Friday, April 4, 2008

MuleCon 2008 Roundup

We held our second MuleCon in San Francisco and what an event! We had around 250 attendees from all over the globe with Norway, Turkey, Japan, UK, Argentina and Sweden, Kula Lumpur and Australia represented (that I can think of). Like last year we had customers presenting their experiences with Mule and MuleSource, technology partners talking about complimentary products and of course the MuleSource team talking out our products, roadmap, deep dive sessions and lots of Q and A. It was great to see the diversity of users and customers in terms of size of customer and role in the organization. It was a real mix bag of people and I think it worked really well; people really seemed to enjoy exchanging use cases, success stories and challenges.
Though I didn’t have time to attend many sessions, I did catch glimpses of a few. Some that stuck in my mind were –

Scripps Network (the guys behind HGTV, FoodTV etc) gave an good case study presentation about how they have used Mule to help them build Rich client applications. What was really interesting is seeing the progress Scripps have made since they came to MuleCon last year and we held and “Open Architecture” session using them as a use case. By the way their use of Mule is a little different from most since they are actually scheduling programming content including HD, so they really need a platform that can route very large messages in a reliable way.

Eugene from LeapFrog gave us a good insight into how to architect Resource Oriented Architectures (ROA) using Mule and demonstrated how this is being used at Leapfrog. I found this session particularly interesting since it sparked some ideas of how we could combine the recently release Mule RESTpack with the patterns Eugene used to help Mule users build ROA.

Jahan More of U1 Technologies gave a good presentation on high performance messaging. U1 is a little known JMS vendor right now that have a very fully featured JMS implementation that goes well beyond the JMS spec (in a good way). I was pretty impressed with what their Ambrosia JMS server can do, and we are currently in the process of certifying it with our QA infrastructure for use with Mule. Those who know me know I am a big fan of AMQP and see it making huge in-roads at all levels of the organisation in future, but I still see a lot of value of having a slid JMS implementation available for people today.

Dan Diephouse gave a great drill down of Mule Galaxy, our Registry and runtime governance product. We saw a lot of interest in this primarily because as the SOA projects in organisations mature more people are faced with the issue of how do you manage, discover and govern the increasing number of services available. Dan also spoke about our new RESTpack and new Web Services support in Mule 2.0.

We ran two panels and both provided lively discussion. The first day panel “Are they Smarter than a 5th Grader” discussed for the most part the impact and adoption of Open source in the Enterprise. The session was hosted by Michael Cote, from Redmonk with Larry Augustin,, from Augustin ventures, Matt Asay from Alfresco, Jason Maynard from Credit Suisse and Dave Rosenberg. M. di Paolantonio provides a good walk though.

The second panel was hosted by Michael again with John Davies, John Davies, Eugene Ciurana, John Rowell and John Gardner. Unfortunately I missed this one, but M. di Paolantonio covered it.

We have a number of excellent technology partners presenting. Uri Cohen from GigaSpaces gave a good overview of the different strategies you can combine Mule and Gigaspaces to build highly available and scale out solutions with Mule and Gigspaces. Alexis Richardson for CohesiveFT (who I’ve blogged before) gave an excellent demonstration of using Mule and RabbitMQ deployed to Amazon EC2 using CohesiveFT’s Elastic Server On Demand (ESOD). Why would you do this? Well, you get a great scaling story when you deploy Mule and RabbitMQ together on EC2 clouds and ESOD makes this very easy; a step in the SaaS direction. John Davies talked about high-performance message processing using Artix Data Services, which seemed to be well received by the slice of our audience from the financial sector.

For those who missed it Mule 2.0 was released during MuleCon. The Mule 2.0 team did a great job of providing a drill down session of what is new in Mule 2.0. There was a lot of buzz from customers and users wanting to get going with Mule 2.0. The last session of the day we decided to combine my session with the “Developer campground”. We kicked off with a demonstration of Mule hot deployment with some oohs and ahhs from the crowd. People were very keen to see when they could get their hands on this stuff. We need to do a lot more QA and tooling around it before its ready for our customers but I imagine we might provide an early access release later this year.

Other members of our team presented the new Mule IDE that works with Mule 2.0. I heard lots of great comments about the progress we've made. Also, our business transaction monitoring product, Mule Saturn, got an introduction. This product is in beta at the moment but seemed to be something our customers really want.

I was overwhelmed by the response of this year's event. If you missed it this year, make sure you don't next year. We will be posting all of this year's content on our website.

p.s. sorry if I missed anyone


Yuen-Chi Lian said...

Thanks a lot for the write-up.

I was reading up the SOA book by Thomas Erl last week, which mentions about service registry being one of the essential components of an SOA architecture.

With Galaxy, we are seeing a more complete picture now.

- yc

Ross Mason said...

Yuen-chi, That is a pretty good book. I don't think I've seen such an exhaustive coverage of SOA. What the book lacks in fun it more than makes up for in detail :)

Travis Carlson said...

You forgot Germany and Israel. This was a real global convergence!