Label Switching and MPLS

Label Switching and MPLS

So far, on our tour of the network layer of the Internet, we have focused exclusively on packets as datagrams that are forwarded by IP routers. There is also another kind of technology that is starting to be widely used, especially by ISPs, in order to move Internet traffic across their networks. This technology is called MPLS (MultiProtocol Label Switching) and it is perilously close to circuit switching. Despite the fact that many people in the Internet community have an intense dislike for connection-oriented networking, the idea seems to keep coming back. As Yogi Berra once put it, it is like deja vu all over again. However, there are essential differences between the way the Internet handles route construction and the way connection-oriented networks do it, so the technique is certainly not traditional circuit switching.

MPLS adds a label in front of each packet, and forwarding is based on the label rather than on the destination address. Making the label an index into an internal table makes finding the correct output line just a matter of table lookup. Using this technique, forwarding can be done very quickly. This advantage was the original motivation behind MPLS, which began as proprietary technology known by various names including tag switching. Eventually, IETF began to standardize the idea. It is described in RFC 3031 and many other RFCs. The main benefits over time have come to be routing that is flexible and forwarding that is suited to quality of service as well as fast.

The first question to ask is where does the label go? Since IP packets were not designed for virtual circuits, there is no field available for virtual-circuit numbers within the IP header. For this reason, a new MPLS header had to be added in front of the IP header. On a router-to-router line using PPP as the framing protocol, the frame format, including the PPP, MPLS, IP, and TCP headers, is as shown in Fig. 5-62.

The generic MPLS header is 4 bytes long and has four fields. Most important is the Label field, which holds the index. The QoS field indicates the class of service. The S field relates to stacking multiple labels (which is discussed below). The TtL field indicates how many more times the packet may be forwarded. It is decremented at each router, and if it hits 0, the packet is discarded. This feature prevents infinite looping in the case of routing instability.

MPLS falls between the IP network layer protocol and the PPP link layer protocol. It is not really a layer 3 protocol because it depends on IP or other network layer addresses to set up label paths. It is not really a layer 2 protocol either because it forwards packets across multiple hops, not a single link. For this reason, MPLS is sometimes described as a layer 2.5 protocol. It is an illustration that real protocols do not always fit neatly into our ideal layered protocol model.

On the brighter side, because the MPLS headers are not part of the network layer packet or the data link layer frame, MPLS is to a large extent independent of both layers. Among other things, this property means it is possible to build MPLS switches that can forward both IP packets and non-IP packets, depending on what shows up. This feature is where the ‘‘multiprotocol’’ in the name MPLS came from. MPLS can also carry IP packets over non-IP networks.

When an MPLS-enhanced packet arrives at a LSR (Label Switched Router), the label is used as an index into a table to determine the outgoing line to use and also the new label to use. This label swapping is used in all virtual-circuit networks. Labels have only local significance and two different routers can feed unrelated packets with the same label into another router for transmission on the same outgoing line. To be distinguishable at the other end, labels have to be remapped at every hop. We saw this mechanism in action in Fig. 5-3. MPLS uses the same technique.

As an aside, some people distinguish between forwarding and switching. Forwarding is the process of finding the best match for a destination address in a table to decide where to send packets. An example is the longest matching prefix algorithm used for IP forwarding. In contrast, switching uses a label taken from the packet as an index into a forwarding table. It is simpler and faster. These definitions are far from universal, however.

Since most hosts and routers do not understand MPLS, we should also ask when and how the labels are attached to packets. This happens when an IP packet reaches the edge of an MPLS network. The LER (Label Edge Router) inspects the destination IP address and other fields to see which MPLS path the packet should follow, and puts the right label on the front of the packet. Within the MPLS network, this label is used to forward the packet. At the other edge of the MPLS network, the label has served its purpose and is removed, revealing the IP packet again for the next network. This process is shown in Fig. 5-63. One difference from traditional virtual circuits is the level of aggregation. It is certainly possible for each flow to have its own set of labels through the MPLS network. However, it is more common for routers to group multiple flows that end at a particular router or LAN and use a single label for them. The flows that are grouped together under a single label are said to belong to the same FEC (Forwarding Equivalence Class). This class covers not only where the packets are going, but also their service class (in the differentiated services sense) because all the packets are treated the same way for forwarding purposes.

With traditional virtual-circuit routing, it is not possible to group several distinct paths with different endpoints onto the same virtual-circuit identifier because there would be no way to distinguish them at the final destination. With MPLS,

the packets still contain their final destination address, in addition to the label. At the end of the labeled route, the label header can be removed and forwarding can continue the usual way, using the network layer destination address.

Actually, MPLS goes even further. It can operate at multiple levels at once by adding more than one label to the front of a packet. For example, suppose that there are many packets that already have different labels (because we want to treat the packets differently somewhere in the network) that should follow a common path to some destination. Instead of setting up many label switching paths, one for each of the different labels, we can set up a single path. When the already-labeled packets reach the start of this path, another label is added to the front. This is called a stack of labels. The outermost label guides the packets along the path. It is removed at the end of the path, and the labels revealed, if any, are used to forward the packet further. The S bit in Fig. 5-62 allows a router removing a label to know if there are any additional labels left. It is set to 1 for the bottom label and 0 for all the other labels.

The final question we will ask is how the label forwarding tables are set up so that packets follow them. This is one area of major difference between MPLS and conventional virtual-circuit designs. In traditional virtual-circuit networks, when a user wants to establish a connection, a setup packet is launched into the network to create the path and make the forwarding table entries. MPLS does not involve users in the setup phase. Requiring users to do anything other than send a datagram would break too much existing Internet software.

Instead, the forwarding information is set up by protocols that are a combination of routing protocols and connection setup protocols. These control protocols are cleanly separated from label forwarding, which allows multiple, different control protocols to be used. One of the variants works like this. When a router is booted, it checks to see which routes it is the final destination for (e.g., which prefixes belong to its interfaces). It then creates one or more FECs for them, allocates a label for each one, and passes the labels to its neighbors. They, in turn, enter the labels in their forwarding tables and send new labels to their neighbors, until all the routers have acquired the path. Resources can also be reserved as the path is constructed to guarantee an appropriate quality of service. Other variants can set up different paths, such as traffic engineering paths that take unused capacity into account, and create paths on-demand to support service offerings such as quality of service.

Although the basic ideas behind MPLS are straightforward, the details are complicated, with many variations and use cases that are being actively developed. For more information, see Davie and Farrel (2008) and Davie and Rekhter (2000).

Frequently Asked Questions