Solace in Capital Markets: Market Data, FX and Trading Messaging

Solace in capital markets is where the company began and where its broker was shaped. Solace was founded in Ottawa in 2001 to run messaging in hardware, and its first buyers were banks distributing market data and trading events to thousands of applications at low, predictable latency. Solace states that 6 of the 10 biggest investment banks and 3 of the top 6 FX trading banks use it (solace.com), and its largest published deployment is RBC Capital Markets. This page explains the problems banks and exchanges use it for, the published customers, and why the same patterns now matter to manufacturers.

It sits under Solace use cases by industry in the Solace cluster.

The problem capital markets brought to Solace

Market data is a fan-out problem with a deadline. One price update from a venue may be needed by an algorithmic trading engine, a pricing service, a risk engine, a surveillance system and hundreds of trader screens, and it loses value with every microsecond of delay. Volumes are spiky: a volatile morning can multiply message rates without warning.

Trading also has transactions that must never be lost. An order, an execution report or a trade confirmation is not superseded by the next one. It must arrive, in order, even if the consumer restarts.

Solace’s answer was one broker that carries both. Direct messaging delivers prices at-most-once with minimal latency. Guaranteed messaging spools orders and trades to queues until each consumer acknowledges them (Solace docs). Hierarchical topics with wildcards let each application subscribe to exactly the instruments it needs, which the topics and wildcards page explains.

Solace capital markets customers, as published

The table lists the banks and exchanges with public Solace case studies. Every description is Solace’s, as of October 2026.

Customer	What Solace states it does there	Pattern
RBC Capital Markets	Backbone for the global low-latency trading platform and market data. Grew from 65 billion to over 118 billion messages a day; about 30,000 application connections; 400+ application teams. Replaced incumbent middleware around 2010	Fan-out, legacy replaced
Barclays	Integrates applications across front, middle and back office on Solace appliances	Enterprise integration
Standard Chartered	FX trading platform. Started in four trading centres (Singapore, Kuala Lumpur, London, New York) and expanded to 11 sites	Many sites linked
LSE Group	Persistent “shock absorber” between trading systems and downstream surveillance, data warehousing, post-trade and billing	Guaranteed buffering
TMX Group	Messaging backbone for the plant that disseminates Toronto Stock Exchange and TSX Venture data feeds	Fan-out to participants
Akuna Capital	Real-time data distribution for a Chicago options market maker, replacing homegrown middleware	Fan-out, legacy replaced
TAB Bank	Event-driven endpoints for fintech partners, complementing a MuleSoft ESB	Events alongside APIs

Direct or guaranteed: how trading flows are split

Flow	Typical delivery mode	Why
Prices and quotes	Direct	The next update supersedes the last; a retransmit is stale and adds latency
Reference data (instruments, calendars)	Guaranteed or cached	Every consumer must hold the same version
Orders and execution reports	Guaranteed	Each message is a commitment that must arrive, in order
Trade confirmations and post-trade	Guaranteed, often partitioned queues	Must not be lost; ordering per account or instrument, with consumers scaled out
Surveillance capture	Guaranteed queue fed by wildcards	Regulators expect a complete record

Ordering with scale-out is covered on the partitioned and last value queues page.

Late joiners are the subtle case. A trader who opens a screen at 10:00 needs the current price of every instrument on it immediately, not whenever each one next ticks. Direct messaging does not keep history, so the design adds a last-value store: Solace Cache for direct topics, or a last value queue that holds only the most recent message. The same need appears in a plant, where a new dashboard or analytics service must start from current machine state rather than wait for the next change.

Beyond the trading floor: banking and insurance

Solace’s financial customers are not only investment banks. TAB Bank, a US bank serving business customers and fintech partners, added event-driven endpoints beside its MuleSoft ESB so partners could react to events instead of polling APIs. Indiana Farm Bureau Insurance announced a hybrid cloud event mesh in October 2025, and Solace’s company page also lists HDFC Bank. The driver here is different from trading. Latency in microseconds matters less than decoupling: letting a core banking or policy system publish once while digital channels, partners and analytics consume at their own pace.

Many trading centres, one routing fabric

Global banks trade from several cities, and each needs local prices fast and remote prices when someone subscribes. Standard Chartered’s growth from four trading centres to eleven sites is the published example. Brokers in each centre are linked by dynamic message routing, which propagates subscriptions so a topic crosses the WAN only when a consumer on the far side asks for it.

This is the pattern that transfers to manufacturing. Swap trading centres for plants and prices for production events, and the design is the multi-site backbone described in Solace in manufacturing: a broker per site for local speed and WAN tolerance, linked for enterprise visibility.

Where it gets hard

• Market data licensing is not a broker feature. Exchanges license data per user and per use. Broker ACLs control which client may subscribe to which topic, but entitlement management and usage reporting to exchanges are usually separate systems. Ask how the two are kept consistent.
• Latency figures are vendor claims. Solace’s appliance page cites about 20 microseconds for direct messaging and up to 12 million messages a second point to point, without naming the model. Software brokers on shared infrastructure will not match it. Test with your own message sizes; the software vs appliance page covers the trade-off.
• Distribution costs more than the data. Solace, citing a 2019 Refinitiv report, states that making market data available across an enterprise can cost up to eight times as much as acquiring it (Solace). A broker reduces duplication, but the licences and the people remain.
• Incumbent replacement is slow. Banks run TIBCO Rendezvous, FTL and EMS, and IBM MQ, often for decades. Migrations need dual running and flow-by-flow testing. See Solace vs TIBCO and Solace vs IBM MQ.
• Slow consumers hurt everyone if unmanaged. A trader screen on a poor connection or an overloaded risk engine must be isolated so it cannot back up the feed for others. Ask how the design discards or buffers for each class of consumer.

The sibling page on retail, airlines and logistics shows the same broker applied to stores, gates and ports. If you are weighing these patterns for your own estate, a modernization roadmap sets the order of work before any product choice.