March 2011 Harvard Business Review featured an article titled "Experts are more persuasive when they're less certain". Zakary Tormale stressed the positive influence of uncertainty in many circumstances. In IT and networking areas, there are also many situations where we're better to humbly accept that we don't know everything – to quote a few:
- Are Facebook and YouTube recreational applications? Probably yes most of the time, but what about training video and customer communities?
- Will acceleration and other redundancy elimination techniques improve critical application quality of experience? Might be the case, but what about side effect of accelerating low critical apps?
- Is MPLS network faster than the local Internet access? It's reasonable to think it is the case in average, but what's happening right now?
- Should SAP, Oracle and other transactional traffic be always prioritized? What if they finally take too many resources and hurt badly other applications?
- How to infer the actual usage of an application from a few tests in pilot phase?
- We can reasonably forecast the average traffic at tomorrow peak time… but what about next second activity?
This is exactly why Autonomic Networking is so important to efficiently implement WAN Governance: you know your business goals and IT priorities that translate into application performance objectives. Let's accept that you can't model and precisely define all details about your global network and traffic patterns: you can trust autonomic networking to learn, react fast and finally match stable performance objectives in a very unpredictable environment.
Conscious uncertainty is certainly an IT key strength.