A Glossary of Standards and Protocols Relevant to Redundant Networks

You now know how to design a network with redundant components. Let's take a look at the building blocks of such networks: the standards and protocols.

Border Gateway Protocol 4. BGP-4 supports interdomain routing (aka inter-autonomous system routing). Border Gateway Protocol (BGP) routers exchange network-accessibility information with other BGP routers. A primary BGP feature is its identification of networks by their Autonomous System Number (ASN). An autonomous system (AS) is a logically contiguous network under one administrative structure that uses one routing policy. You can also use BGP to exchange routing information between routers within an AS. This variation of BGP is called the Interior Border Gateway Protocol (IBGP). Similarly, you can use BGP to exchange routing information between ASs. This variation of BGP is called the Exterior Border Gateway Protocol (EBGP).

Classless Inter-Domain Routing. CIDR lets you specify network addresses more flexibly—and with less wasted network address space and route table space—than you do with the IP address class concept (e.g., Class A, B, C, and D address spaces).

Fast EtherChannel. A proprietary Cisco Systems link-aggregation (aka port-trunking) scheme, Fast EtherChannel permits multiple parallel active links between two devices. Link aggregation lets you scale total transmission bandwidth between the two devices and load-balance the links. An advantage of Fast EtherChannel is rapid reconvergence (less than a second) when an active link fails.

Hot Standby Routing Protocol. HSRP is Cisco's proprietary implementation of Virtual Router Redundancy Protocol (VRRP). Like VRRP, HSRP lets you configure a fully redundant outbound communication path.

IEEE 802.1D Spanning Tree Algorithm. You can use multiple parallel connections to interconnect switches that support Spanning Tree Algorithm. STA disables all but one connection and reenables an alternative connection when the active link fails.

IEEE 802.3ad Link Aggregation Control Protocol. The Link Aggregation Control Protocol (LACP) standard, which grew out of Fast EtherChannel, offers the advantages of link aggregation in a multivendor environment. Two modes are available: Static and Dynamic. In the common Static Mode, multiple adapters attach to the same switch (with STA off). Static Mode can also provide fault tolerance in the event of a switch failure, as in the following configuration: Exactly two NICs connect to two switches (with STA on)—one NIC is active until the link fails, at which point the other NIC takes over. In Dynamic Mode, multiple adapters in a server are split between two teams, each team connecting to a different switch. One team is active at a time, offering the benefits of scalable bandwidth and load balancing. Should the active switch fail, traffic fails over to the team on the other switch.

Internet Router Discovery Protocol. Described in Internet Engineering Task Force (IETF) Request for Comments (RFC) 1256, Internet Router Discovery Protocol (IRDP) provides a method for routers to periodically advertise their presence and for clients to request an immediate advertisement (typically at startup). IRDP lets client systems dynamically locate default gateways.

Virtual Router Redundancy Protocol. VRRP works in conjunction with BGP to achieve network fault tolerance in the event of a router failure. VRRP introduces the concepts of a virtual router ID and a virtual router IP address that fail over to (i.e., are reassigned to) a backup router when the primary router fails. You can configure a physical router with more than one virtual router ID, and you can configure several physical routers with the same virtual router ID. For each virtual router ID, only one physical router is active at a time. When it fails, another router with the same virtual router ID assumes the active role.