MuleSource subscribers already get MuleHQ, a set of monitoring and management tools, but the firm is gearing up to launch a more sophisticated and broader set of management and deployment tools in the fourth quarter, though details are at this stage still sketchy.
MuleSource CTO and co-founder Ross Mason explained the firm's business model in a Computer Business Review interview. Mason first founded the Mule open source ESB project in 2003 and then joined Dave Rosenberg, now MuleSource CEO, to found MuleSource Inc in 2006.
The business model is selling services and support, said Mason, We know there are over 1,000 customers using Mule in production and not all of them are buying support. We can offer them value-added services, as well as indemnities [against hypothetical legal action for copyright or patent infringement] and warranties.
But Mason added that the firm will also make money from, Management tools that are not open source. We'll do commercial software later on that helps people monitor and manage their Mule implementations even more effectively.
MuleSource is well funded, especially when you consider that today it has less than 50 staff. The firm took on a second round of funding totaling $12.5m in May this year, adding that to the $4m it raised in October 2006. The first round was from Hummer Winblad Venture Partners and Morgenthaler Ventures, while the Series B was led by Lightspeed Venture Partners.
MuleSource CEO Dave Rosenberg told us yesterday that the firm raised the additional funding, Because we could. He said there was enormous interest in the firm as it had large customers already using the Mule ESB, and it could demonstrate that customers were willing to pay for services and support.
So why did the firm take on a second round of funding so soon after the first? I've always wanted a yacht, joked Mason. Actually we didn't need the second round now but there was this huge buzz around the company after the first round: we were getting big customers and lots of traction, and anyone close to the market can see there is a huge opportunity there.
Rosenberg added that, We think there have been over a million downloads of Mule, but clearly the question is not just about downloads but about dollars - who is willing to pay. But the sizes of the deployments are becoming enormous: we now have over 10 people using over 200 servers, and H&R Block [a US-based tax and financial software and services outsourcer] will have Mule on 13,000 servers next year.
There are over 1,000 companies using Mule in production and we also think that with what's gone on in SOA [services-oriented architecture], it's increasingly clear that the ESB is key, said Rosenberg.
Mason said that the firm will use the funds to increase its headcount a little in sales and support, invest in version 2.0 of the Mule ESB which is due out soon, and develop its commercial management tools further.
But isn't an open source ESB like Mule largely developed by the open source community rather than by MuleSource's development team? Actually no, it's a misnomer that the community will write you a product, says Mason. They will validate what you do, certainly, and may write their own extensions, but you need to write the core product and do all of the necessary QA testing and so on to make sure it works with other things.
Mason said MuleSource can already offer 24x7 support to Mule customers, with offices in San Francisco, Atlanta, the UK, Malta, as well as offices just opened in Sydney Australia and Kuala Lumpur, Malaysia.
Only founded in 2006, it's clearly early days for MuleSource. One should not forget that there are plenty of open source ESB rivals such as ServiceMix, Fuse ESB (formerly Celtix), JBoss ESB and the Sun-backed Open ESB to name a few. Equally, commercial ESB vendors are rivals too because although downloading Mule is free, most customers will pay at some stage for services, support and warranties if they start scaling it up in the enterprise, just as they pay for commercial ESBs.
But if the company's claim that there are already 1,000 customers using Mule in production - many with over 100 servers - are to be believed, its prospects of fashioning a business from open source roots look more credible. It's clearly also well funded and that should help it get its infrastructure in place.
Enterprise middleware vendors like IBM, Oracle and BEA would argue that an ESB is only one element in an SOA suite, and for that reason MuleSource only solves a small part of the SOA problem. That is surely true, but it is also true that not every customer, and certainly not every project, is a full-blown SOA implementation. For standards-based, open-source, lightweight heterogeneous integration challenges, Mule is clearly answering many companies' needs already. Now all MuleSource needs to do is turn more of them into paying customers.