Skip to content
Terminal Ledger Desk
Directory

繁體中文

文章與來源記錄

瀏覽文章和機構來源記錄。文章內文以英文提供,請直接核對原始來源。

已發布研究簡報
208
核心專題集群
22

文章與來源記錄

ZH-HANT · TERMINAL-LEDGER-DESK

文章與來源記錄

瀏覽文章和機構來源記錄。文章內文以英文提供,請直接核對原始來源。

Matching Engine Clob Colocation Ptp

Smart Order Router (SOR) Engineering: Latency-Equalized Sweeps & Poisson Fill Models

Simultaneously firing child IOC slices from Secaucus to Cboe (2 us), Nasdaq Carteret (211 us), and NYSE Mahwah (250 us) lets fast local fills tip off remote venues; institutional SORs equalize arrival timestamps and optimize net-of-rebate Poisson fill probabilities.

核心提要:Staggering outbound child order departures by (tau_max - tau_v) synchronizes arrival across NY4, Carteret, Mahwah, and IEX—lifting sweep fill rates from 61.5% to 94.2%.

Marcus Thorne, CQF
Matching Engine Clob Colocation Ptp

PCIe Gen5 TLP DMA, Intel DDIO & AMD Cache Stashing: Avoiding Leaky L3 Misses

When a kernel-bypass NIC DMAs inbound market data across PCIe, Intel DDIO injects frames directly into 2 of 11 CPU L3 cache ways—unless bloated RX rings trigger the Leaky DMA cliff and force 85-nanosecond DRAM misses.

核心提要:Intel DDIO and AMD TPH/SDCI inject inbound PCIe DMA packets directly into CPU L3/L2 cache (~16 ns), bypassing 88-ns DDR5 DRAM round-trips.

Marcus Thorne, CQF
Matching Engine Clob Colocation Ptp

Wire-to-Wire Latency Telemetry: HdrHistogram, Coordinated Omission & Optical Taps

Internal application timers lie by up to 50x during microburst stalls due to Coordinated Omission and unread NIC ring queues; institutional latency audits correlate passive optical wire taps into lock-free HdrHistograms.

核心提要:Starting internal timers after recv/poll hides every microsecond packets sat unread inside the NIC RX ring; always anchor t0 to NIC hardware RX timestamps or external optical taps.

Marcus Thorne, CQF
Mt5 Ctrader Bridge Dealer Architecture

MT4 Single-Process Lock Contention vs MT5 Distributed Server Cluster Scaling

Under NFP tick storms, monolithic MT4 servers serialize quote updates, plugin hooks, and chart history requests behind shared process locks—forcing multi-server fragmentation that MT5 solves via four decoupled server roles.

核心提要:Monolithic MT4 servers couple client sockets, historical chart queries, and synchronous C++ plugin DLLs inside one process—causing severe NFP queue blowups.

Marcus Thorne, CQF
Mt5 Ctrader Bridge Dealer Architecture

MT5 & cTrader Access Server Proxies: DDoS Scrubbing, Anti-Sniffing & Backbone Routing

Exposing a core Trade Server IP directly to the public internet invites 400 Gbps volumetric DDoS extortion; multi-region Access Server proxies cloak origin racks inside LD4/NY4 while accelerating TLS handshakes over private backbones.

核心提要:Core MT5 and cTrader Trade Servers must sit on RFC 1918 private VLANs in LD4/NY4 with zero public internet routes, reachable exclusively via whitelisted Access POP tunnels.

Marcus Thorne, CQF
Mt5 Ctrader Bridge Dealer Architecture

Liquidity Bridge Architecture: oneZero Hub, PrimeXM XCore & Centroid Datapaths

Institutional liquidity bridges sit at the heart of Equinix LD4, NY4, and TY3—translating retail MT5/cTrader tickets into sub-millisecond multi-LP FIX 4.4 sweeps while enforcing markup matrices and credit limits.

核心提要:Enterprise bridges (oneZero Hub, PrimeXM XCore, Centroid) colocate bare-metal inside Equinix LD4/NY4/TY3 to cross-connect directly into Tier-1 LP FIX 4.4 engines.

Marcus Thorne, CQF
Mt5 Ctrader Bridge Dealer Architecture

A-Book STP, B-Book Internalization & C-Book Hybrid Dealing Desk Classifiers

Modern institutional brokerages rarely operate pure 100% A-Book or blind 100% B-Book silos; real-time quantitative risk engines score account Sharpe, markout toxicity, and aggregate net exposure to dynamically hedge tail risk.

核心提要:Natural two-way client flow internalization crosses 68–82% of volume internally with <1ms fills, zero LP Last-Look rejections, and 81% lower Prime Broker clearing fees.

