Topology-Based (CEF) Switching

Matching a packet’s intended destination address with the IP routing table requires router CPU cycles. Enterprise routers can contain tens of thousands of entries and can match the same number of incoming packets against these entries. In an attempt to make this process as efficient as possible, Cisco has created various switching methods. The first is called process switching and it uses the route lookup and best match method already outlined. This method was improved upon with fast switching. A list of IP addresses of recently forwarded packets is generated by the router as well as the Data Link Layer headers that were copied if the IP address matched. Cisco Express Forwarding (CEF) was created as an improvement on fast switching. CEF runs on current models of Cisco routers by default.

Cisco Express Forwarding (CEF)

CEF operates at the data plane and is a topology-driven proprietary switching mechanism that creates a forwarding table that is tied to the routing table (i.e., the control plane). CEF was developed to eliminate the performance penalty experienced due to the first-packet process-switched lookup method used by flow-based switching. CEF eliminates this by allowing the route cache used by the hardware-based Layer 3 routing engine to contain all the necessary information to the Layer 3 switch in the hardware before any packets associated with a flow are even received. Information that is conventionally stored in a route cache is stored in two data structures for CEF switching. These data structures provide optimised lookup for efficient packet forwarding and are referred to as the FIB and the adjacency table.

NOTE: It is important to remember that even with CEF, whenever there are routing table changes, the CEF forwarding table is also updated. While new CEF entries are being created, packets are switched in a slower switching path, using process switching, for example. All current models of Cisco routers and current IOS use CEF.

Forwarding Information Base (FIB)

CEF uses a FIB to make IP destination prefix-based switching decisions. The FIB is conceptually similar to a routing table or information base. It maintains a mirror image of the forwarding information contained in the IP routing table. In other words, the FIB contains all IP prefixes from the routing table.

When routing or topology changes occur in the network, the IP routing table is updated, and those changes are also reflected in the FIB. The FIB maintains next-hop address information based on the information in the IP routing table. Because there is a one-to-one correlation between FIB entries and routing table entries, the FIB contains all known routes and eliminates the need for route cache maintenance that is associated with switching paths, such as fast switching and optimum switching.

Additionally, because the FIB lookup table contains all known routes that exist in the routing table, it eliminates route cache maintenance and the fast-switching and process-switching forwarding scenarios. This allows CEF to switch traffic more efficiently than typical demand-caching schemes.

The Adjacency Table

The adjacency table was created to contain all connected next hops. An adjacent node is a node that is one hop away (i.e., directly connected). The adjacency table is populated as adjacencies are discovered. As soon as a neighbour becomes adjacent, a Data Link Layer header, called a MAC string or a MAC rewrite, which will be used to reach that neighbour, is created and stored in the table. On Ethernet segments, this header information is the destination MAC address, the source MAC address, and the EtherType, in that specific order.

As soon as a route is resolved, it points to an adjacent next hop. If an adjacency is found in the adjacency table, a pointer to the appropriate adjacency is cached in the FIB element. If multiple paths exist for the same destination, a pointer to each adjacency is added to the load-sharing structure, which allows for load balancing. When prefixes are added to the FIB, prefixes that require exception handling are cached with special adjacencies.

Accelerated and Distributed CEF

By default, all CEF-based Cisco Catalyst switches use a central Layer 3 switching engine where a single processor makes all Layer 3 switching decisions for traffic received on all ports in the switch. Even though the Layer 3 switching engines used in Cisco Catalyst switches provide high performance, in some networks, having a single Layer 3 switching engine do all the Layer 3 switching does not provide sufficient performance. To address this issue, Cisco Catalyst 6500 series switches allow for CEF optimisation through the use of specialised forwarding hardware. This is performed using either Accelerated CEF (aCEF) or Distributed CEF (dCEF).

Accelerated CEF allows a portion of the FIB to be distributed to capable line card modules in the Catalyst 6500 switch. This allows the forwarding decision to be made on the local line card using the locally stored scaled-down CEF table. In the event that FIB entries are not found in the cache, requests are sent to the Layer 3 engine for more FIB information.

Distributed CEF refers to the use of multiple CEF tables distributed across multiple line cards installed in the chassis. When using dCEF, the Layer 3 engine (MSFC) maintains the routing table and generates the FIB, which is then dynamically downloaded in full to each of the line cards, allowing for multiple Layer 3 data plane operations to be performed simultaneously.

In summation, dCEF and aCEF are technologies that implement multiple Layer 3 switching engines so that simultaneous Layer 3 switching operations can occur in parallel, boosting overall system performance. CEF technology offers the following benefits:

• Improved performance – CEF is less CPU-intensive than fast-switching route caching. More CPU processing power can be dedicated to Layer 3 services, such as Quality of Service (QoS) and encryption, for example.
• Scalability – CEF offers full switching capacity at each line card in high-end platforms, such as the Catalyst 6500 series switches, when dCEF mode is active.
• Resilience – CEF offers an unprecedented level of switching consistency and stability in large dynamic networks. In dynamic networks, fast-switching cache entries are frequently invalidated due to routing changes. These changes can cause traffic to be process-switched using the routing table rather than fast-switched using the route cache.

Configuring Cisco Express Forwarding

Enabling CEF requires the use of a single command, which is the ip cef [distributed] global configuration command. The [distributed] keyword is only applicable to high-end switches, such as the Catalyst 6500 series switches, that support dCEF. The following output shows how to configure CEF on a lower-end platform, such as the Catalyst 3750 series switch:

The following output illustrates how to enable dCEF on the Catalyst 6500 series switches:

NOTE: There is no explicit command to configure or enable aCEF.