The financial value of Open Source

What’s the cumulated financial value of the today existing open source code base? It’s a big number, for sure. If you take published numbers and multiply them in a simple exercise, you might come close to the truth. Blackduck claims to have indexed a couple of billions of lines of code in its repository. Assuming they only index 70% (they mention 185’000 projects) and “a couple” is “5” then we might end up with 7 to 10 billion lines of code. If we multiply this with the estimated value (or cost to develop) of one line of code (let’s take 15 USD for the exercise here), we end up with a total value of USD 100 to 150 billion. What is interesting here is that only a little piece of this is funded by Venture Capital (the 451 Chaos blog mentioned 2.95 billion USD total investment (and most of it goes into marketing and sales, rather than into product development) until today), so where is the rest coming from? I would assume that some of it comes from enterprises and large organizations, but this is probably max 20-30% of it. The bigger share actually comes from individuals and small companies, people either sharing what they have done with a social idea behind or planning to build or extend a business around it.Another interesting question is whether these “investments” will go up or down during the tough times we are going through. I actually think they are likely to increase because of two reasons. First, there’s a lot of development capacity available in large companies, but little money to spend for buying software. Companies are trying to keep their staff busy, so it’s well possible that they invest in some piece of software and share it with others. Secondly small companies will suffer of less demand from their customers, they will have time available to invest into the future. As the open source approach has proven to be a good vehicle to publish and market software, many will take this route.What does this mean for all of us? Well, yes, we can expect more good software coming from the Open Source world and the innovation will continue.

Planning to redesign and improve EOS Directory – need your input!

After almost one and a half years of existence, it’s time to think about a redesign of EOS Directory. – Not only feels the current design a bit old fashioned, there are also a number of usability aspects to be corrected. And not all the functionalities and features were really – successful.We would of course appreciate input and feedback when – going for the redesign. So if you have good ideas on how to make EOS Directory more effective for you, then please drop us an email to already have some ideas. Looking at the analytics for example it seems that more than 76% of all the users are having a browser with at least 1280 pixels in the horizontal, no need therefore to focus on 1024. We have even seen quite a significant number of people with more than 3000 pixels! There are few people with iPhones and the likes, but if we can easily do it, we will prepare a specially rendered access for small devices.The focus in the redesign will be put on a more useful directory section while we will also improve the content around what makes Open Source enterprise ready. – We certainly will simplify the forum section and try to make user input simpler. We plan to add a newsletter and a project specific alert function. – We would also like to integrate some external data (e.g. ohloh, sourceforge and the likes) if possible.But again, if you have good and implementable ideas, don’t hesitate to send us an email. – Thanks for your contribution

Is Open Source more cost effective? Is Alfresco 96% cheaper than Sharepoint, Documentum or OpenText?

