Chronicles of a Wandering Mind

Build vs buy has long been a critical decision point for large organizations. Service providers in particular have often chose to build in house solution to manage the infrastructure for variety of reasons:

• package solutions were not able to cope with the size (scale and performance)
• unique requirements that not met by packaged solutions
• maintaining control over how the solution evolves
• strategic differentiation from competitors

Obvious disadvantages of the in-house build solutions is the potential high cost of ongoing develoment and maintenance, dependency on small number of people, etc. It is understandable that developing every solution in-house for a single organization is a costly proposition regardless of how big the organization is.

Like it or not, most service providers have been driving towards using packaged solutions whenever there is a viable solution in the market. Problem with the packaged solutions is the proprietary nature of them. Hard to find skill sets in proprietary products, lack of innovation by the vendors, feature set not meeting the requirements, scalability challenges to meet the needs of large organizations, and integration hurdles significantly diminish the perceived cost savings and strategic value of these solutions.

Open source solutions offer an alternative that may be best of both worlds. Organizations can can collaborate in developing solutions that meet their common requirements, giving organizations the opportunity to share the cost of development and still have control over the solution, etc. Open nature of the open source solutions make integration much easier as well.

Yet there are no significant open source projects in IT management driven by these large organizations. Most open source solutions is product of companies or handful of individuals, and typically target the SMB market. Service providers invest massive resources for in-house development, often have multiple teams with dozens of developers. Moving the development of these solutions into open source projects would offer significant benefit to these organizations. I’ve seen projects in “maintenance mode”, meaning none of the original developers are around anymore and no one knows how the solution works, and no enhancement is possible.

There is clearly a failure here, inability to collaborate, corporate barriers to share, etc. Whatever it is, it’s preventing these organizations to take advantage of what open source development model would have to offer.

So what does all this have to do with the clouds? There are significant parallels between the needs of cloud providers and traditional service providers in terms of IT management solutions. Scale, multi-tenancy, strategic role of the tools, need for flexibility, open interfaces etc. mean that it is not likely that cloud providers can use packaged solutions offered by Big 4 and others.

The good news is that these companies have open source in their DNA, and already use it heavily, typically using open source components to build internal solutions.They have not been so far collaborating to build solutions specific to clouds as open source, but it is more likely to happen than traditional service providers.

It is already safe to say open source software is the de facto choice in the public clouds. What remains to be seen is whether this meme will transfer through the firewalls into the private clouds. My guess is that it will. Simply because open source tools will be ahead in meeting the management needs in the cloud from scalability to integration. Thought leadership is important and open source tools like Hyperic is paving the way for the open source tools to play a leading role in mangement in the cloud. Will the cloud providers play a more active role in development of open source tools for the management in the clouds? It would serve them to learn from the troubles faced by traditional service providers.