Why would someone choose to look at a cloud solution?
How is it different than typical outsourcing arrangements?
What would stop someone from using a cloud solution?
I have been reading some of the work Oliver Williamson has done on contracts and while I am in no way an economist and much of it is only fuzzily understandable (to me) there are some really interesting points he makes that help me understand what cloud might bring to the table. I thought I would mention a few of them in this post and see if anyone has any thoughts.
The first item I came across was his comment that "all complex contracts are unavoidably incomplete", and by extension, as outsourcing is complex, the contractual arrangements are also complex and also incomplete. As a related point, although a contract dispute might eventually end up in court, the aim of the contract is more to act as a guide to how to resolve disputes and to set up the terms of the engagement than to ever actually use the costly (and unreliable) dispute resolution offered by the courts. That is, the contract is a framework, not an accurate indication of real working relations.
The second was his discussion on how and why companies might make the decision to vertically integrate (do everything in house) vs. the choice to use outside agencies. The lower the transaction cost, the more likely the firm will buy, as opposed to build the solution. And what determines the transaction cost?
Asset specificity - the more unique a solution, the more sense it makes to do it in house (I am rephrasing based on my understanding to make this more relevant to cloud computing). So, general purpose technologies will make more sense to have as a cloud computing resource ( and when you look at where cloud computing started you can see this quite clearly). Unique solutions (Asset specificity) would include more contract complexity and the problem in identifying a market price. You might argue that test and development is much more similar across companies than production environments, which would have wider variations in service level agreements. If the risks and safeguards add too much cost then companies will tend to bring this in-house. (that was a condensation of multiple pages of text, so it might be confusing - or just plain wrong. You can read the article referenced below).
This leads to the following guideline - try cloud, then try a hybrid solution, and have recourse to internal resources as a last resort. You might specify this a bit further by expanding the cloud to try SaaS, then try PaaS , then try IaaS.
What does this all mean? Well, I think it means that as contract complexity for cloud offerings goes up as asset specificity goes up, so the more commodity like components of a company are more obvious cloud candidates (e-mail, servers, payroll - you can think up your own list) and the next step might be to take apart complex components into separate pieces so the more standard parts can be commoditized and able to use the cloud more easily. This starts to sound like service oriented architecture (SOA) - the buzz word from LAST decade.
Here is the article I was reading:
|Williamson, Oliver E. 2002. "The Theory of the Firm as Governance Structure: From Choice to Contract." Journal of Economic Perspectives, 16(3): 171–195.|