Marcus Thorne, CQF
Mt5 Ctrader Bridge Dealer Architecture

Virtual Dealer Plugins vs True Market Execution: Auditing Asymmetric Slippage

Rogue dealing plugins hold incoming client orders for 200 to 1,500 milliseconds—filling instantly if the market moves in the dealer favor while requoting or slipping if the market moves in the client favor.

核心提要:An asymmetric 500-ms Virtual Dealer plugin acts as a free lookback option against the trader, invisibly draining ~0.68 pips per leg even when advertised spreads read 0.0!

Marcus Thorne, CQF
Mt5 Ctrader Bridge Dealer Architecture

MT5 Execution Modes & Fill Policies: Instant, Request, Market, Exchange & IOC/FOK

Hardcoding ORDER_FILLING_FOK inside an MQL5 Expert Advisor causes instant 10030 (Unsupported Filling Mode) rejections when deployed on an IOC/Return ECN broker—requiring bitmask inspection of SYMBOL_FILLING_MODE.

核心提要:Inspect SYMBOL_FILLING_MODE bitmasks dynamically in every MQL5 EA (preferring IOC on ECN bridges, FOK second, Return third) to eliminate Retcode 10030 rejections.

Marcus Thorne, CQF
Mt5 Ctrader Bridge Dealer Architecture

Decoding MT5 & cTrader Client Journal Logs: Millisecond TCA, Requotes & Off Quotes

Every MetaTrader 5 MQL5/Logs/*.log file and cTrader Journal records a multi-line millisecond state trace—separating client WAN transit from Trade Server queuing, bridge rejections, and LP Off Quotes (10021).

核心提要:Subtracting MT5 "accepted" timestamp from "done" timestamp (Delta_2) isolates pure broker Bridge + LP execution latency from your personal internet ping (Delta_1).

Marcus Thorne, CQF
Mt5 Ctrader Bridge Dealer Architecture

Tick-Filtering Gateways: Off-Market Spike Suppression & Quote Aggregation Throttles

A single erroneous 80-pip rogue quote from one Liquidity Provider can falsely trigger thousands of resting Stop-Loss and Stop-Out liquidations unless caught inline by multi-source median Z-score tick filters.

核心提要:Spatial cross-LP median + Median Absolute Deviation (MAD) filters block 99.98% of single-LP bad ticks in 3.4 us while adding 0.0 ms of lag to genuine multi-LP NFP jumps!

Marcus Thorne, CQF
Mt5 Ctrader Bridge Dealer Architecture

Spotware cTrader Open API (Protobuf) vs cTrader FIX 4.4: Quotas, Auth & Latency

Spotware cTrader uniquely ships every live retail and institutional trading account with both OAuth2 Google Protobuf Open API access (50 msg/sec) and direct FIX 4.4 QUOTE/TRADE credentials—each suited to distinct quant workloads.

核心提要:Spotware cTrader grants built-in per-account FIX 4.4 (QUOTE + TRADE) and OAuth2 Google Protobuf Open API access with zero $50k institutional deposit gates.

Marcus Thorne, CQF
Mt5 Ctrader Bridge Dealer Architecture

MT5 Netting vs Hedging Position Ledgers: Orders, Deals, CloseBy & Margin Math

When MetaQuotes launched MT5 with institutional Netting-only accounting, retail FX EA developers revolted until Build 1325 added Hedging—creating two distinct server ledger modes with radically different margin and CloseBy behavior.

核心提要:MT5 strictly separates Orders (instructions), Deals (immutable executions), and Positions (net or ticket inventory linked via persistent POSITION_IDENTIFIER).

Marcus Thorne, CQF
Mt5 Ctrader Bridge Dealer Architecture

OTC FX Last-Look Hold Windows: FX Global Code Principle 17 & Firm Liquidity TCA

Disclosed OTC FX streams quote razor-thin 0.1-pip spreads guarded by 2 to 35 millisecond Last-Look latency shields—forcing institutional bridges to weigh headline tightness against rejection information leakage.

核心提要:Headline top-of-book FX spreads are meaningless without Last-Look rejection rates and hold times: a 0.30-pip Firm CLOB beats a 0.10-pip 30ms-LL stream during volatility.

Marcus Thorne, CQF
Mt5 Ctrader Bridge Dealer Architecture

Margin Stop-Out Liquidation Engines: Tick Evaluation, Smart Partial Close & NBP

When a flash crash drives thousands of accounts below 50% Margin Level simultaneously, server liquidation engines must evaluate floating cross-currency equity in nanoseconds and unwind positions without triggering self-reinforcing cascades.

核心提要:Pre-solving exact liquidation prices into symbol-inverted min/max heaps lets trade servers detect Margin Stop-Outs across 85,000 accounts in 4 nanoseconds per tick.

