnecessary processing step.

Transparent Fill Reporting

Every matched fill generates a nanosecond-timestamped execution report containing price, quantity, queue position, and latency data — giving traders and brokers complete audit-level visibility into how every order was processed.

Architecture

Engineered Without Compromise for Speed and Determinism

Ultency is built on a lock-free, event-driven processing architecture — using in-memory order book management, zero-copy data pipelines, and kernel-bypass networking to eliminate every unnecessary processing step between order receipt and execution confirmation. The result is a matching engine where latency is measured in nanoseconds, not milliseconds — and where every fill is the deterministic outcome of a transparent, reproducible algorithm.

Lock-Free Order Book Engine

Manages all open order queues in memory without mutex locks, eliminating thread contention that introduces latency variance in conventional matching systems. Every order book operation completes in constant time with zero blocking, regardless of queue depth or concurrent order volume.

Zero-Copy Data Pipeline

Routes order data from receipt to matching without memory copying operations, reducing processing overhead at every stage of the execution path. Data flows through shared memory buffers using pointer-based access patterns that eliminate redundant allocation and deallocation cycles.

Kernel-Bypass Network Layer

Bypasses the operating system network stack using DPDK and RDMA protocols to achieve direct hardware-level packet processing. Incoming order packets are read directly from the network interface card into user-space memory, eliminating kernel interrupts and context switches.

In-Memory Price Level Management

Maintains all price levels and order queues in RAM with deterministic access times, avoiding disk I/O latency entirely. Price level lookup, insertion, and deletion complete in O(1) time using a pre-allocated hash-map structure optimised for order-book access patterns.

Event-Driven Processing Model

Processes all order events in strict temporal sequence using a single-threaded event loop, guaranteeing execution order determinism. No event is processed out of sequence, and the absence of thread scheduling ensures that matching outcomes are reproducible under identical input conditions.

Hardware Clock Synchronisation

Synchronises the Ultency engine clock to nanosecond precision using PTP over dedicated network infrastructure, ensuring accurate timestamping of every matched event. All execution timestamps are traceable to a UTC reference source with a maximum divergence of ±50 nanoseconds.

Matching Logic

Absolute Fairness. Zero Ambiguity. Every Match.

Ultency implements a strict Price-Time Priority matching algorithm — where orders are matched first by the best available price and then by the chronological sequence in which they were received at the engine. Every market participant receives equal and predictable treatment regardless of order origin, size, or account type — making execution outcomes transparent and reproducible.

Price-Time Priority (FIFO) Model

Orders are matched strictly by best price first and then by the chronological sequence of receipt — ensuring that the earliest order at the best price always receives the first fill. This eliminates any possibility of queue-jumping or preferential order sequencing.

Pro-Rata Allocation for Iceberg Orders

When an iceberg order's visible quantity is matched, Ultency allocates the remaining hidden quantity proportionally across resting orders at the same price level. This ensures that large hidden liquidity is distributed fairly without disadvantaging smaller participants.

Partial Fill & Remainder Handling

When an incoming order cannot be fully matched at the best available price, Ultency executes the available portion immediately and places the remainder in the order book at the limit price. Traders receive an instant partial fill report followed by the remainder's queue position.

Self-Match Prevention (SMP) Controls

Ultency detects and prevents orders from the same account or trading entity from matching against each other — eliminating the risk of wash trades that can trigger regulatory flags. SMP rules are configurable per account with cancel-resting, cancel-incoming, and cancel-both modes.

Minimum Fill Quantity Enforcement

Traders can specify a minimum fill quantity on any order — Ultency will only execute the order if the available counterparty liquidity meets or exceeds the minimum threshold. Orders that cannot satisfy the minimum fill are held in queue rather than executed at an unacceptable partial size.

Technical Note: Unlike legacy retail matching systems that employ last-look rejection windows — allowing liquidity providers to reject orders after seeing the price — Ultency processes every order on arrival with no rejection mechanism available to any counterparty. Additionally, Ultency enforces symmetric slippage: positive slippage (price improvement) is passed to the trader at the same rate as negative slippage, eliminating the asymmetric slippage practices common in dealer-model execution venues.

Speed

Execution Speed That Makes Latency a Non-Factor

Ultency achieves consistent sub-microsecond order-to-match latency — measured from the nanosecond the order arrives at the engine input to the nanosecond the matched fill confirmation is dispatched. Latency variance is maintained within a tight P99/P50 ratio of under 1.6x, eliminating the unpredictable execution timing that affects strategies sensitive to fill timing and queue position.

