OpenBSD may be one of the world's most reliable and secure operating systems available, but the hardware it runs on is still prone to the occasional break down. There are many steps we can take to ensure that your network is available at all times. At the center of this is a decision about the amount of downtime that can be tolerated by your operation. For example, a company that relies entirely on web sales for its income may take a "zero downtime" approach to reliability.

Although this might be a laudable goal, it isn't very realistic, unless your company is in a position to spend a large amount of money duplicating all of your network. If a NIC or a switch should fail, there will be network downtime, although with good diagnostic routines and the availablility of replacement hardware, the downtime can be absolutely minimized.

Here is a table of some approaches to network redundancy to help you decide what will work for you.

Should you keep spare parts on hand? Most businesses can tolerate the loss of internal networking for up to a day without a tremendous loss of productivity. Some businesses will find this loss unacceptable.

Your business is unique in this area, too, and this should be a part of evaluating your current and future networking needs.