The concept of the shim DIF over Ethernet
For migrating from current TCP/IP over Ethernet networks to RINA, the concept of a shim DIF was introduced. A shim DIF consists of shim IPC processes which are not fully functional IPC processes, but just present a layer in the current internet model as if it was a regular DIF. The task of a shim DIF is to put as small as possible a veneer over a legacy protocol to allow a RINA DIF to use it unchanged. Applications in the current internet model can keep on using the existing layers in this way, while applications using the RINA API can use the advantages RINA has to offer. Within IRATI, the shim DIF over Ethernet has been defined, which the prototype will use.
An Ethernet shim DIF spans a single Ethernet “segment”. This means relaying is done only on the MAC addresses. This also means we assume the number of users of the same shim DIF is small. It is not the case that all stations on an Ethernet “segment” are by default members of the same shim DIF. Each shim DIF is identified by a VLAN (IEEE 802.1Q) id, which is in fact the shim DIF name. Each VLAN is a separate Ethernet Shim DIF. All the traffic in the VLAN is assumed to be shim DIF traffic. All members of a VLAN are assumed to be members of the same shim DIF.
Instead of implementing its own directory mechanism, the Ethernet shim DIF reuses ARP in request/response mode to perform this function. ARP resolves a network layer address into a link layer address; that is, in the context of the shim DIF, mapping the application process (AP) name to a shim IPC Process address, exactly the function that the directory in RINA provides.
If you want to know more you can take a look at the shim DIF over Ethernet specification, available as part of IRATI’s D2.1. It can be found in the Deliverables area.