~85 ns

Order Receipt & Validation

Incoming order packet is read from the kernel-bypass network buffer, deserialised, and validated against schema, instrument, and session rules.

~120 ns

Order Book Insertion

Validated order is inserted into the in-memory price level queue at the correct position based on price-time priority sequencing.

~280 ns

Price Level Matching

The matching engine evaluates the incoming order against all resting orders at eligible price levels, executing fills in strict FIFO sequence.

~180 ns

Fill Confirmation Generation

Matched fill details — including price, quantity, counterparty reference, and nanosecond timestamp — are assembled into an execution report.

~130 ns

Execution Report Dispatch

The completed execution report is serialised and dispatched to the originating session via the kernel-bypass output buffer for immediate delivery.

Industry Comparison:The median end-to-end matching latency for retail trading platforms in 2026 ranges from 5–50 milliseconds — approximately 6,000x to 60,000x slower than Ultency's sub-microsecond profile. Even among institutional-grade matching engines, Ultency's P99 latency of 1.2 microseconds places it in the top percentile of independently benchmarked execution venues operating in the FX and multi-asset class space.

Order Book

Deepest, Most Accurate Order Book in Real Time

Ultency maintains a fully consolidated real-time order book for every instrument on the RTX5 platform — aggregating liquidity from all connected brokers and liquidity providers into a single unified depth-of-market view. Every price level and queue position is updated in memory at matching speed as orders are added, cancelled, modified, and executed — ensuring that the order book displayed to traders is always an exact representation of live market conditions.

Real-Time Consolidated Order Book

Maintains a single unified view of all bids and offers across every connected liquidity source — updated in-memory at matching speed with every order add, cancel, modify, and execution event.

Multi-Tier Liquidity Aggregation

Aggregates order flow from multiple brokers, liquidity providers, and internal crossing pools into a single depth-of-market view — ensuring the order book reflects the deepest available liquidity at every price level.

Level 2 Depth Display (10 Levels)

Publishes ten price levels of bid and offer depth for every instrument — displaying aggregate quantity, number of orders, and implied order composition at each level in real time.

Order Book Imbalance Indicator

Calculates and publishes the bid-ask volume imbalance ratio in real time — giving traders an instant quantitative signal for short-term directional pressure at the top of the order book.

Historical Order Book Snapshot Archive

Stores microsecond-resolution snapshots of the full order book at configurable intervals — enabling post-trade analysis of liquidity conditions at the exact moment any order was matched or placed.

Order Book Replay for Strategy Testing

Allows traders and quant teams to replay historical order book data at full speed or in slow motion — testing how strategies would have performed against real market microstructure conditions.

Execution Quality

Every Order Treated with Absolute Execution Equality

Ultency is designed on a principle of execution neutrality — where no order receives preferential treatment based on the trader's account size, broker relationship, or order frequency. Every fill is processed by the same deterministic algorithm with the same latency and the same price-time priority rules applied without exception — making Ultency one of the most auditably fair matching engines operating in the multi-asset trading space.

No Last-Look Rejection Policy

Ultency does not implement any last-look window — every order that passes pre-trade validation is immediately admitted to the matching queue without any opportunity for counterparty rejection. This eliminates the single largest source of execution uncertainty in conventional dealer-model platforms.

Symmetric Slippage Enforcement

Price improvement (positive slippage) is passed to the trader at exactly the same rate and magnitude as negative slippage — Ultency does not capture or asymmetrically distribute price improvement. Every slippage event is logged and reported with full transparency.

Requote Elimination Architecture

Ultency&apos;s deterministic matching model eliminates the technical conditions that produce requotes in legacy systems — every valid order receives an execution or a queue placement on first submission. The concept of &quot;requote&quot; does not exist in Ultency&apos;s order processing flow.

Best Execution Price Guarantee

Every matched fill is executed at the best available price in the consolidated order book at the nanosecond of matching — with the fill price, order book state, and matching timestamp all recorded for independent verification. No order is filled at any price other than the best available.

Full Fill Transparency Reporting

Every execution generates a detailed fill report containing the fill price, executed quantity, queue position at time of match, latency from receipt to fill, and counterparty liquidity pool reference. These reports are available in real time and archived for post-trade audit.

