Solace Standard Edition vs Enterprise: Limits and Production Use

The Solace Standard Edition is the free version of Solace Event Broker Software (formerly PubSub+ Standard Edition). It has the full feature set but is capped at 1,000 client connections, a 10,000 messages per second rate limit and one Message VPN, with no disaster recovery replication. The Enterprise Edition lifts those caps to 200,000 connections, unthrottled rates and five VPNs, and adds replication. The question most people search for is whether Standard can run in production, and the two sources disagree: Solace’s 2018 launch called it “production-ready”, while the current documentation marks it not production deployable.

This page gives the full limits table, shows both sources side by side, and explains where Standard is genuinely useful. It belongs to the Solace Event Broker overview in the Solace cluster.

Solace Standard Edition vs Enterprise: the limits

This is the comparison table from the Solace documentation, as of October 2026.

Feature	Standard Edition	Enterprise Edition
High availability with redundancy	Yes	Yes
Disaster recovery with replication	No	Yes
Dynamic Message Routing	Yes (the table adds “100+ connections”)	Yes
Message replay	Yes	Yes
Microgateways	Yes	Yes
Upgradable	Yes	Yes
User-configured Message VPNs	1	5
Direct and guaranteed messaging rate	10,000 messages/s limit	Unthrottled
Client connections	Up to 1,000	Up to 200,000
Maximum message spool	800 GB	6 TB
Maximum queued messages	240 million	3 billion
Production deployable	No	Yes
Price	Free	Not published
Support	Not offered	Commercial support subscription

The rate limit is the cap most teams hit first. The table applies it to both direct and guaranteed messaging. We read it as a limit on messages published into the broker, before fan-out; confirm that with Solace for your design. A plant that publishes tag changes by exception from a few lines will fit comfortably; a site that streams high-frequency vibration or vision data through the broker will not.

The “production deployable” question, with both sources

This is the contradiction worth understanding before anyone builds on Standard.

Source	Date	What it says
Solace launch release	24 May 2018	Headline: “Solace Launches Free Production-Ready Version of its Industry-Leading Message Broker”. Denis King, then COO, called it “the exact same product as our Enterprise edition”, limited to “a reduced number of connections” and rate-limited to 10,000 messages per second
Solace docs, edition table	Current, October 2026	“Production Deployable”: No for Standard, Yes for Enterprise
Solace software page	Current, October 2026	“Standard edition is not to be used in production environments”, with no production support, and Enterprise described as “always production-deployable”

Our reading. The product did not get worse. What changed is how Solace positions the free edition: in 2018 it was a route into production for small workloads, and today it is presented as an evaluation and development edition. The current Solace pages are the ones that bind a buyer today, and they say no. Whether the licence text itself prohibits production use, or only withholds support, we have not verified. Read the licence that ships with the version you install, and ask Solace for a written answer if a production use of Standard is being considered.

The practical answer is the same either way. A broker that carries production orders, quality holds or maintenance work orders needs a support contract and, usually, disaster recovery. Standard has neither. That alone rules it out for anything a plant would miss on a Monday morning.

What the Solace Standard Edition is good for

• Learning and prototyping. It runs in Docker on an engineer’s laptop with the full feature set: topics, queues, replay, MQTT, JMS and REST.
• Development and CI environments. Teams can run the real broker in pipelines rather than a mock.
• Lab and pilot work that stays under 1,000 connections and 10,000 messages per second, and that nobody depends on.
• Testing HA designs. Because HA is in Standard, a team can build and fail over a three-node group before paying for anything. The HA and disaster recovery page explains what to test.

Sizing a pilot against the Standard limits

A pilot that outgrows Standard halfway through loses weeks. A quick sizing pass before installing avoids that. An illustrative pilot: one packaging hall with six lines, each with a gateway publishing about 400 tags by exception, plus an MES, a dashboard server and a quality application.

1. Connections. Six gateways, three applications with a few sessions each, a handful of engineers’ test clients and one bridge to a second broker: well under 100. Standard is comfortable.
2. Message rate. If each tag changes on average every two seconds, six lines produce about 1,200 messages per second. Fan-out to three consumers is delivered by the broker, but the published ingress stays near 1,200. Comfortable, with headroom for bursts at shift change.
3. Spool. If the quality application is offline for a weekend and its queue collects 50 messages per second of 1 KB each, that is roughly 8.6 GB over 48 hours. Far inside 800 GB.
4. Tenancy. One Message VPN is fine for a pilot, but note now whether production will need OT and IT separated.
5. Recovery. If the pilot must survive a site loss, it has already left Standard.

Now change one assumption: the line team wants vibration data at 100 samples per second from 40 motors. That alone is 4,000 messages per second per line, and the 10,000 cap is gone on the third line. The rate limit, not connections, is usually what ends a Standard pilot in manufacturing. Treat high-frequency data as a separate design question, and consider batching it at the edge before it reaches the broker.

The 90-day Enterprise evaluation

To test Enterprise limits, install Standard and enter a free 90-day evaluation key. The documentation is blunt about the end: upgrade to a purchased Enterprise licence before day 90, or “the event broker will cease to function after 90 days, and all configuration will be lost”. Export your configuration, or keep it as code with the Solace Terraform provider, from day one.

Moving from prototype to production

When a pilot succeeds, the choice is between Enterprise software, which you run, and a Solace Cloud service, which Solace runs. Cloud service classes run from 250 to 100,000 connections, and 200,000 on request, as of October 2026 (Solace pricing). Neither price is published. The editions and pricing page covers what a quote is built from, and a modernization roadmap is where we would place the broker decision alongside the rest of the integration plan.

Where it gets hard

• Counting connections honestly. Every MQTT client, JMS session pool, REST consumer and bridge counts. A plant with a few hundred devices plus applications can pass 1,000 sooner than expected.
• One Message VPN means no tenancy. Standard cannot separate OT and IT traffic into different VPNs, which matters for the Message VPN security model.
• No disaster recovery. Without replication, a site loss means rebuilding the broker and losing whatever was spooled.
• The routing qualifier. The edition table lists Dynamic Message Routing for Standard with the note “100+ connections”. Its exact meaning is not explained in the table; ask before designing a mesh around Standard nodes.
• The licence question is unresolved in public. See above.

---

Related: Solace Event Broker · Software vs Appliance · Kubernetes and OpenShift · Solace hub