OB Group

Operated Brokerage Infrastructure

We build and operate the infrastructure behind trading businesses.

Infrastructure, engineered. Platforms, administered.

Intelligence supervises. The desk decides. Execution delivers.

You build the business — we make sure the technology is never the bottleneck.

The platform we operate

Private cloud and servers, the MT5 and MT4 platforms themselves, databases, monitoring and security — built, secured and run around the clock. The production stack and the daily operations, off your plate.

01Private cloudKVM, Terraform, Ansible, Equinix
02MT5 / MT4 administrationGroups, symbols, swaps, Manager API
03DatabasesMySQL, MariaDB, PostgreSQL, ClickHouse
04Monitoring & securityGrafana, Prometheus, Vault, firewalls
0524/7 operationsSLAs, incident response, change management
ProvisionMonitorRespondHardenProvision

A host drops, a disk fills, a service degrades — monitoring catches it, the on-call team responds, and the fix ships fleet-wide as code. The platform is provisioned, monitored and hardened continuously — the broker runs on infrastructure that is operated, not a one-time setup.

We build and operate the platform itself.

We run dedicated, low-latency infrastructure, install and administer the trading platforms, and operate the whole production stack around the clock — as one accountable operator running it as one engineered system.

Colocation & low-latency infrastructure

Dedicated, low-latency environments at Equinix and other major financial data centres — colocation topology, redundant network structure and disaster-recovery governance, built for institutional brokerage operations.

Private cloud, provisioned as code

Bare-metal and virtualized deployments stood up with Terraform and Ansible. Every host is reproducible from one command, not configured by hand.

MT5 & MT4, administered

MT5 and MT4 operated across the full server stack, from deployment and tuning to live migration. Trading conditions are configured end to end: symbol sets, swap and commission schemas, groups and sessions. Manager-API automation handles accounts, credits, routing and reconciliation.

High availability & disaster recovery

Hot-standby replication and failover automation across the trading core — the MT5 trade and history servers and our own FIX stack both run redundant — with backups and recovery procedures.

Databases, streaming & observability

MySQL, ClickHouse and PostgreSQL operated, replicated and backed up. Live MT5 deal and position data streams over Kafka, and a Grafana, Prometheus, Loki and ELK stack carries metrics, log aggregation and alerting across the whole fleet.

Continuous operations & security

24/7 monitoring, defined SLAs, incident response and change management under one permanent contract — with defense-in-depth at the edge, host and data layers (firewalls, VPN, SSH hardening, Vault-managed secrets, TLS/PKI) and a full change and access trail.

What the broker gets

The broker stands on an institutional-grade platform without building, staffing or coordinating it. One team builds it, secures it, runs it 24/7, and is accountable when it matters.

  • Equinix colocation
  • Terraform + Ansible IaC
  • MT5 · MT4 administration
  • HA + failover automation
  • ClickHouse · Postgres · Kafka
  • Grafana · Prometheus · Loki · ELK

Three Proprietary Instruments

On the platform we run three instruments of our own — operated, never sold. They make the operation supervise, decide and execute as one self-correcting loop.

MonitorDecideAdjustVerifyMonitor

A trader the platform flags as toxic is re-routed by the desk. A risky group has its markup or leverage adjusted. The change is verified — then the cycle repeats. The broker gets a system that corrects itself, not five separate tools.

The Trading Intelligence Platform supervises.

Every deal, scored. You see who is profitable at your expense, why, and what to do about it — in minutes, not days.

It answers three questions
01Who is toxic?
02Why are we losing money?
03What should be routed differently?

Risk classification

A Profitability × Risk threat matrix — Critical, Elevated, Watch, Safe — recommends an action on every account: A-book, B-book or monitor. With trader drill-down, symbol risk and classification history.

Toxic-flow detection

Named toxic-flow patterns across profitability, consistency, duration and timing — each with distribution and per-trader detail.

Manipulation & fraud

Latency-arbitrage detection by tick-sync correlation, collusion across account pairs and clusters, account linking via shared CID and IP, and login forensics for geo-impossible and VPN/datacenter access.

P&L attribution

Company P&L with A/B-book split, hidden markup, slippage, commissions and spread cost — plus credit-abuse and stop-out-compensation detection with a configurable taxonomy.

Behavioral analytics

ML trading-style clustering, full trader profiles, retention and multi-account segments, registration-quality scoring and fast-growth detection.

Real-time operations

A live dealer board — execution-speed analysis, gap monitoring, request-flood surfacing and time-of-day analytics.

What the broker gets

The dealing desk and risk team stop flying blind. Who is toxic, who to A-book, who to restrict, and where the money actually goes — answered with evidence, and fed straight back into dealing.

