One other overarching issue that plagues most, if not all, traditional DCA solutions is their inefficient architectural approach to system management. They all use either procedural or image-based methods. The image-based technique essentially makes an image of the entire file system and clones the drive, hogging both network and storage capacity. This “make-a-copy-of-everything” methodology typically breaks down whenever hardware is changed, and it leaves no audit trail. It also requires administrators to spend time and energy updating every copy of the OS and to patch every instance of an application. Similarly inefficient procedural approaches are highly labor-intensive as they automate installation and change processes for operating systems, applications, or patches. Further, undoing a change or rollback to the previous state becomes more and more challenging over time as more changes are made, particularly as application and hardware platforms evolve. Despite these limitations, enterprises are often forced by the traditional DCA solutions to choose one approach or the other to manage multi-OS environments. Both approaches require significant manual configuration effort, making them slow and cumbersome. Unfortunately, the traditional DCA vendors can’t serve multiple OS without that compromise. In short, both image and procedural approaches are less than optimal (and too slow and cumbersome) to address the needs of a truly agile Linux data center.

The bottom line? IT departments that choose Linux for their data centers lose many Linux advantages unless they choose an automation solution that is specifically engineered to support Linux environments. Besides being designed to handle the challenges presented by Linux, these solutions help sysadmins manage system dynamics better overall. One way to see this difference is to compare the aforementioned procedural or image-based approaches to system management with state-based system management, an innovative approach to addressing these issues.

During a server’s active life, its operating system and application files are constantly evolving or changing “state.” When a server fails, it’s difficult to return that system to its exact previous state; add in the inherent complexities of Linux and the labor and time required by the image-based and procedural methods and this becomes even more challenging. State-based provisioning treats system state changes as if they were data and specifically addresses the need to manage a large number of Linux servers in constant flux or changing state. How can this be done? By virtualizing the entire Linux software stack (the OS, middleware, application, and configuration), many systems can be easily and effectively maintained as just a handful of systems, which provides a foundation for operational scalability and robustness in Linux systems. Other more tangible benefits of this approach include portability in the IT environment (which, in turn, means provisioning or migrating machines in minutes), easier and less cumbersome rollback, and the negation of the need to image-copy across the network (which offers a major speed advantage).

The important thing to remember is to choose a DCA solution that’s appropriate to the enterprise environment, whether multi-OS or solely Linux. The Big 4 frameworks vendors, in their quest for breadth, sell complex software that treats Linux as an afterthought, while delivering cost, complexity and poor integration with Linux. Independent DCA vendors either ignore Linux or deliver incomplete solutions. Linux distribution vendors offer point solutions for their own flavor of Linux and don’t offer comprehensive DCA capabilities. Ideally, enterprises with data centers that run on Linux should opt for a DCA solution that is designed specifically for Linux and specifically to handle the management challenges that Linux presents. To fully leverage Linux capabilities (and future enhancements as seamless extensions) while minimizing the potentially deleterious impact on data center productivity, enterprises have to look outside the box and consider pure-play Linux DCA solutions; they answer the modern challenges of Linux-dominated data centers without bringing along the historical baggage of inefficiencies and hidden costs associated with traditional solutions.