Building a stronger open source product

It has been fascinating to watch the Alfresco community grow over the last few years. We really had no idea what the shape would be when we started and who would adopt our product and our project. Since 2005, we have been joined by some great world-class companies, household brands and some of the best names in small and medium size businesses. We are very pleased with the adoption of both the open source and enterprise products. What we are puzzled about though is that some of the biggest enterprises in the world (and I mean Fortune 50 and even Fortune 10) are only using the open source version of the product. We have designed our enterprise services specifically to cater to this type of customer.

Some open source systems try to prevent companies using their free version by either crippling their non-enterprise products or by letting their open source versions run into a destabilized state. Back in 2006, we came to the conclusion that we didn't want to hinder the open source version of the product. To hinder the product would make it difficult for certain governments to use our product and would encourage the community to build around imposed limitations. We don't want to provide you with a crippled version, because it doesn't really do the product justice when you try it. And we are trying to provide you with one of the biggest benefits of open source, eliminating lock-in to proprietary software. In fact, in the latest release of Alfresco, we went well out of our way to incorporate many of the changes that we have been putting into place for our enterprise customers into the Labs version of Alfresco.

MySQL has been experimenting with changes to their business model recently that attempts to draw a line on what is fair in open source. After all, it is the revenue that is generated from the enterprise subscription that helps fund, grow and improve the freely available open source version. When Matt, John and I met with Jonathan Schwartz last year, he said that he felt that it is important to have a completely open source core and system, but that in order to have the tools to run MySQL in production or in a high scale, high availability environment, it is only fair to have those tools be available as enterprise and for purchase. Interesting concept, but does it work?

We took some time earlier this year to consider what was fair and what are the core principles to which we want to adhere. We tried to determine where it is right to charge for a service or a function and where do we defend a capability as open source. If we are held to account, these are the principles that I expect we can apply with transparency, consistency and fairness:

• We must insure that customers using our enterprise version are not locked into that choice and that open source is available to them. To that end, the core system and interfaces will remain 100% open source.
• We will provide service and customer support that provides insurance that systems will run as expected and correct problems according our promised Service Level Agreement
• Enterprise customers will receive fixes as a priority, but that we will make these fixes available in the next labs release. Bugs fixed by the community are delivered to the community as a priority.
• We will provide extensions and integrations to proprietary systems to which customers are charged. It is fair for us to charge and include this in an enterprise release as well.
• Extensions and integrations to ubiquitous proprietary systems, such as Windows and Office, will be completely open source.
• Extensions that are useful to monitor or run a system in a scaled or production environment, such as system monitoring, administration and high availability, are fair to put into an enterprise release.

We started with Alfresco 3 to put extensions to proprietary databases such as Oracle or SQL Server into the enterprise release only, while extensions to MySQL, Ingres and other open source databases were available in open source. Now with the Alfresco Enterprise 3.1, we will be adding system monitoring capabilities and easy clustering administration that will only be available as part of the enterprise version. This does not prevent the open source version from being a very usable or even scalable system. However, we believe it provides an incentive for those large enterprises that have not chosen the enterprise system to do so, because it significantly reduces their costs of deployment and scalability, as well as providing them the help and support they need for deployment. These enterprise subscriptions help provide the resources that make a stronger and more functional open source release. The enterprise subscriptions also insure that production systems will be up and running. We still provide this enterprise system at a cost that is still less than 10% of proprietary systems.

We want to make both our enterprise and community users successful. The more people download, install and use the community version successfully, the more they will put it into production and look to an enterprise subscription as an insurance policy for that production system. The enterprise subscription is designed to save time and money and be more cost effective than supporting the labs product yourself. Alfresco is in a unique position to offer this. In all situations, we want our users to be able to choose the best option for them.

We are making these changes in a way that is based on a set of principles that are fair and accountable. We believe in open source and making it freely available and providing choice of not just proprietary systems, but between enterprise and open source. We think the rest of the open source world is heading in a similar direction, because this is what makes open source stronger in the long run. However, we are interested in what you think. Please drop me a line with your thoughts, ideas and concerns.

Comments

[Arabic Translator]

The basis for the Opensource phenomenon is that services can actually drive more revenue than code. If you are the defacto source for the code then you can control the knowledge associate with the most profitable services. This is a smart strategy.

