What is it about Linux and open source that these vendors can use to create a better business? For starters, Linux is free. Not free as in "free beer," but free as in "free speech." No one owns Linux, which is good if you want to use it to change your business. You don't want your business beholden to any third party that you don't control, or you may find your business held hostage by an erstwhile "partner" in the future. Second, Linux is flexible. It can be tailored to maximize the performance of your application. There are also an almost unlimited number of free software utilities that can be readily added to Linux to enhance the value of your application. Apache, Tomcat, and Struts are just a few good examples. Third, Linux runs on industry standard hardware, so there's no threat of customer lock-in from proprietary hardware vendors.

Given these lessons from the market, it still is not obvious how independent software vendors might win their freedom. There are legacy issues to consider. It's impractical to discontinue support for existing deployments on "independent" systems. And there's the issue of revenue recognition. Moving to a subscription model probably means deferring revenue, which in turn creates a mismatch between revenue and operating expenses on an income statement. Finally, not all applications can be delivered as multi-tenant, on-demand services over the network, and it does not make sense for all software vendors to consider delivering a hardware appliance solution because their customers do not want to add another hardware platform to receive the value of the application.

One transition approach for software vendors to consider to lose their independence and gain their freedom is the software appliance concept. Think of a software appliance as a hybrid of the on-demand application model and the network appliance model. The customer receives an integrated solution that combines the application with a streamlined operating system. The software appliance might install natively on the hardware via an installation CD or DVD, or it could be delivered to the customer in a virtual machine format that runs atop another OS or atop a standard virtualization layer like VMware's ESX server. Maintenance for the entire solution is received from the application provider via a simple web user interface. Since maintenance comes from the application provider, it is pre-tested and certified in the exact environment the customer has deployed, so there are no more mismatched maintenance streams from various vendors.

Many will argue that Windows-based apps benefit from the concept of a "universal" platform. Not true. Microsoft doesn't allow the OS to be stripped of all components except those required to support your application (Microsoft even argued in court that a browser was an integral part of an OS), nor does it let you pass judgment on its maintenance stream before releasing it to your customer base. Microsoft has a "one size fits all" mentality that's inconsistent with the concept of a software appliance. The operating system should be secure, reliable, small, and practically invisible. It should also be free so that your application is never held hostage to the technical or economic whims of another vendor. These are not attributes that describe Windows.

Parting Thoughts
Linux, open source infrastructure, and virtualization provide application vendors the historic opportunity to lose their independence and free their customers from the hassles of assembly and maintenance of complex software solutions. The initial transition is likely to require an investment in an optimized port for Linux, and some decisions regarding when and how to transition customers from legacy platforms will certainly be difficult.

It is inevitable, however, that customers are demanding lower starting costs and faster time to value from application vendors, and the legacy method of delivering software applications is not going to be acceptable. The great news is that engineering expense that used to be spent on context issues can be redirected to investments in product features, sales, and marketing. Lower cost of entry also tends to expand the available market, opening up customer opportunities that were previously unreachable due to the expense of complex assembly and maintenance routines. It's time to stop thinking of Linux and open source as a threat or simply another application port. It's time to embrace the open source trend as a strategic opportunity to improve the economics and market position of the software application business.