A couple of things happened during the past week that, on the surface, have little in common. They speak to a bigger truth, however: IT managers and their bosses prepare for the unexpected -- and realize that it is impossible to have all contingencies covered. The organizations that do the best job of generally protecting their infrastructure will end up ahead.

The first event was the loss of YouTube over much of the globe. The Pakistani government ordered ISPs not to allow customers to the site because it was offering a video thought to offend Islam. The ISPs did this by ordering the Border Gateway Protocol servers that route traffic across ISP boundaries to send YouTube requests to nonexistent addresses. The problem was that this "advice" was disseminated to BGP routers around the world, and the site eventually disappeared.

The second event was a widespread power outage yesterday in Florida. Engineers are still trying to figure out precisely what happened. The working theory is that the failure of a relatively small piece of equipment started a chain reaction that eventually cut the juice to about 475,000 Florida Power & Light subscribers and about the same number of customers from other providers. Russell Shaw points out that the loss of power by almost 1 million certainly included some VoIP casualties, and is asking for feedback from readers who were affected in this way.

Backup powering was a much bigger deal in the early days of VoIP. Now, the ubiquity of cell phones somewhat lessens the requirement for immediate backup phone service. Rick McCharles, a telecommunications consultant outside of Toronto, takes an even-handed look at whether residential VoIP is a better bet than legacy phone service. He is very complimentary to VoIP -- before saying that he opts for legacy services. In a section on security, McCharles points out that unlike emergency calls on traditional networks, VoIP E911 calls may go through a call center before reaching the public service answering point. He also points to other weaknesses in the way devices are powered during emergencies.

This post at Network World says that Florida power outage may be construed as a signal for companies to consider their VoIP disaster recovery plans. This is especially true, the writer says, in considering the redundancy of the link between the enterprise and carrier networks. The writer is aware of some companies that use four or five VoIP carriers in addition to legacy approaches.

It is important to understand, in a general sense, that no service is guaranteed. It may be something as superfluous as YouTube -- or as important to a business as a mission-critical SaaS application or a small business VoIP service. Indeed, no service is more mission-critical to some organizations than Research in Motion's BlackBerry, which suffered its second annual upgrade-caused outage earlier this month. The bottom line is to be as prepared as possible, but to have the wisdom to understand that total preparation is impossible.