Ultency's execution quality metrics — including fill rate, slippage distribution, and latency percentiles — are published quarterly in RTX5's Execution Transparency Report, available to all registered users and institutional partners.
Scalability

Scales Without Degradation Under Any Market Condition

Ultency is engineered to maintain consistent matching latency and throughput regardless of order volume — handling millions of order events per second during extreme market volatility events such as central bank announcements, earnings releases, and flash crash scenarios. No degradation in matching speed, order book accuracy, or fill confirmation timing occurs even under sustained peak loads.

Horizontal Scaling Architecture

Ultency distributes order book processing across multiple independent matching engine instances — each handling a partitioned set of instruments — allowing throughput to scale linearly by adding nodes without any architectural code changes. Inter-node communication uses a low-latency message bus with sub-microsecond propagation for cross-partition order routing.

Peak Load Order Buffer Management

During extreme order volume spikes, Ultency&apos;s ring-buffer architecture absorbs incoming orders without dropping or throttling — maintaining ordered queueing even when instantaneous arrival rates exceed steady-state throughput capacity. The buffer automatically drains at maximum matching speed once the spike subsides.

Graceful Degradation Under Extreme Load

Under unprecedented load conditions that exceed all buffer and scaling capacity, Ultency applies configurable rate-limiting on a per-session basis rather than rejecting orders arbitrarily. Affected sessions receive explicit rate-limit notifications with retry guidance while all queued orders continue to be matched in strict sequence.

Multi-Instance Failover Clustering

Multiple Ultency instances operate in an active-active cluster configuration with real-time state replication — ensuring that any single node failure triggers automatic workload redistribution to surviving nodes in under 50 milliseconds. No orders are lost during failover and no matching gaps occur.

Real-Time Throughput Monitoring Dashboard

A dedicated operations dashboard displays live throughput metrics — including orders per second, matching latency percentiles, buffer utilisation, and node health — enabling infrastructure teams to monitor engine performance continuously. Automated alerts trigger when any metric deviates from predefined thresholds.

Stress Test Results: Ultency was stress-tested with simulated peak loads of 12 million order events per second sustained over a 4-hour window — replicating the order flow characteristics of simultaneous NFP release, FOMC announcement, and multi-asset flash crash scenarios. Matching latency remained below 1.2 microseconds at P99 throughout the entire test period with zero order loss and zero buffer overflow events.

Risk Engine

Institutional-Grade Risk Controls at the Execution Layer

Ultency incorporates a pre-trade risk validation layer that processes every incoming order against a configurable set of risk parameters — including margin availability, position size limits, instrument trading restrictions, and account-level exposure caps — before the order is admitted to the matching queue. No order reaches the book that would breach pre-defined risk boundaries, protecting both the trader and the broker from uncontrolled exposure.

Margin Availability Check

Validates that the trader's available margin is sufficient to cover the notional exposure and margin requirement of the incoming order before admitting it to the matching queue.

Maximum Position Size Enforcement

Rejects any order that would cause the trader's total position in a single instrument to exceed the broker-configured maximum position size limit.

Instrument-Level Trading Restriction Filter

Applies instrument-specific trading restrictions — including session-based halts, regulatory suspensions, and broker-defined exclusion lists — preventing restricted orders from entering the book.

Account Exposure Cap Validation

Calculates the aggregate exposure across all open positions and pending orders for the account, rejecting any new order that would breach the broker's defined total exposure ceiling.

Maximum Order Frequency Rate Limiter

Enforces a configurable maximum order submission rate per session — preventing excessive order traffic from individual accounts that could indicate erroneous automation or market abuse patterns.

Anomalous Order Pattern Detection

Monitors order flow in real time for patterns indicative of spoofing, layering, or momentum ignition — flagging suspicious activity for compliance review and optionally blocking further orders from the flagged session.

Broker Integration: All six pre-trade risk controls are fully configurable via the RTX5 Broker Risk Management API — allowing brokers to set, update, and monitor risk parameters at the firm, group, and individual account level in real time. Changes to risk parameters take effect on the next order submission with zero engine downtime.

Connectivity

Every Professional Execution Protocol, Connected

Ultency supports connectivity via industry-standard trading protocols — enabling direct integration with broker OMS systems, algorithmic trading platforms, DMA providers, and prime brokers through documented and certified connection interfaces. Full session management, heartbeat monitoring, and automatic reconnection handling are built into every protocol adapter.

