Editor’s noteWe’ve renamed our product suite. The VMR is now referred to as Solace PubSub+. See our products page for more information.

***

Most recent performance numbers are available here.

The recent release of the Solace Virtual Message Router (VMR) Enterprise Edition dramatically expands the applicability of Solace message routers into a wide range of new application use cases. It does so by enabling the deployment of Solace routing and persistence functionality into a variety of deployment environments including corporate datacenters, public and private clouds, remote field offices, and Internet of Things environments. You can learn all about Solace message router technology and the VMR in particular on our technology page.

Since the announcement of the Solace VMR, people have been asking about its performance, particularly as it compares to our message router appliance. This blog post summarizes results which are now available in the performance section of our developer portal as an article that details the throughput performance of both the VMR and the 3560 appliance running release 7.1.1 of SolOS. There are so many dimensions to performance testing that you can never cover them all in one short article, so I focused on the Solace message routers and aims to help you understand their throughput and fan-out performance. We can look at other aspects of messaging system performance in other posts.

This is the first GA release of SolOS to include support for the Solace VMR. SolOS is the messaging technology that powers Solace Message Router Appliances and now also powers the Solace VMR, so the Solace VMR’s performance will improve dramatically over time as the Solace R&D team optimizes SolOS for the x86 hardware architecture. That said, we believe you’ll be impressed with the performance of this initial release. High performance is in our DNA at Solace, and you can rest assured it will be the same with the VMR as it evolves as well.

Summary of Results

(See recently updated results here.)

   Virtual Message Router 3560 Message Router Appliance
Messages Per Second
Non-persistent (Direct) point-to-point 260,000 7,900,000
Non-persistent (Direct) fan-out 2,900,000 26,400,000
Persistent message point-to-point 40,000 450,000
Persistent message fan-out  200,000 2,500,000

These results show that the Solace VMR offers very fast, scalable performance. In general, it behaves a lot like the 3560 appliances, just at lower message rates. That isn’t surprising since the VMR, which runs on commodity hardware, is based on the same underlying software, SolOS, that runs our appliances. Solace Message Router Appliances still offer the highest possible throughput and performance with our purpose-built hardware, so if you want to support massive messaging capacity or scale across many applications in a compact footprint, this hardware is for you.  If you need less performance, want to deploy messaging into small offices or IoT, or scale your system horizontally in a private or public cloud, then the VMR is for you. The beauty is you can mix and match or switch from one to another without any impact on your applications or management infrastructure.

Check out the full article for detailed results, descriptions of test methodology, full test parameters.  As new versions of SolOS are released the full article will be updated with new results so it can serve as a reference for Solace message router performance going forward.

Mark Spielman

Mark joined Solace in 2004 and currently works as a Senior Product Owner responsible for technical integration of partner’s products with PubSub+ Cloud. Prior to the his current position Mark held various other roles within Solace including Development Lead in the Professional Services team where he helped customers architect their applications and integrate the Solace products into their existing environments and as both a QA Project Lead and Software Designer in the Quality Assurance group helping guide feature development and ensuring new features surpassed customer’s expectations. Mark holds a B.A.Sc. in Computer Engineering from the University of Waterloo.