Customers of Internet Service Providers (ISPs) are increasingly interested in running applications such as voice over IP, video games, and commercial transactions. This new range of applications cannot tolerate poor network performance (high delays or low available bandwidth) or network instability (periods of loss or variation in delay or available bandwidth). Unfortunately, routine events such as equipment failures or planned maintenance cause routing changes, which may lead to transient service disruptions or persistent performance problems. Operators of ISP networks are faced with the challenge of minimizing routing disruptions using current routing technology, which offers little control. In this paper, we discuss routing disruptions from an ISP perspective. First, we describe the causes and effects of routing changes. Then, we provide a set of network design guidelines and operational practices that network operators can use to reduce the impact of routing changes in their network.