FIX Protocol 4.2 / 4.4 / 5.0 Support

Full support for FIX protocol versions 4.2, 4.4, and 5.0 SP2 — enabling certified connectivity for any OMS, EMS, or algorithmic trading platform that communicates via the Financial Information eXchange standard.

WebSocket API for Algo & HFT Clients

A low-latency WebSocket API providing full order management, real-time execution reporting, and order book streaming — optimised for algorithmic trading systems that require persistent, bidirectional communication with the matching engine.

RTX5 Native Connection Protocol

A proprietary binary protocol delivering the lowest possible latency between the RTX5 terminal and the Ultency matching engine — with pre-serialised order formats and zero-overhead framing for maximum throughput efficiency.

Broker OMS Direct Integration

Pre-built integration adapters for leading broker order management systems — enabling seamless order flow routing, position synchronisation, and account management between the broker's existing infrastructure and Ultency.

Prime Broker & DMA Gateway Connectivity

Dedicated gateway connections for prime broker and direct market access providers — supporting sponsored access models, give-up arrangements, and multi-venue order routing through Ultency's matching infrastructure.

Drop Copy Feed for Post-Trade Reporting

A real-time drop copy feed that mirrors all execution reports to a designated compliance or middle-office system — enabling parallel trade capture, position reconciliation, and regulatory reporting without impacting matching engine performance.

Analytics

Complete Visibility Into Every Execution Event

Ultency generates a comprehensive execution audit trail for every order lifecycle event — from receipt and validation through matching, partial fills, and final confirmation. All data is accessible through the RTX5 execution analytics dashboard and exportable for post-trade analysis, regulatory reporting, and strategy performance evaluation.

Full Order Lifecycle Audit Trail

Captures every state change in an order's lifecycle — from initial receipt and validation through queue insertion, partial fills, modifications, and final completion — with nanosecond timestamps at each transition.

Fill Quality Score per Order

Assigns a composite fill quality score to every executed order — factoring in price improvement, slippage, fill speed, and queue position to give traders a single metric summarising execution quality.

Slippage Distribution Report

Generates a detailed slippage distribution histogram for all executed orders — broken down by instrument, order type, and time period — showing both positive and negative slippage events with full transparency.

Order Rejection Reason Log

Logs every pre-trade validation rejection with the specific reason code — including margin insufficient, position limit exceeded, instrument restricted, and rate limit triggered — enabling traders to diagnose and resolve issues immediately.

Execution Latency Histogram

Publishes a real-time latency histogram showing the distribution of order-to-fill times across all matched orders — with P50, P95, P99, and P99.9 percentiles displayed for continuous execution speed monitoring.

Regulatory Trade Reporting Export (MiFID II, EMIR)

Exports execution data in the formats required by MiFID II transaction reporting, EMIR trade reporting, and equivalent regulatory frameworks — ready for direct submission to approved reporting mechanisms.

Algo Trading

Purpose-Built for Algorithmic and High-Frequency Trading

Ultency's deterministic matching model, sub-microsecond latency profile, and high-throughput architecture make it the ideal execution environment for algorithmic strategies that depend on consistent fill timing, predictable order queue position, and rejection-free order processing. From low-frequency systematic strategies to latency-sensitive HFT models, Ultency provides the execution infrastructure that professional algo traders require.

Deterministic Fill Timing for Time-Sensitive Strategies

Ultency's single-threaded event loop guarantees that order processing time does not vary based on system load, time of day, or concurrent user activity — giving time-sensitive strategies the execution consistency they require. Fill timing predictability is maintained at the nanosecond level, enabling calibration of strategy models against real matching engine behaviour.

Order Queue Position Predictability

Ultency's strict FIFO model means that once an order is confirmed in the queue, its position will never change except through normal matching — no priority re-ordering, no queue reorganisation, and no hidden order bumping. Algorithmic strategies can reliably model their queue position and expected fill probability.

Mass Quote & Multi-Leg Order Support

Ultency natively supports mass quote submission for market makers and multi-leg order entry for spread and options strategies — processing multiple related orders in a single atomic transaction. Each leg is validated and matched independently while maintaining execution consistency across the entire multi-leg structure.

Co-Location Service for Minimum Network Latency

RTX5 offers rack space co-location within the same data centre facility as the Ultency matching engine — reducing network round-trip time to under 10 microseconds for co-located clients. Cross-connect cables are provided directly to the matching engine network switch.