  • ClickHouse analytics backbone
  • named toxic-flow pattern library
  • latency-arbitrage · collusion · account-linking
  • A/B-book P&L attribution
  • ML behavioral clustering
  • real-time dealer board

The Smart Dealing Desk decides.

Your own FIX gateway and dealing desk — not a rented bridge. We route, hedge and price on your behalf, with the logic owned and observable.

A-book / B-book routing

Deterministic per-session routing — each trading-server session maps to an explicit destination, decided per group, not per trade by luck. Move a group between A-book and B-book by live reload, with no session restart.

Direct market access

A-book orders go straight to the chosen LP as market or IOC, with platform ↔ venue symbol translation and a full order-state machine tracked by ClOrdID.

Internalization engine

A B-book engine per server, pricing off the aggregated NBBO, with a realistic per-asset-class execution model — wall-clock and volatility-aware holds plus size-based slippage tiers.

Markup management

Per-category B-book markup baked into the synthetic fill, and an A-book markup manager that re-applies the broker's spread on the fill leg — closing the leak a raw LP fill would otherwise create.

Exposure & hedging

Net-exposure tracking, hedging decisions, margin-call and stop-out thresholds, and spread, swap, news and holiday policy — operated by an in-house desk.

Operated as code

Routing, markup and slippage configs are version-controlled and hot-reloaded without dropping sessions, with per-category metrics on the same monitoring stack.

What the broker gets

The broker controls exactly how each group is executed and keeps the dealing margin a bridge vendor would take — logic owned and observable, not a black box.

  • deterministic per-session A/B routing
  • 3-stage slippage model per asset class
  • per-group A-book + B-book markup
  • full order-state machine
  • live config hot-reload
  • own gateway — not a resold bridge

The FIX Gateway executes.

We wire your trading servers to the outside world — data feeds, liquidity providers and exchanges — and aggregate it all into one clean price.

Own FIX gateway

A proprietary FIX 4.4 gateway built and operated by us — not a resold third-party bridge — carrying both market-data and execution sessions to each venue.

Direct venue connectivity

Direct sessions to exchanges, liquidity providers, banks and prime brokers across both traditional and digital-asset markets, added per engagement.

Quote aggregation

Cross-venue NBBO with stale-venue failover in seconds, per-venue contribution metrics and bad-tick rejection. If a feed dies, trading continues on the others.

Feed into the platform

A datafeed gateway pushes live aggregated prices straight into the trading platform over an auto-reconnecting FIX session, and the same feed serves every connected platform and the dealing engine.

Symbol & spec mapping

Each venue's symbols, digits and contract sizes are mapped to the platform's, translating platform ↔ venue conventions before go-live.

Secure transport

stunnel TLS termination for FIX-over-TLS venues, certificate management, and FIX conformance certification against each venue's spec.

What the broker gets

The trading servers are fed one clean, aggregated, multi-venue price with order pipes ready — no third-party aggregation, no intermediary latency, no dependency on an external bridge.

  • own FIX 4.4 gateway
  • direct exchange & LP sessions
  • cross-venue NBBO + failover
  • symbol & spec mapping
  • stunnel TLS + FIX certification
  • LP sourcing & credit setup

How We Operate

We do not deliver software and step away. We operate your stack permanently — and we are accountable for what it produces.

Ongoing Operations
  • 24/7 infrastructure monitoring and incident response
  • Execution and routing performance oversight
  • Exposure and risk position supervision
  • Toxic flow detection and client behavior monitoring
  • P&L analytics, fraud alerting and structured reporting
  • Liquidity quality and pricing review
Strategic Projects
  • Full infrastructure design or migration
  • Routing and execution architecture restructuring
  • FIX connectivity and exchange integration
  • Client base audit — toxic flow, fraud patterns and behavioral profiling
  • Custom ML models for broker-specific detection tasks
  • Risk framework design and stress testing
Engagement Structure

Engagements are scope-based and structured around operational complexity and scale. Initial reviews are complimentary.

Client Portfolio

We maintain a selective client portfolio to ensure dedicated oversight and long-term alignment.

Selective engagements. Long-term alignment.

Closed by Design

Everything we run was shaped while operating real brokerages, not in theory — so the broker stands on a stack where every part has already earned its place under real conditions.

The infrastructure is the foundation: what has to be in place, what fails under load, what keeps a brokerage live when it matters. We design it from that experience and operate it ourselves, end to end.

The instruments on top — execution, dealing and intelligence — work at the level of the individual account: many different clients, each routed and priced on its own merits. Detail like that isn’t bought off a shelf. It’s built by running the thing it serves. So we operate the instruments ourselves — never licensed, never handed over.

Request an Operational Review

An operator-grade review of your trading backend — execution, dealing and risk. Initial reviews are complimentary.