Marcus Thorne, CQF
Mt5 Ctrader Bridge Dealer Architecture

PAMM, MAM & Copy-Trading Engines: Master-to-Child Allocation Latency & Slippage

Executing 1,500 separate follower Market orders sequentially when a Master signal fires sweeps every LP rung and leaves Follower #1,500 4.5 pips behind; institutional PAMM and Block-MAM engines aggregate pre-trade into one LP sweep.

核心提要:Sequential 1,500-account EA copiers self-sweep the LP order book and inflict up to 4.5 pips of structural slippage on tail followers.

Marcus Thorne, CQF
Mt5 Ctrader Bridge Dealer Architecture

FX Swap Rollover Engines: Triple-Wednesday T+2 Settlement & Tom-Next Batch Math

Spot FX settles T+2 (and T+1 for USD/CAD), forcing automated 17:00 EST rollover engines to accrue 3-day weekend interest on Wednesday afternoons (Thursday for T+1 pairs) via institutional Tom-Next forward points.

核心提要:Triple-Wednesday FX swaps occur because Spot FX settles T+2: rolling past Wednesday 17:00 EST advances value date from Friday across the weekend to Monday (3 days).

Marcus Thorne, CQF
Mt5 Ctrader Bridge Dealer Architecture

WebAssembly & WebGL Browser Trading Terminals: Holding 16.6ms 60-FPS Frame Budgets

Mutating React virtual-DOM nodes on 500 ticks/sec freezes browser trading terminals in 45-millisecond garbage-collection and layout-reflow jank; WebAssembly WebWorkers paired with WebGL2 instanced rendering hold <2ms frame times.

核心提要:Never parse high-frequency market data or mutate HTML DOM tables on the browser Main Thread; offload WebSockets and order-book math into a Rust/C++ WebAssembly WebWorker.

Marcus Thorne, CQF
Mt5 Ctrader Bridge Dealer Architecture

Prime-of-Prime Net Open Position (NOP) Gates: Credit Throttling in FX Bridges

When a broker exhausts its Prime-of-Prime Net Open Position (NOP) credit ceiling during a one-way Gold or Yen rally, upstream FIX orders instantly fail with OrdRejReason=3 unless cross-LP netting and pre-warning throttles engage.

核心提要:Running multiple bilateral Prime-of-Prime accounts with an NOP-blind SOR locks gross margin on opposing Long LP-A / Short LP-B positions and triggers variation-margin calls!

Marcus Thorne, CQF
Mt5 Ctrader Bridge Dealer Architecture

MQL5 High-Speed EA Architecture: OnTick Drop Behavior & OrderSendAsync Pipes

Blocking inside MQL5 OnTick with synchronous OrderSend freezes your Expert Advisor for an entire network round-trip while incoming ticks silently drop; OrderSendAsync dispatches 8+ parallel orders in under 50 microseconds.

核心提要:MQL5 drops incoming NewTick events whenever OnTick is still executing—making blocking synchronous OrderSend calls blind during volatile bursts.

Marcus Thorne, CQF
Mt5 Ctrader Bridge Dealer Architecture

Mobile Trading Terminals: iOS/Android Socket Suspension, QUIC & Push Execution

Mobile operating systems freeze background TCP sockets within 5 to 30 seconds of screen lock, causing 2.5-second TLS reconnect delays when traders tap a push alert unless QUIC 0-RTT session resumption and state delta-sync engage.

核心提要:Because iOS and Android suspend background sockets within seconds of screen lock, mobile terminals must combine TLS 1.3 / QUIC PSK session resumption with sequence-delta catch-up (<180 ms wake).

Marcus Thorne, CQF
Mt5 Ctrader Bridge Dealer Architecture

Trading Terminal Binary Hardening: Certificate Pinning, DPAPI/Keychain & Anti-Tamper

Commodity infostealer malware targets plaintext config files and unencrypted trading terminal credential stores; institutional client hardening enforces Secure Enclave/DPAPI hardware binding, SPKI TLS pinning, and scope-split API keys.

核心提要:Never store FIX passwords or OAuth tokens in plaintext .env files or browser localStorage; bind secrets to TPM 2.0 DPAPI, macOS/iOS Secure Enclave, or systemd-creds.

Marcus Thorne, CQF
Mifid Rts6 Pretrade Kill Switch Controls

SEC Rule 15c3-5 Market Access Control: Ban on Unfiltered Sponsored Access & Annual CEO Attestation

Adopted after the May 2010 Flash Crash, SEC Rule 15c3-5 eliminated unfiltered sponsored access across US equities and options—mandating pre-trade credit, erroneous-order, and regulatory gates under exclusive broker-dealer control.