My colleague Jeff Potts analyzed a recent white paper coming from Alfresco, comparing the cost of Open Source versus Proprietary solution. According to Jeff, this is a good read. Here’s his assessment:If you are evaluating ECM solutions, particularly if you are interested in cost, you need to take a look at Alfresco’s TCO Whitepaper. In it, Alfresco uses licensing numbers they snagged from the United States government to compare the first year costs of their solution with EMC/Documentum, OpenText, and Sharepoint.When the whitepaper came to my attention, I expected it to be Marketing hype, full of soft numbers and exaggerated claims. While readers must take the paper with a grain of salt considering the obvious bias of the source, Alfresco does a good job of avoiding Marketing speak for the most part and simply laying out the facts. The whitepaper shows line item detail for licensing and support for the first year. If you want to include supporting infrastructure (OS, application server, database) in your analysis those are provided for you as well.The paper shows that for document management plus collaboration and integration with SharePoint, you’d have to pay EMC/Documentum $863,937.98 for a 1000 user configuration as opposed to $318,738 for SharePoint and $33,500 for Alfresco for similarly-sized systems with equivalent functionality. Those numbers exclude the supporting infrastructure software.So what’s the fine print? Here are some considerations…The numbers Alfresco used are from a government price list. It isn’t clear to me whether those numbers are “list” or are a negotiated, reduced rate, but from my past experience with Documentum, I’d say they are closer to list. I don’t think it is likely that anyone would actually pay $800k for a 1000-user Documentum system. Even if you were to negotiate 50% off of those numbers, though, the difference is still significant.A portion of the “first year’s cost” is maintenance and that recurs every year. For Alfresco you are only paying for maintenance, so the entire $33.5k will be due every year. Using the numbers from the whitepaper your Documentum maintenance bill would be about $115k every year. I think in all cases, the maintenance is probably understated for what typical clients will pay because most will want “top shelf” SLA’s. The numbers used here are for lower levels of service.The legacy vendors have 1000’s of product configuration options. The line items Alfresco chose to include for the Documentum configuration look roughly right, but with so many options you can’t say with certainty that what’s listed is what everyone who needs a 1000-user document management system built with Documentum will use. So tweak the table using the quote your vendor gave you and come to your own conclusions.Alfresco showed a 2-CPU configuration for their 1000-user config priced at $33,500 which included a test server. Then they showed a “high availability” config with a $9,250 up-charge. But they didn’t double the procs. If you’re going to be HA, you’ll need at least two of everything. While they did double the test server procs, they didn’t double the production server procs so the HA version of the 1000-user config should be more like $76,250, in my opinion. Incidentally, it isn’t clear to me what you get for that extra $9,250. I have an open question with the Alfresco folks to clarify both issues.What about services? Honestly, it’s usually a wash. There are things you can get done faster because you can see the source code but there are other things you may end up spending more time on. When it comes to services, the primary value of open source is in the ability to spend less on the software and still end up getting something closer to what you actually need through customizations (See “Why Open Source?”).Obviously, big decisions like this should never be made on cost alone. Documentum, FileNet, SharePoint, and Alfresco aren’t perfectly interchangeable. You still have to figure out which one is a better fit for you along all sorts of dimensions. But the stark analysis Alfresco is providing is likely to get a lot of attention from buyers who are particularly price-sensitive in today’s market.

Can open source companies be successful in 2009 and 2010?

A recent blog entry by John Bennett posted the question, whether many open source vendors will continue to struggle, and some well-known vendors will shut down. This triggered some thinking on my side, as I hadn’t thought about this recently.The question on whether companies can make money with open source has been discussed broadly and lengthy. And yes, there are many good examples of companies being able to make money and attract investors. We all know RedHat, JBoss, MySQL, SugarCRM, Alfresco, MuleSource to name just a few. Well, we do not know exactly how much money they are making, with the exception of RedHat of course. Many so called successful open source company actually just made money by selling the company. Obviously it takes more than a couple of developers, ideally working for free, and a good idea to come up with a convincing business model. And the current recession will even more clearly point out who the winners and who the loosers are. What is clear is that pure services revenue don’t drive very high margins and don’t create the interesting multiples VCs are hot on. What is also clear is that combining the open source and the proprietary software product model is a challenge. And, what we have learned the hard way is that sales and marketing is expensive and difficult to get right, even if we leverage all the cleverness of modern marketing 2.0 techniques. So, stay tuned. In 2009 and 2010 we will most probably see a major shakeout in this space. Only the strong ones will survive, not that different to other markets.

Open Source Year 2008 in Review – More Adoption, Success, Innovation, and Alternatives