Smart Order Routing Integration

Ultency integrates with RTX5's smart order routing engine to enable algorithmic strategies to route orders across multiple liquidity pools based on configurable rules — including best-price, fastest-fill, and minimum-impact routing modes. Routing decisions execute at matching engine speed with no additional latency.

Institutional

Deploy Ultency as Your Execution Backbone

RTX5 offers Ultency as a licensed matching engine component for brokers and institutions — deployable as a standalone order matching system or integrated directly into an existing OMS and risk management infrastructure. Full API documentation, dedicated integration support, and customisable risk parameter configuration are included with every institutional licence.

White-Label Matching Engine Deployment

Deploy Ultency under your own brand as a fully white-labelled matching engine — with custom branding on all client-facing execution reports, analytics dashboards, and compliance documentation.

Custom Risk Parameter Configuration API

A comprehensive REST and FIX-based API for configuring all pre-trade risk parameters at the firm, group, and individual account level — including margin rules, position limits, and instrument restrictions.

Multi-Asset Order Book Licensing

License Ultency's order book technology for FX, equities, commodities, indices, and crypto asset classes — with each asset class configurable as a separate matching engine partition with its own risk and matching rules.

Dedicated Co-Location Infrastructure

Institutional clients receive dedicated rack space, power, and cross-connect allocations within RTX5's primary data centre — with physically segregated network paths to the Ultency matching engine.

Institutional SLA with 99.999% Uptime Guarantee

A contractual Service Level Agreement guaranteeing 99.999% matching engine availability — with defined escalation procedures, credit rebates for SLA breaches, and 24/7 dedicated institutional support.

Brokers, prop firms, and financial institutions interested in deploying Ultency as their core matching engine are invited to contact the RTX5 institutional team. We provide a full technical evaluation environment, custom deployment proposals, and dedicated integration engineering support for every institutional partner.

Resilience

Engineered to Keep Running When It Matters

Ultency is deployed on a geographically distributed infrastructure with active-active redundancy — where multiple matching engine instances operate simultaneously across separate data centre locations. Automatic failover completes in under 50 milliseconds in the event of any single node failure, ensuring zero order loss and continuous matching availability even during infrastructure incidents.

Active-Active Multi-Node Clustering

Multiple Ultency matching engine instances operate simultaneously in an active-active configuration — each processing a partitioned set of instruments while maintaining full state synchronisation with all other nodes. If any node fails, its workload is redistributed across surviving nodes without any order loss or matching interruption.

Sub-50ms Automatic Failover

In the event of any single node failure, automatic failover completes in under 50 milliseconds — with all affected order books fully reconstructed from the replicated state and matching operations resumed at full speed. No manual intervention is required at any point in the failover sequence.

Geographically Distributed Data Centre Deployment

Ultency is deployed across geographically separated data centre locations — ensuring that a complete facility failure at any single site does not interrupt matching engine availability. Cross-site replication maintains state consistency with sub-millisecond lag between regions.

Real-Time Order Book State Replication

Every order book mutation — including order adds, cancels, modifications, and matched fills — is replicated in real time to all standby nodes using a low-latency binary replication protocol. Standby nodes maintain an exact replica of the primary order book state at all times.

Zero Order Loss Guarantee on Failover

Ultency guarantees that no order is lost during a failover event — all orders that were accepted by the failing node are preserved in the replicated state and are available for matching on the replacement node within the 50-millisecond failover window. A post-failover reconciliation report confirms zero loss.

Compliance

Built to Satisfy the Strictest Regulatory Audits

Ultency generates a complete, immutable, nanosecond-timestamped audit log of every order event processed by the engine — satisfying regulatory record-keeping requirements under MiFID II, EMIR, Dodd-Frank, and equivalent frameworks. All logs are stored in a tamper-proof archive and accessible via the RTX5 compliance reporting API for on-demand regulatory examination.

Nanosecond-Stamped Immutable Order Log

Every order event is recorded in a write-once, append-only log with nanosecond-precision timestamps — creating a tamper-proof audit trail that cannot be modified, deleted, or retroactively altered by any user or system process.

Regulatory Trade Reporting Integration

Pre-built integration with approved reporting mechanisms for MiFID II, EMIR, and Dodd-Frank — enabling automated submission of trade reports in the required formats within the mandated reporting timeframes.

