Enterprise software is under attack. Traditional infrastructure players like BEA are seeing their core products replaced with free open source projects, while traditional application vendors like Oracle/Siebel are being displaced by SaaS. But is this a slugfest with only one winner? Will SaaS and open source ultimately turn against each other for dominance of the software business model - WWF Smack Down style - where the once united tag team, after conquering their opponent, starts to fight between themselves? Actually I think not. Just as I don't see traditional enterprise software disappearing completely, I believe both SaaS and open source are naturally suited to different parts of the market and in fact will coexist quite naturally.

One of the more intriguing aspects about open source is the huge variation in capabilities. As an example, consider the Apache web server. With dominant market share, the consensus engineering opinion seems to be it's the best web server on the market, regardless of cost. At the opposite extreme you have something like OpenOffice, a replacement for MS Office. Despite the huge latent demand for a free office suite, OpenOffice still has not hit its stride, even after many years. While OpenOffice does have some ardent fans, I believe that most business users would agree that OpenOffice cannot be used as a replacement for MS Office in its current state.

Open source infrastructure software is, generally speaking, a viable substitute to its commercial counterparts. The infrastructure category includes Web servers like Apache, application servers like JBOSS/Tomcat, and databases like PostgreSQL. There are plenty of distinctions to be made within that category. For example, some might argue that Apache is better than MS IIS, JBOSS is about the same as BEA WebLogic or IBM WebSphere, and perhaps PostgreSQL is not quite as good as Oracle or IBM DB2. But despite those subtle differences all would be considered roughly equivalent to their commercial counterparts.

Contrast this with open source application software. There are many open source content management systems (e.g., Alfresco), CRM systems (e.g., SugarCRM), ERP systems (Compiere), and many others. None of these has matched the same level of success as their commercial counterparts. OpenOffice is a good example. Office functionality is arguably the most important single business application, and superficially, at least, OpenOffice appears to tick all the boxes. However, there are some issues that are sufficient to prevent broader adoption. First, given that 99% of all office docs in existence today are MS format, being able to handle these is a sine qua non. Unfortunately, importing a formatted document or presentation into OpenOffice usually requires a round of repositioning and font resizing to make it look right. The amount of time that takes quickly nullifies any MS Office license fee saving (and for many, that's a sunk cost anyway). Second, with no proper anti-aliasing on drawn objects, presentations in particular look poor with jagged edges in comparison to MS Office. For the same reason it makes sense for sales departments to spend more on a brighter projector to use with their customers; no one is going to present a chart with jagged edges, when with MS office they can have smooth lines.