Netmask edges

Unlike IP, XIP does not have a analogue of network masks to identify which addresses are locally reachable. Just having a list of hosts attached to the local network to mimic netmasks requires to scan all nodes in a DAG, for example, consider a packet destined to AD_1->AD_2->HID_1->SID_1 whose last-node field points to AD_1; without the scan, the packet would be forwarded to AD_1 even when HID_1 is on the LAN because AD_1 would not be in the list of locally attached hosts.

A similar problem happens when the intent, SID_1 in the previous example, is available on a host forwarding the packet. This similar problem was solved in the design with the shortcut routing addressing style. In fact, one can also handle the lack of network masks adding high-priority edges from the entry node to HIDs in a given address. These extra edges, here called netmask edges, avoid the scan of all nodes in the address, and does not require another element/solution in XIP.

In spite of the fact that XIP does not distinguishes shortcut and netmask edges, they should be seen/handled differently. Whereas shortcut edges are intended to query the network, netmask edges are intended to look for local hosts. This difference must be handled by applications; likely the DAG library that will be used by applications to compose addresses.