IT executives frequently use the cloud as a way to break free from vendor lock-in. Companies end up tied to one vendor because of a need to reduce costs -- and a commodity-driven approach to infrastructure acquisition would be one path to achieve this. It is very easy to blame the vendors for high costs because they are the ones presenting large invoices, usually on an annual basis, with very little room for negotiation. Why would they negotiate? Consumers are already locked-in with their sole source technology. It seems reasonable that one way to break free of the sole source technology is by creating a competitive environment.
There are two milestones in a business' IT lifecycle where vendors must compete: pre-implementation technology selection (where vendor lock-in occurs) and post-implementation purchasing.
Some components, such as the server hardware, are standard across the market. However, critical components -- like the content management system -- are highly differentiated, which causes consumers to commit to a single vendor. Not all decisions at this phase are tied to a vendor, but they can still lead to significant lock-in. Choices of operating system and programming language are not necessarily vendor-specific, but they can limit future options.
Vendor lock-in that started at the beginning of a project does not become a pain point until you hit the purchasing phase of the IT business lifecycle. It quickly becomes evident that the major cost drivers are the non-commoditized portions and those vendors, being sole-sourced, are in the best position to negotiate prices to their advantage.
This becomes particularly problematic in the later years of a service's lifecycle, when memories of big productivity gains have faded, leaving only the cost concerns. To add insult to injury, the commoditized parts tend to decrease in cost over time as Moore’s Law and other exponentially improving technologies, such as networking and storage, make these components a smaller percentage of the ongoing cost model, bringing even more focus on sole-sourced vendors.
The obvious approach would be to mandate an end to vendor lock-in. But this won't solve all problems; remember the principle that correlation does not imply causality. Many enterprises seem to have organizational structures that foster an incomplete analysis cost versus value of a system's components. Supply management is forced to focus solely on reducing costs, which does not necessarily assess the value that approach delivers as a whole and by extension, its sole-sourced components.
These components are selected to maximize the system's value at the time of its implementation. But what was appropriate at the start of the project may not always seem so throughout the life of that project -- here is where the analysis gets messy. The cost of the component has to be warranted by the value it delivers. One of two situations will come from this analysis -- either the component still provides value that warrants the cost, in which case the bill should be paid, or it does not, in which case it needs to be replaced. But the latter is expensive and requires a great deal of work. To justify a large project that begins as a cost-reduction effort, new value has to be added, which increases the complexity and cost of the project.
However, had the original service been architected to allow companies to replace one component without reconstructing the entire implementation, companies could realize a cost reduction in the later years of the lifecycle. At the end, vendor lock-in is an implementation issue. There have been many architectural patterns over the years (n-tier client/server, service-orientation and loose coupling) that have allowed companies to swap out components, but the achievements in this area have been limited because of a lack of architectural vision and rigor.
This brings us back to the cloud. Certainly the cloud has provided a way to more efficiently deliver undifferentiated technology such as email, but there is still a large degree of lock-in since moving from one Software as a Service vendor to another is always a big project. But attempting to use Infrastructure as a Service to achieve vendor independence will lead to disappointing ROI.
In other words, moving a gnarly mess in your data center to a cloud data center is not going to free you from vendor lock-in. It will only add a vendor to the list of those you’re locked into. The only way to be free from vendor tyranny is to focus on agile architectures that allow the technology to evolve over the lifecycle of the project.
About the author
Mark Eisenberg is a multi-faceted executive with more than 20 years of experience and is current with latest tech trends, business development and sales, from startups to major corporations. He is an early champion of the cloud at Microsoft, yielding a unique ability to comment both on the high-level strategies and the specific technical issues on which companies should be focused.