Best Execution Evidence Generation

Automatically generates best execution evidence packages for every trade — including the order book state at the time of execution, available price levels, matching latency, and fill quality metrics — satisfying MiFID II RTS 28 requirements.

Suspicious Order Pattern Flagging

Real-time surveillance algorithms monitor order flow for patterns indicative of market manipulation — including spoofing, layering, and cross-venue manipulation — and flag suspicious activity for immediate compliance review.

Compliance Report Export in Required Formats

Exports compliance and audit data in all major regulatory formats — including FpML, ISO 20022, and custom CSV/XML schemas — with configurable report scheduling for daily, weekly, and on-demand regulatory submissions.

Tech Specs

Complete Ultency Technical Specification

The full Ultency technical specification is available below for engineering teams evaluating integration or deployment.

SpecificationValue
Order Matching Latency (P50)< 0.8 μs
Order Matching Latency (P99)< 1.2 μs
Order Matching Latency (P99.9)< 2.1 μs
Maximum Order Throughput4,200,000 orders/sec
Supported Order TypesMarket, Limit, Stop, Stop-Limit, Iceberg, FOK, IOC, GTC, GTD, MOO, MOC
Supported ProtocolsFIX 4.2/4.4/5.0 SP2, WebSocket, RTX5 Native Binary, Drop Copy
Order Book Depth (Price Levels)Unlimited (10 levels published to L2 feed)
Maximum Simultaneous Instruments50,000+
Clock Synchronisation Precision± 50 ns (PTP IEEE 1588v2)
Failover Time< 50 ms (active-active clustering)
Data Persistence ModelIn-memory primary, synchronous replication to standby
Deployment OptionsOn-Premise, Co-Location, Private Cloud, Hybrid
Operating System SupportLinux (RHEL 9+, Ubuntu 22.04+), kernel-bypass DPDK 23.11+
API Documentation FormatOpenAPI 3.1, FIX Repository, Interactive Sandbox
FAQs

Every Technical Question About Ultency, Answered

Expand any question below for a detailed, technically precise answer about the Ultency Matching Engine.

Ultency is a proprietary, high-performance order matching engine purpose-built for the RTX5 platform. Unlike standard retail matching systems that process orders through multi-threaded architectures with variable latency and optional last-look rejection windows, Ultency uses a single-threaded, lock-free, event-driven processing model that matches every order deterministically in under 0.8 microseconds at median. There is no last-look rejection, no asymmetric slippage, and no order priority manipulation — every order is processed by the same algorithm with the same rules.
Benchmarks

Performance Independently Measured and Publicly Reported

RTX5 publishes quarterly Ultency performance benchmark reports — independently verified by a third-party financial technology auditor — covering order matching latency distributions, throughput under simulated peak load, fill quality statistics, and execution neutrality metrics across all instrument classes processed by the engine during the reporting period.

Q1 2025

P50: 0.78μs
Latency
4.1M orders/sec
Throughput
99.94%
Fill Quality

Introduced hardware clock synchronisation upgrade achieving ±50ns UTC accuracy — reducing timestamp jitter by 65% compared to Q4 2024.

Q2 2025

P50: 0.76μs
Latency
4.15M orders/sec
Throughput
99.95%
Fill Quality

Deployed zero-copy data pipeline v2, eliminating the last remaining memory copy operation in the order processing path and reducing P99 latency by 8%.

Q3 2025

P50: 0.74μs
Latency
4.2M orders/sec
Throughput
99.96%
Fill Quality

Completed 12M orders/sec sustained stress test over a 4-hour window with zero order loss — validating horizontal scaling architecture under extreme simulated conditions.

Q4 2025

P50: 0.72μs
Latency
4.25M orders/sec
Throughput
99.97%
Fill Quality

Achieved the lowest-ever P99.9 latency of 1.9μs following kernel-bypass networking stack upgrade to DPDK 23.11 with hardware RSS offloading.

Full Reports Available: Complete quarterly benchmark reports — including raw latency histograms, throughput time series, and execution neutrality analysis — are available for download on the RTX5 documentation portal. Institutional clients may request raw execution data under NDA for independent evaluation by their own technology and compliance teams.

Experience Ultency Execution Quality From Your First Trade

Sub-microsecond matching speed, zero last-look rejection, and full execution transparency reporting — available to every RTX5 user from the moment you place your first order.

<0.8μs Order Matching Speed99.999% Uptime SLA18M+ Orders Processed Daily