TRILL

TRILL was designed as a successor to the Spanning Tree Protocol, both having been created by the same person, Radia Perlman. The catalyst for TRILL was an event at Beth Israel Deaconess Medical Center which began on 13 November 2002. The concept of Rbridges [sic] was first proposed to the Institute of Electrical and Electronics Engineers in 2004, who in 2005 rejected what came to be known as TRILL, and in 2006 through 2012 devised an incompatible variation known as Shortest Path Bridging.

General overview

TRILL switches, or RBridges, run the IS-IS link-state routing protocol amongst themselves. In this protocol, connectivity is broadcast to all the RBridges, so that each RBridge knows about all the other RBridges, and the connectivity between them. This gives RBridges enough information to compute pair-wise optimal paths for unicast, and calculate distribution trees for delivery of frames either to destinations whose location is unknown or to multicast or broadcast groups. In contrast to Spanning Tree Protocol, which ensures a loop-free network topology by blocking otherwise active ports, TRILL establishes paths over all active links.

IS-IS was chosen because:

• it runs directly over layer 2, so it can be run without configuration [no IP addresses need be assigned], as it is inclusive of a subset of the Connectionless-mode Network Service (CLNP).
• it is easy to extend by defining new type–length–value (TLV) data elements and sub-elements for carrying TRILL information.

To mitigate temporary loop issues, RBridges forward based on a header with a hop count. RBridges also specify the next-hop RBridge as the frame destination when forwarding unicast frames across a shared-media link, which avoids spawning additional copies of frames during a temporary loop. A reverse-path forwarding check and other checks are performed on multi-destination frames to further control potentially looping traffic.

The first RBridge that a unicast frame encounters in a campus, RB1, encapsulates the received frame with a TRILL header that specifies the last RBridge, RB2, where the frame is decapsulated. RB1 is known as the "ingress RBridge" and RB2 is known as the "egress RBridge". To save room in the TRILL header and simplify forwarding lookups, a dynamic nickname acquisition protocol is run among the RBridges to select two-octet nicknames for RBridges, unique within the campus, which are an abbreviation for the six-octet IS-IS system ID of the RBridge. The two-octet nicknames are used to specify the ingress and egress RBridges in the TRILL header.

The TRILL header consists of six octets: the first two octets include a six-bit decrementing hop count, plus flags; the next two octets contain the egress RBridge nickname; the final two octets contain the ingress RBridge nickname. For multi-destination frames, the "egress RBridge nickname" specifies a distribution tree for the frame, where the (nick)named RBridge is the root of the distribution tree. The ingress RBridge selects which distribution tree the frame should travel along.

Even though RBridges are transparent to Layer 3 devices, and all the links interconnected by RBridges appear to Layer 3 devices to be a single link, RBridges act as link routers in the sense that, in the forwarding of a frame by a transit RBridge, the outer Layer 2 header is replaced at each hop with an appropriate Layer 2 header for the next hop, and the hop count is decreased. Despite these modifications of the outer Layer 2 header and the hop count in the TRILL Header, the original encapsulated frame is preserved, including the original frame's VLAN tag.

Multipathing of multi-destination frames through alternative distribution tree roots and equal-cost multi-path routing (ECMP) of unicast frames are supported. Networks with a more mesh-like structure benefit to a greater extent from the multipathing and optimal paths provided by TRILL than networks with a more tree-like structure.

A host with multiple interfaces running the Internet Protocol requires that each of the interfaces [or the group of interfaces in a team] have a unique address, whereas when using TRILL the multiple-interface host can have a single IP address for all of its interfaces attached to a common broadcast domain — akin to the case of a network service access point address (NSAP) on an end system in CLNP.