The MPLS PE (Provider Edge) router or network element plays a key role in MPLS infrastructure. The PE router is both the interface between the customer-facing network and the MPLS core, and the point where customer data is given an MPLS label and/or the label is removed. Data enters the MPLS network through the PE router, traverses the network, and then exits through another PE router.
“After receiving customer data, the PE must be able to impose a unique label to the traffic source and forward such information across an MPLS cloud to a remote destination; at the other end of the cloud, another PE will receive such information, remove the MPLS labeling and deliver data back to the customer,” said Jose Santos, lab director of the University of Colorado at Boulder’s Telecommunications Program.
A single MPLS PE router can be shared with more than one customer and is a cost effective way to aggregate traffic, according to Santos. Because of that feature, he added, the MPLS PE router is responsible for maintaining distinctions between the information received from every customer which interconnects with the MPLS cloud through that router.
According to Santos, MPLS PE routers can function in multiple ways, depending on the MPLS service sold to the customer:
- In case of MPLS IP VPNs, the PE router will maintain different routing tables for each one of its customers using features such as VRFs (Virtual Routing and Forwarding); and will permit the exchange of IP packets between local and remote destinations.
- Services such as VPLS and Pseudo wires (PW) will have comparable forwarding information on a per-customer and/or per-service basis. In this case, the PE router will permit the exchange of Ethernet frames, VLAN trunks and/or emulated circuits between locations.
- In “flat” enterprise MPLS networks (no VPN service), the PE router will impose labels to permit label lookup in the core instead of IP routing table lookup.
For MPLS provisioning inside the provider network, Santos said, the MPLS PE router will need:
- A label distribution protocol to signal and exchange information with other MPLS devices (P routers and PE endpoints); some of the options available are LDP, TLDP & RSVP-TE. BGP is also used for label distribution in case of L3VPNs.
- MPLS forwarding capability: Label information stored in components such as Label Information Base (LIB) and Label Forwarding Information Base (LFIB) , which will tell the router what labels to use to encapsulate customer data, as well as what interfaces to forward this information to.
- IP knowledge (a routing table, a routing protocol ( OSPF, BGP in case of L3VPNs) & IP addresses on its ports). It is important to highlight that Label distribution protocols operate on top of IP, Santos said, so the PE has to run a routing protocol and be able to reach via IP all other MPLS nodes, before using an LDP and then enabling MPLS services.
MPLS PE: Putting It All Together
An MPLS PE router will receive customer data and perform label imposition in single (one label) and stack form (more than one label), depending on the service offered to the customer.
How will the PE router know what labels to use? Santos said that a PE router is responsible for generating the labels for the customer networks where they terminate (see MPLS tutorial); they will exchange such labels with other PE’s (customer reach-ability information) and P routers (core reach-ability information) using a label distribution protocol. That label information is stored in the LIB and LFIB tables. Each time the PE router receives data from a customer, it refers to those tables to determine the label which should be imposed and the forwarding path for that labeled packet to reach its destination. At the end of this process, Santos said, the MPLS PE router will know how to reach other remote destinations via a labeled path to another PE.
