When network connections are associated with ephemeral terminations in the MG@ certain resources are used. A failure in the MG may result in the loss of certain groups of resources@ whilst other groups of resources remain unaffected. The resources that share a risk of failure are called a shared risk group. A shared risk group is simply a grouping of the network resources (e.g.@ IP interfaces) sharing the same risk of failure. Each shared risk group could@ for example@ contain a number of IP interfaces located on the same piece of hardware. If a resource failure occurs@ only terminations in the shared risk group corresponding to the failed resource are affected while other terminations are not. The MGC may select different shared risk groups when the redundancy of network resources has to be controlled@ or when the MGC wishes to specify which group of resources are to be used. With this functionality the MGC is given control of the resources used by ephemeral terminations in the MG. The MGC may order the MG to use resources from a selected Shared Risk Group@ or not to use resources from a Shared Risk Group. This gives an advantage over using Termination ID schemes to link resource groups in that the MGC can specify certain resources to avoid in a CHOOSE ($) Termination ID scenario. The capability to order the MG to use diverse resources for different connections is particularly useful when a pair of connections is used for a primary and secondary link e.g.@ for signalling transport. Figure 1 shows an example of a scenario where two shared risk groups have been defined and are used by ephemeral terminations. Context 3 represents a Primary Link and Context 5 is its secondary link. The Shared Risk Group Package may be used whenever the behaviour described in the procedures section of the package is required.