However, as you point out there also needs to be a the intent to help the greater good. Cheap beer is good if it is brewed well and doesn't have you running to the bathroom every ten minutes.

[r4 sdhc]

Such a true clear and concise observation.Brilliant post about building a stronger open source product.

[Tim Neugebauer]

@Ronny:
I am with you in that point. It´s about service and opportunity costs. What resources can be saved by using Alfresco Enterprise Level software version and/or service level subscribtion in comparison with community version usage.

But it is also a question of assurance and concept. What can be developed in better quality and at better costs within "Alfresco Headquarter" and what not. I believe in that kind of "dual business model" with high quality open source version and enterprise version with enterprise features. If large enterprises do not use enterprise version there is something wrong with consultancy or communication... :-)

In addition: ECM and Knowledge management needs such a lot of consultancy, process management and change management that there is enough to do for Alfresco integrators and ECM consultants.

Kind regards.

[Ronny Timmermans]

Open source allows to quickly reuse, develop and debug software by more than one company. However, what really matters, is getting a great product at a low cost. Open Source delivers on this, as it breaks the proprietary barriers of reusing and improving software.

But the question, John, is how you pass on that value to your customers? Your customers have the tendency not to pay for what you have done (putting together a great ECM, releasing it in open source, and thus forsaking the secrecy & exclusivity premium that you could collect), but for what you will do. What does Alfresco sell? A subscription to support, an insurance as you put it, i.e. your expertise. This is much better than the "right to use" (which is as closed source as proprietary can get). If your support is great, consultative, pro-active, your customer will repeatedly enjoy an on-demand expert, available within 4 hours, to solve a problem that would costs them days (of an engineer, out of production, etc..). This is sellable value. Your people, network (in or outside) and tools (that you do not have to share :-)) will drive companies to subscribe.

Paying Alfresco must be the cheapest option to your users (in any perspective: per user, per document, per search, per CPU, per incident): Any other competitor, including the internal staff (surfing on the waves of the 'community' edition), partners or plain "downtime", should be more expensive.

Maybe we can push innovation on support: preemptive, auto-diagnostic capabilities that can be offered in enterprise as a service in which every warning and error of Alfresco is sent to Alfresco (compliant with security rules), so YOU can fix before users harass the IT manager. In the end, he is paying for a good nights' rest :-).

It's not about free beer. It's about cheap beer, of a consistent quality, chilled, and available on demand. Like Trappist!

[Michael Tiemann]

I take the macro view when I consider what's fair. The macro view says that globally the world spends $3.4T USD on information and communication technology, and that $1T USD is wasted on efforts that are abandoned before ever reaching production or on "bad software" that is late, functionally defective or deficient, or both. That $1T USD per year is a big waste, especially now that money actually matters again. (The full blog posting is here: http://opensource.org/node/384 )

I think what's fair is to be paid to deliver a solution that costs a nickel and saves a dollar of waste.

I believe that if a F10 or F50 company is using your software and not paying the terms that you ask, either they or you are not clear on where the true waste is in that F10 or F50 company and how you can help them clean it up right quick.

Having been at Red Hat for 10 years (and in open source for 20 years), I've seen 1000s of "transformations" where IT executives have realized that spending money on good open source technologies, people, and practices has a multiplier effect, not only on their own budgetary performance, but on what is possible for the company itself.

That is what is so exciting about open source! Companies who "buy" the "free as in beer" part of open source while dumping hundreds of millions of dollars on broken proprietary products are getting what's fair: degrading IT performance that leads to degraded company performance. Neither capitalism nor open source were designed to reward stupidity, nor ignorance.

But for those that are willing to spend that nickel and to retool from proprietary lock-in to open source software...great is their reward! And rapidly does it improve!!

[João Martins]

I agree and understand that you have to have revenue to develop the software, but sometimes the cost of licensing the enterprise version of an open source software is too expensive.

I think the open source version and enterprise must always be the same source, and should be released at the same time. However, the enterprise version has the modules for "Enterprise" and the bugs fix by Alfresco, like you suggest.

So also will be easier to start a company with the open source version and then move to the enterprise version.