Colocation for B3: technical requirements and checklist to integrate infrastructure into the market environment

Orders that arrive late at the B3 matching engine or are rejected during market peaks are not merely technical incidents: they create missed opportunities, execution risk and commercial disputes. In many cases the cause is unpredictable latency, high jitter or network paths that traverse shared transit when dedicated routes were needed.

This guide provides a practical technical and contractual checklist that trading, infrastructure and compliance teams should require from a colocation before migration: how to measure, validate and negotiate connectivity focused on B3, which topologies favor execution consistency, and which SLA clauses protect your operation.

What operational risk should you mitigate first?

Picture a volume spike where algorithmic strategies need predictable latency to maintain spreads and avoid slippage. If a link presents high jitter or alternate routes cross congested providers, the outcome is delayed orders or retransmissions that produce losses. In a colocation environment, risk increases if cross‑connects are absent, multiple carriers are not available, or routes to B3 are not validated in pre‑production.

Which metrics matter — and how to turn them into requirements?

For connecting trading infrastructure to B3, relevant metrics are not just averages: percentiles and consistency are key.

• Round‑trip time (RTT): measure percentiles (e.g., p95/p99) to understand variation under stress. Require reporting over market windows.
• Jitter: packet delay variation; affects buffers and retransmission behavior in low‑latency protocols.
• Packet loss: even low loss rates can trigger retransmissions that significantly change arrival times of orders.
• Consistency: route stability during critical periods; frequent route changes indicate shared transit or unstable BGP policies.

Avoid arbitrary numeric thresholds without baseline: derive targets by comparing reference measurements from your production site to B3 points of presence, collecting percentiles and market time windows to set acceptable limits.

Recommended connectivity topology for trading with B3

There are clear trade‑offs between cost and predictability. In order of operational preference for order execution:

1. Direct cross‑connect to the B3 interconnection equipment within the same data center (minimizes route variability).
2. Direct peering with providers that host gateways to B3 (when physical cross‑connect is not feasible).
3. Multicarrier with physically diverse paths; deterministic failover policies.
4. Shared IP transit only as a last resort, with strict contractual quality limits.

Each option carries operational implications: cross‑connects require physical coordination and maintenance processes; peering demands BGP configuration and filters; multicarrier requires SLA management and failover testing.

Minimum technical checklist for a colocation RFP focused on B3

Use this checklist as a baseline for the RFP. It focuses on elements that impact latency, predictability and availability for trading connectivity.

• Description of physical ports available (capacity per port, media type, latency between racks and interconnection points).
• Cross‑connect options: physical presence of patch panels for direct connections to B3 or to providers that interconnect it.
• List of carriers and the ability to implement multicarrier with independent paths; carrier failover policies.
• Peering policy and the possibility to establish direct BGP sessions with relevant providers.
• Details on existing SLAs: availability, cross‑connect restoration times, maintenance windows and notification lead time.
• Change and test procedures for maintenance: windows, rollback and notification.
• Capability to perform latency tests and collect metrics pre‑contract (p95/p99, jitter and loss) and support tests during market hours.
• Physical security and access control at rack/cage level; compliance requirements for trading operations.
• Redundancy plans for power and cooling, including whether redundant network paths are physically segregated.

Essential contractual clauses for trading operations

Beyond availability, negotiate clauses that cover network quality and interconnection operations:

• Cross‑connect availability SLA: maximum downtime for physical connections and prioritization procedures in case of failure.
• Network performance SLA: delivery of regular reports on latency, jitter and loss; agreements on measured percentiles (p95/p99) and remedies when applicable.
• Maintenance windows and procedures: minimum notices, options for out‑of‑market maintenance and rollback agreements.
• Support for testing and pre‑production access: permission to run scripts, capture traffic and validate routes with B3 before migration.
• Guarantees on logical or physical segregation: how redundant paths will be handled to avoid single points of physical failure.

How to validate routes and performance in pre‑production

Effective validation requires tests in representative market conditions:

• Test scripts: application‑layer and ICMP/TCP latency tools to measure RTT, jitter and loss; run tests during simulated peak windows.
• Traffic capture and BGP path analysis: validate AS‑path and identify if any hop traverses high‑variability shared transit.
• Failover tests: simulate loss of a carrier or cross‑connect and measure failover time and consistency.
• Percentile measurements: collect p50/p95/p99 in windows of 5 to 30 minutes to capture variation.

Document everything in a technical report to be annexed to the contract, detailing validated routes, test times and collected metrics.

Trade‑offs and limits: when to choose cost versus predictability

Cost reduction often means accepting shared transit or less physical redundancy — which impacts execution predictability. Conversely, investing in cross‑connects and multicarrier reduces variability but requires contractual and operational coordination. The decision should be driven by the risk profile of the trading strategy: latency‑sensitive strategies typically justify higher infrastructure investment.

What a colocation provides and what it does not replace

A colocation supplies space, power, connectivity and interconnection options. It does not replace trading governance, algorithm design, application monitoring or internal failover policies. Successful integration requires cooperation between your network, ops and the colocation contracting teams.

Operational FAQ

What is a cross‑connect to B3 and why is it important?

A cross‑connect is a physical link between your equipment and the interconnection panel that provides access to B3 or to providers that interconnect it. It reduces hops and route variability, increasing predictability of order delivery.

What is the difference between direct peering and IP transit for market connectivity?

Direct peering exchanges traffic between networks without transit providers, reducing hops and often improving consistency. IP transit routes traffic through providers, which can introduce variability from congestion outside your control.

Does MVX provide latency tests before contracting?

You can request a diagnostic and proof of concept to validate routes and behavior in pre‑production. To start a consultative diagnostic, contact a specialist at fale com um especialista.

What are the risks of using shared links for production orders?

Shared links may suffer variability from third‑party traffic, causing jitter, loss and route changes that affect market orders. Consider dedicated links or multicarrier for critical strategies.

Operational next step with MVX

If your immediate concern is to reduce variability in orders sent to B3, request a connectivity diagnostic and proof of concept. MVX can help map routes, validate cross‑connects and prepare a technical annex for your RFP. To begin, fale com um especialista and request the checklist tailored to your trading environment.

Quick decision checklist

• Have you measured p95/p99 between your site and B3 during market hours?
• Is a physical cross‑connect to the interconnection point available?
• Are there physically independent multicarrier options?
• Does the contract include network performance SLAs and adequate maintenance procedures?
• Did you validate failover and BGP routes in pre‑production?

Answering these questions reduces the risk of operational surprises and provides a solid basis for technical and contractual negotiation with the colocation provider.