Stacking it up with open source
Dave Rosenberg's post provoked some interesting thoughts about what I consider to be the next significant wave in the evolution of the open source software industry, packaged infrastructure software stacks. The dawn of their emergence is being heralded by the strong growth of lower cost open source infrastructure components in the operating system (Linux, FreeBSD), middleware (JBOSS, FUSE), and database (MySQL, PostgreSQL) categories. The next logical step consists of identifying solution stacks that meet vertical and horizontal market needs.
These solution stacks can be defined as logical groupings of open source software that are not necessarily compiled into pre-integrated packages (even if it is helpful if they do). The immediate question is how this is of any value? The answer lies in the fact that there has yet to be a consensus in terms of how to best apply multiple working open source products towards streamlining multi-layered issues across an enterprise.
For example, the need to mobilize applications and data along with the requirement to integrate collaborative functions for an organizational workforce has highlighted the need for mobile infrastructure and web-based access to scheduling, calendars, etc. Seeing how these two exhibit a good bit of overlap in terms of their scope and requirements, it only makes sense to investigate the viability of relevant open source solutions. A combination of Linux, Zimbra and Funambol could be positioned as a potential stack. Because the stack does not exist in the traditional sense of a shrink-wrapped, end product, and because each component can be evaluated as an individual, suddenly it becomes a moot point that there isn't a comprehensive platform available.
Less direct forms of these solution stacks are already being used across the world when there is an immediate need to cut the acquisition costs related to enterprise software. As an example, it is becoming increasingly popular for teams within data centers to roll their own management system/toolset using a mix of open source and proprietary offerings. However, in the coming future infrastructure stacks will increasingly be considered not only because they are simply a 'cheaper alternative,' but also due to the fact that they introduce a level of flexibility and choice that has the potential to operate along the lines of introducing a valuable number of growth factors.
When the value proposition(s) for stable, mature open source software is combined with other open source pieces, proprietary software has its room to maneuver (see: margins) further decreased. This fact is what makes the acquisition of JBOSS by Red Hat all the more relevant. Further down the line as Red Hat pushes closer to the $1 billion mark, a possible RHEL stack coupled with direct support for JBOSS is going to make a strong statement within the enterprise server market. It will also most likely help RHEL gain an edge in its market share battle against Novell's SUSE as well as hold off the fast approaching Ubuntu.
However, most open source vendors do not have the cash (or desire/need) to purchase a top open source project, like a JBOSS, for the purpose, of among other things, pursuing the prospect of rolling out a stack sometime down the line. So their approach will have to depend on identifying where the best opportunities are in terms of positioning a specific product as a viable member of a given stack, even if that stack has yet to materialize. Association (as vague as the term may seem) with other members of a logical stack has a valuable self-reinforcing tendency that helps validate the value of each member as well as its capability to meet the needs of specific area. Both of which are healthy for the open source software industry now, as well as further down the line.
Comments