As many people reading this blog will know I’ve been a member of the Apache Software Foundation since 2002. Membership is awarded by existing members in recognition of contributions to the success of the foundation as a whole. Those who speak to me about this know that I consider it “a big thing”. I refer to it as my “badge of honour”, I believe it is the one thing I have that truly singles me out as an “expert” in this open source thing.
But just why do I feel this way?
I guess it is because, as a member, I am acutely aware of how the ASF generates massive value from an open development model. It’s sometimes hard to illustrate just how much value there is in the ASF since it is a non-profit and therefore cannot be sold, like MySQL, for masses of cash.
I usually point to the fact that the ASF manages the Apache HTTPD project (the web server that serves somewhere between 50% and 75% of the worlds internet pages). But for me this feels a little hollow since I don’t work directly on the HTTPD project.
Today was the Members meeting. In that meeting we were treated to some wonderful stats:
The foundation has, at the time of writing, 1765 committers, that is people who have write access to at least one of the 65 Apache projects.
Collectively these committers have made 662,663 changes to the various Apache projects in Subversion.
For the month of May 2008 the ASF web servers received an average of 5,995,132 requests per day with a further 1,572,052 requests to the subversion servers and 479,636 requests to the issue trackers.
In terms of data transfer that amounts to an average of 2.908 TB per day from the web servers and another 21 GB per day from subversion and svn.
(For stats junkies, there is more info on the ASF web stats)
From a mail processing point of view there were 1,147,191 connections to the ASF mail servers on May 30th alone (an average day by all accounts), over 98% of these were rejected (presumably because they were spammers).
Impressive as those stats are they become all the more impressive when you consider that the Apache Software Foundation does all this with an expenditure of only $181,202 in the 2007 fiscal year.
How does the ASF do this?
Simple, they rely on a fully transparent model that recognises contributions of individuals rather than commercial entities. There are no ties to any commercial bodies and therefore there is no possibility of a commercial body taking control of a project for their own interests and at the expense of other people.
Everyone is able to contribute and everyone is rewarded for that contribution. As a result, lots of people do contribute, regardless of their technical skills – documentation, user support and bug reporting are equally valued alongside coding in the ASF.
Since the ASF can maintain 65 world class projects on a budget of well under $200,000 I find it hard to understand why some people in the academic community struggle with the concept of open source, and in particular open development, as a model for sustainability.
I suppose I should note that the ASF illustrates only one model for open source sustainability, we discuss others in our wiki. Similarly, the ASF is not the only organisations that survives through open development. However, with these stats it is one hell of an example, don’t you think?
These are very interesting numbers Ross. I assume the ‘requests’ include page hits and downloads. Do you have figures for total downloads (over the same time period that the commits is given)? I would be very interested in those?
Thanks, James
James
Downloads are not included in these stats (either requests or bandwidth) since all ASF downloads are from an extensive set of mirror servers rather than the core web servers.
Unfortunately this makes it difficult to collect accurate download figures.