A mapping between virtual processing units and physical processing units allows for an increase or decrease in the number of physical processing units installed in a computing device without disrupting the computing device's network communications. At startup, each of the virtual processing units can be allocated the necessary network communication management structures for supporting network communication. The networking subsystem can continue to use the virtual processing units as it would physical processing units, namely packets can be routed through the appropriate processing unit. However, because the networking subsystem is only aware of the virtual processing units, the mapping between the virtual processing units and the physical processing units can continually adjust for changes in the number of physical processing units installed, allowing the number of installed physical processing units to be modified without affecting the operation of the networking subsystem or ongoing network communications.