And we are hosting a messaging and integration conference in DC in May 2011. Its titled CamelOne. You can read more here http://camelone.com. And notice some of the speakers who are from large companies using Fuse ESB in production. And they have much love so they want to come to this conference and give presentations.
Let me know if you are in need of more information.
Sonic ESB, and all the other commercial ESB's, are roughly equivalent from a technical capabilities perspective. There are some slight differences that sometimes are relevant based on the specifics of your use case, but generally decisions are based more on comfort with a vendor's stack and largely on do you already own.
As Claus said, Fuse ESB / Apache ServiceMix's Apache open source license tends to be the number one deciding factor for most of the enterprises I talk to: initial up front cost savings by avoiding the software license purchase, and eventually the flexibility and sense of ownership your development gains though deep access to the code and community.
To get a little more technical, ServiceMix (Fuse ESB) includes Apache CXF and Camel for its Web Services and Integration frameworks respectively. Many other open source and commercial products (like Sonic) have adopted CXF and Camel as well for their ESBs - a case and point for not much capability difference if they use the same frameworks... ServiceMix has much more aggressively adopted OSGi versus the other ESB products - though they all have an OSGi roadmap - so if full use of OSGi is a plus for you, then that would be a plus for ServiceMix...
Don't get me wrong in that I'm not implying that ServiceMix (or Sonic ESB) don't solve important and difficult problems; FuseSource supports a number of large and small companies using ServiceMix in support on many big revenue generating applications and services. Its just that there aren't obvious capability differences between the products, so factors like the license matter more. That's why we think the open source model works well for this space as it lets folks like yourself try first and see if it really does solve a meaningful problem for your enterprise, which might even mean run in production for a while.
Now of course our biggest differentiator is Claus, but that's a different discussion