2008 was an important year for Open Source and a successful one in addition. We have seen more adoption, more commercial success, more innovation, more collaboration and more options for the IT buyer. And it’s not the end, more success is still to come. The following paragraphs are summarizing what we have seen in the last 12 months.Accelerated Adoption of Open SourceOpen Source has clearly become mainstream for enterprises. Everybody from Gartner to IDC and to published about the continuous adoption of Open Source by the Enterprises.The financial crisis and the threatening recession helped to promote Open Source even more and position it as a recipe to deal with lower IT budgets.Open Source adoption continued to be strong in Europe specifically. Germany, France and Spain alone can compete with the US in terms of adoption rates.Security remained a major concern with commercial software and has driving adoption of Open Source alternatives such as Firefox on the cost of Microsoft’s Internet Explorer.Finally forced the consolidation in the commercial IT world – e.g. Oracle buying BEA, acquisitions by IBM or OpenText in the ECM space, the consolidation in the business intelligence domain in general – many decision maker to rethink his technology choices and to look at Open Source alternatives.Successful commercial Open SourceThe leading open source vendors continued to prosper. Alfresco, SugarCRM, RedHat/JBoss, etc. all announced good progress and seemed to be continuously beating their business plans.A quite large number of Open Source vendors have received additional funding, including OpenBravo, Alfresco, Optaros, DotNetNuke, SugarCRM, EnterpriseDB, GreenPlum, JasperSoft, Open-Xchange, Nuxeo, besides many others. The big question of course is for how long the cash injections need to last, given the difficult times in the economy.For faster growth and more impact in the market, Open Source providers have acquired other companies and technology providers. Examples here are SpringSource buying Covalent or RedHat buying Qumranet.Foundations continue to be a key driver for sustainable open source development. Outside of the well known established foundations such as Mozilla or Apache, the Django Software Foundation was established to foster the development of Python based web framework Django.Open Source Enterprise Content Management of high interestEven more than in the years before Open Source Content Management Solutions have been pushing back their commercial alternatives.Web 2.0 and Enterprise 2.0 initiatives in large companies more and more involved evaluations of open source stacks (i.e. Alfresco, Liferay, Drupal, etc.)The interest of IT decision makers can also be interpreted from the fact, that the top 3 technologies researched on EOS Directory in 2008 are all Open Source ECM solutions.Innovation and pushing the envelope by Open SourceIn 2008 we have seen a number of influential new releases of top ranked open source projects: Firefox 3.0 (8.3 million downloads in 24 hours), Open Office 3.0 (almost 18 million times downloaded in less than two months) to name just two.Exciting new open source projects were launched, such as Google Chrome (browser) or DimDim (web conferencing).The rise in interest in mobile business application can be clearly attributed to the success of Google Android and its proprietary alternative iPhone, that for whatever reasons is also loved by many Open Source geeks.With the success of cloud computing a number of relevant open source technologies surfaced, such as Eucalyptus, Globus Nimbus, RESERVOIR, OpenNebula, etc.It was a productive year with no doubt. More than 80% of all projects tracked on EOS Directory published new releases, many of them multiple times.Open Source communities have continued to produce and invest, as a matter of fact twenty-three percent of all downloadable Open Source code was released or renewed in 2008 according to some research of BlackduckOpen Source and Open StandardsOpen Standards have again received strong attention, are still rising in importance and are heavily supported by Open Source. At the same time Open Standards are driving interest for Open Source technologies: CMIS and the fast adoption by open source players such as Alfresco, Magnolia or Joomla is one example. OpenID was adopted by large players such as Google, IBM, Microsoft, VeriSign and Yahoo!. Microsoft finally decided to support OpenDocument format. New standards such as openAuth, DiSo, OpenSocial are generating more and more interest.Traditional commercial vendors adopting Open SourceIf you can’t fight them, join them. We have seen a number of acquisitions of open source companies by traditional IT vendors. Sun acquired MySQL and bought VirtualBox (Innothek), Novell acquired SiteScape, etc.Many commercial vendors tried a hybrid strategy, among them Iona or Nokia who acquired Symbian/Trolltech), Microsoft putting more emphasis on Open Source with CodePlex and the (osi certified) Microsoft Public License, Adobe (, BT (acquiring Osmosoft, striking a large deal with SugarCRM).Some vendors were opening their software for enlarged market reach, i.e. Oxid eSales following the example of Ingres and many others before.Solving the Support “problem”Support, seen as one of the barriers for Open Source adoption in the past, has become less of an issue over time. With self support offerings, as promoted by Sourcelabs for example, new options for the IT buyer increase the already well established choices. New companies focusing on services around specific technologies receive both funding and customer interest, such as the company Acquia offering software bundles, extensions, support and professional services around the Open Source technology Drupal.Hopes that remained unheardNot all wishes were fulfilled in 2008. There is still no modular mainstream java based CRM platform, we still see no clear winner in the Content Management domain, actually there have been more options added during the last twelve months than removed. There’s no single leading technology visible to develop Ajax applications or to implement portals. And there are still too many choices for web programming frameworks the confused enterprise developer is confronted with. And, did we really need yet another internet browser? But choice is always better than monopolies, so let’s continue to be “open”.

Do Excellence, Enterprise Readiness and Popularity go together in Open Source?