核心提要:SEC Rule 15c3-5 outlawed unfiltered ("naked") sponsored access across all US securities exchanges and Alternative Trading Systems (ATSs).

Marcus Thorne, CQF
Mifid Rts6 Pretrade Kill Switch Controls

MiFID II RTS 6 Compliance: Annual Self-Assessment, 2x Stress Testing & DEA Governance

Commission Delegated Regulation (EU) 2017/589 (MiFID II RTS 6) codifies Europe strictest algorithmic trading rulebook—mandating annual validation reports, 2x peak-volume stress tests, and real-time Direct Electronic Access controls.

核心提要:MiFID II RTS 6 governs both proprietary algorithmic firms and Direct Electronic Access (DEA) providers across all EU trading venues.

Marcus Thorne, CQF
Mifid Rts6 Pretrade Kill Switch Controls

Inline FPGA Pre-Trade Risk Gateways: Cut-Through FCS Stomping & Sub-Microsecond Credit Collars

To enforce SEC Rule 15c3-5 and MiFID II RTS 6 without surrendering queue priority in NY4 or Aurora DC3, Tier-1 prime brokers deploy cut-through FPGA gateways that inspect SBE/OUCH frames in 310 nanoseconds.

核心提要:Cut-through Layer-1/2 FPGAs enforce 12+ pre-trade risk rules in ~310 nanoseconds with <16 ns jitter, preserving colocation queue competitiveness.

Marcus Thorne, CQF
Mifid Rts6 Pretrade Kill Switch Controls

Dynamic Pre-Trade Price Collars & Fat-Finger Guards: NBBO Reference Anchoring vs Static Ticks

Static percentage price collars anchored to yesterday close fail catastrophically during earnings gaps and CPI releases; institutional gateways compute asymmetric bid/ask collars off live multicast NBBO.

核心提要:Static "Previous Close +/- X%" collars fail on earnings gaps while still permitting catastrophic 500-bps order-book sweeps during calm markets.

Marcus Thorne, CQF
Mifid Rts6 Pretrade Kill Switch Controls

Gross vs Net Intraday Pre-Trade Credit Accounting: Open-Order Exposure & Cancel-Replace Inflation

Counting only filled trades toward intraday credit limits invites clearing insolvency; counting naive sums of bids and offers starves market makers. Learn exact worst-case directional and in-flight 35=G credit accounting.

核心提要:Per-symbol worst-case directional exposure MAX(|Pos + OpenBuys|, |Pos - OpenSells|) captures exact physical sweep risk without double-penalizing two-sided market makers.

Marcus Thorne, CQF
Mifid Rts6 Pretrade Kill Switch Controls

Exchange Cancel-On-Disconnect (COD): Heartbeat Grace Windows Across CME, Nasdaq & Eurex

Cancel-On-Disconnect is the market maker safety net when switches crash or fiber snaps—yet silent black-hole timeouts, 5-second grace windows, and GTD/GTC order exclusions routinely leave "protected" books exposed.

核心提要:Passive Cancel-On-Disconnect fires quickly (~150 us) only when a live OS kernel sends TCP FIN/RST; silent power or switch black holes wait for full heartbeat expiration.

Marcus Thorne, CQF
Mifid Rts6 Pretrade Kill Switch Controls

FIX Tag 35=q OrderMassActionRequest & Venue Firm-Level Kill-Switch APIs Benchmarked

Iterating through 4,000 individual OrderCancelRequest (35=F) messages trips exchange message throttles right when seconds matter; FIX 35=q and SBE MassAction529 purge entire books atomically in under 1.5 milliseconds.

核心提要:Never rely on O(N) individual OrderCancelRequest (35=F) loops for emergency halts; venue message throttles delay full book cleanup for multiple seconds.

Marcus Thorne, CQF
Mifid Rts6 Pretrade Kill Switch Controls

Self-Match Prevention (SMP): Cancel-Oldest, Cancel-Newest & Decrement-and-Cancel Mechanics

When Strategy A buys while Strategy B rests an offer under the same firm tax ID, unintentional wash trades violate FINRA Rule 5210 and CFTC Rule 1.38—making exchange-native SMP mode selection critical.

核心提要:Wash trades between independently run pods sharing one legal entity violate FINRA Rule 5210, CME Rule 534, and EU MAR unless prevented by exchange-native SMP.

Marcus Thorne, CQF

機構來源紀錄

機構來源紀錄

瀏覽匯入的機構、監管機構及市場紀錄。紀錄可能不完整或已過時,請向原始來源核實目前資訊。

文章與來源記錄 | Terminal Ledger Desk | Terminal Ledger Desk