Starting today, I would like to introduce you to the content related to the IS-IS protocol. This article mainly covers the IS-IS network model, history, router level, region division, and IS-IS address.
1. Introduction to IS-IS
The so-called is-is intermediate system-to-intermediate system, which means intermediate system to intermediate system.
As we all know, the TCP IP network model and the OSI model are widely available at the moment. In the field of computer networks, most of the routing protocols and functions are based on the TCP IP network model, but the IS-IS protocol is a protocol based on the OSI network model, so it is slightly different from other routing protocols in the implementation of various functions.
In the TCP IP network model, we call the routing device that implements network-layer communication routers, and in OSI, it is called the intermediate system, which is the origin of the intermediate system in the IS-IS protocol.
Similar to the function of OSPF, IS-IS is also a dynamic routing protocol, which has many similarities and differences with OSPF, so it is recommended to learn Xi OSPF before Xi learning the IS-IS protocol, so that the learning Xi of IS-IS can be smoother.
IS-IS was originally based on the OSI7 layer network model and supported CLNP (Connectionless Network Protocol, equivalent to the network layer protocol of OSI), but because the OSI7 layer network model was not adopted on a large scale, the IS-IS protocol was not used on a large scale. However, since IS-IS is a Layer 3-like network protocol on top of the data link layer, and IS-IS is more flexible, the IS-IS engineering team extended IS-IS to support both IPv4 and IPv6. Therefore, the IS-IS protocol, which integrates IPv4 and IPv6 network functions, has been widely used in the live network. In addition, IS-IS has the advantages of strong scalability, fast convergence speed, and simple working mechanism, so it is deployed by ISPs in large numbers.
At present, in the enterprise network environment, the IS-IS protocol is rarely used, and most of the OSPF protocols are deployed, but in the ISP operator network, the IS-IS protocol and the OSPF protocol are in an equal position.
2. IS-IS address
NSAP (Network Service Access Point) is the network layer address specified in the OSI model, similar to the IP address in the TCP IP model.
The NSAP address structure is shown in the following figure:
NSAP consists of IDP (Initial Domain Part) and DSP (Damian Specific Part). The IDP and DSP lengths are variable, with a maximum of 20 bytes and a minimum of 8 bytes at the NSAP main plant. In fact, the maximum length of an NSAP address is longer than that of an IPv6 address, so it supports more addresses than an IPv6 address.
The IDP is equivalent to the main network number in the IP address, which is specified by the ISO and consists of AFI (Area and Format Identifier) and IDI (Initial Domain Identifier), AFI represents the address format of the address allocation authority, and IDI is used to identify the domain.
DSP is equivalent to the subnet and host address in the IP address, and consists of three parts: high order DSP, system ID, and SE. High Order DSP is used to distinguish regions, which is equivalent to the subnetwork number in the IP address, System ID is used to distinguish hosts, which is equivalent to the host number in IPv4, and SEL is used to indicate the service type, similar to the protocol number in the IP protocol, which is used to distinguish the upper-layer protocol type.
The AFI and ADI fields of the IDP, together with the High Order DSP field of the DSP, form the network number field of the NSAP, also known as the area ID of IS-IS. (Note that there are two concepts of region in IS-IS, in addition to the area here, there are also the concepts of backbone area and non-backbone area, which will be discussed later in this article) The length of the area ID is 1-13b.
In fact, since the SYTEM ID and SEL lengths are fixed at 6b and 1b. Therefore, when calculating the area id, only the 7b from the bottom is required, and the remaining address is the area id.
When configuring IS-IS, although you do not need to configure an NSAP address on each interface of the router, you do need to configure a special NSAP address, the .NET address, for each router running the IS-IS protocol. A .NET address is called a Network Entity Identifier and is a special type of NSAP address that has an all-0 SEL portion.
3. IS-IS router level and region
There are three levels of routers running the IS-IS protocol, which are Level 1, Level 2, and Level 1 2. The IS-IS protocol can establish a Level 1 and Level 2 neighbors between two routers. (abbreviated as L1, L2 and L1 2 in later descriptions).
The IS-IS protocol states:
L1 routers can only establish L1 neighbors
An L2 router can only establish L2 neighbors
The L1 2 router establishes an L1 neighbor with the L1 router in the same region, and establishes an L2 neighbor with the L2 router (whether it is in the same region or not). Establish L1 neighbors with L1 2 routers in the same region, and L1 and L2 neighbors with L1 2 routers in different regions.
In summary, the corresponding level of router can establish the corresponding level of neighbors, but the establishment of L1 level neighbors requires that the two routers be in the same zone, while the establishment of L2 level neighbors does not require the two routers to be in the same zone.
Based on the IS-IS router neighbor relationship, IS-IS can be divided into two regions—the backbone region and the non-backbone region. The area formed by the IS-IS neighbors of L2 is the backbone area, and the area formed by the IS-IS neighbors of L1 is the non-backbone area. The boundaries between the backbone and non-backbone areas are on Level 1-2 routers. Since IS-IS does not have a VLINK mechanism similar to OSPF, the backbone region of IS-IS must be contiguous.