Doing some analysis of the EOS Directory listed projects and user feedback in terms of implicit popularity feedback (project page views) we came up with the following interesting correlation between Popularity and Excellence:

There clearly seems to be some correlation between the excellence/quality of Open Source projects and the popularity on EOS Directory.So “being better” pays off in Open Source. Well, this is what you expect anyway, right. Interesting though is the category “business applications”. As analyzed in an earlier blog entry, business applications are clearly the most popular on EOS Directory, despite their partially less excellent ratings. However the need for alternatives to the incumbent proprietary technologies is so big, that enterprises more and more look for Open Source options here, pushing up the popularity of this category.On the EOS home page the most popular five projects are listed, but this is basically the ranking since the creation of EOS Directory. But what are the most popular 20 projects of the last 12 months?# Project1 Nuxeo EP 52 KnowledgeTree3 Alfresco4 Pentaho5 vtiger CRM6 Darwin Kernel7 Nagios8 phpBB9 Apache10 Tomcat (Apache)11 MySQL12 PostgreSQL13 SugarCRM14 Drupal15 Python16 Trac17 Lucene (Apache)18 Java19 Zimbra20 DjangoThis list actually tells a lot. Not only represent these 20 projects almost one third of all the “project page views”, the popularity also shows the interest in specific categories and subcategories in EOS Directory. Clearly, business applications and in particular Enterprise Content Management and CRM are very popular. The fact that 40% of the projects are business applications clearly shows the interest of enterprises in deploying these in their business. This is clearly a change compared to 2006 or 2007. And of course the financial crisis and the recession “help” here. The three most popular projects are all ECM technologies, this is telling as well.

Open Source on the Move – EOS updates over the last 12 months

Over the last week EOS Directory has been updated by adding the most recent learnings from projects and research. Never before an Open Source Directory for Enterprise usage has been more relevant and needed! EOS is THE resource to help enterprises make more of their IT budgets during the expected downturn.EOS Directory has been around now for quite some time, exactly 16 months actually. In the last 12 months, over – 249’000 – projects have been investigated on EOS directory and there’s quite a bit of anectodal evidence that people find EOS Directory valuable as a research resource.Today EOS lists over 350 projects (up 52 since a year ago), for example 8 application servers, 30 frameworks, 11 programming languages, 7 rules engines, 17 systems management tools, 24 CRM/ERP/eCommerce solutions, 28 ECM solutions, 7 ETL toolsets, 4 search engines and 6 business process and workflow management platforms. The category application development and infrastructure seems with 151 projects and an average enterprise readiness rating of 2.4 to be the most mature, 31 infrastructure solutions with an average enterprise readiness of 2.0 and 98 business solutions with an enterprise readiness of 1.9 document that here’s room to catch up. Interestingly the “business solution” category seems to be the most interesting one for the visitors, with more than 1’300 project detail views per solution since the beginning of EOS Directory.To add some statistics: During the last 12 months 64 new projects were added, 11 were removed (Celtix M2A, Centric CRM, ERP5, Ferret, jEDit, JFreeReport, Majordomo, OpenFTS, OpenRico, PostNuke, Xfire). 288 are still the same as 12 months before and 254 of these (88%) developed further (changed release number). 61 projects have a higher overall rating than before, 52 a lower one, 14 improved their “trend rating”, while 31 received a more negative trend rating than before. We list 19 4-Enterprise-Readiness-Stars, that’s 5% of all projects and includes Apache webserver, Tomcat, MySQL, Python, Java, PHP, Hibernate, Jboss AS, Spring, RedHat and Suse Linux, Firefox and jQuery.So, let me end this message, saying THANK YOU to all who helped making EOS Directory a valuable resource and keeping it updated! And this includes both the consultants at Optaros, industry experts with valuable input and feedback, as well as the visitors and users of EOS Directory.

Open Source Catalogue 2009 (in German) published

While EOS Directory is an online resource, many people like offline resources and easy to consume “catalogues”. For the last three years we have published the Open Source Catalogue on an annual basis. We started with 250 projects, the latest 2009 edition now lists more than 350 projects. The ratings and descriptions (translated) are taken from EOS Directory of course. The Open Source Catalogue 2009 also includes an introduction on how Open Source is influencing Enterprise IT. As said it’s written in German and available as downloadable PDF document.

WordPress, the Movie

Many WordPress bloggers, myself included, have been eagerly awaiting the impending release of 2.7. It’s already in its third beta release (the current 2.7 Beta 3 was released on Nov. 15th) and should be ready for production release toward the end of November. Jane Wells posted this video to show off some of the features coming in 2.7, including a significant dashboard redesign:The WordPress community and Automattic have always done a great job of getting behind the project not just with code contributions but also with "marketing" contributions, helping spread the word about new releases and new features. What if every significant new release of an open source project came with an embeddable, distributable video screencast walking through the new features? Might lead to better upgrade activity across the community.