One of the common criticisms of deploying open source components and emerging stacks is that they are disjointed, hard to integrate, and lack a common management control capability. Today's open source server components amount to a helter-skelter pile of autonomous bricks, and the already overtaxed corporate IT staff is left to the yeoman's work of trying to somehow fashion a solid wall from them -- ultimately at high total long-term cost. This logic has been a cornerstone of how the likes of Microsoft have defended the cost-benefits analysis of choosing Windows Server System over Linux.
This argument, however, is under a steadily building and momentous assault. In a matter of months -- not quarters or years -- many of the hurdles to a smooth, cost-effective, and IT staff-friendly migration to and general use of holistic open source platforms will be whittled away. In the place of the piles of bricks will be pre-configured and tested sections of "wall," with orderly inventories of defined and properly licensed code, off-the-shelf indemnification from insurers, and with an ecology of globally-placed professional services resources to cost-effectively assemble (and remotely monitor) the IT stacks of the users' choices.
This rapid maturation of open source components into refined stacks is due to a huge pile of venture capital money being poured into the creation of the required services mortar to make open source components into mission-critical-ready and tightly aligned solutions. The entrepreneurial, managerial, technical, marketing, and strategic business acumen that has been drawn to this task is staggering. Some of the best minds, with A-team success records in the commercial software business, have left or been recruited from such companies as BEA, Sun, Microsoft, Oracle, and IBM. These folks are now running open source ecology support start-ups; one can imagine their options will not end up under water.
What might an open source solutions ecology look like? I've been briefed lately by a group of companies that gives us a strong sense. There are several players to choose from in each category, and I am by no means anointing winners or best-of-breed players. It's still too early for that. I simply wish to show that the arguments around open source's vulnerabilities and weaknesses arethemselves vulnerable, and growing weaker.
First SpikeSource and SourceLabs are providing cross-functionality assurance benefits, inventory and configuration management solutions over a slew of pre-configured open source-oriented stacks. They include stacks for application run-times, web serving, LAMP, data, J2EE, content management, web services, and CRM. Look soon for more on management and virtualization.
Red Hat, too, is beefing up its management capabilities with an expansion of its management network breadth and depth.
Having automated configuration and installers, coupled with monitoring, remote management and remediation, will go a long way to making IT managers embrace open source solutions.
Optaros is creating a multi-level professional services capability to support the deployments and operations of such open source stacks. And these solutions are being refined to technology categories. I expect further refinement into vertical industry-specific solutions.
And Palamida is providing what I see as IT governance for open source by allowing companies to identify, manage, and inventory the code they have in place (or jettison that which they should not). Such auditing and portfolio management can be extended to all code, commercial or open source licensed, for many benefits, from regulatory and Sar-Box needs to gaining indemnification insurance. (I'm told that a large, old, global insurance company based in London is working on just such insurance.)
In addition to the services now offered by vendors like these, there's more to make open source more readily acceptable by IT departments. What also knocks the wind out of the open-source-is-an expensive-jumble-not-a-solution set of arguments is that these stacks don't necessarily sit on Linux. In increasing cases, the stacks are on Windows, using Windows services as additional mortar, and attracting the legions of Visual Studio developers who know a good total solution when they see it. The same is the case for Solaris, Open Solaris, HP-UX, and IBM AIX. Unix developers with OSes in place are very keen to exploit the value of open source stacks, on Linux or on Unix.
What the new maturity around open source stacks is really saying is: Let the needed solution at hand dictate the chosen components (commercial and/or open source), rather than let the components you are stuck with dictate the solution. This is all good news.