ISIS

[Ratheesh mv]

When routing the packet within an area, the routers use the System-ID of the router for delivery. When routing packets between areas, the Area-ID is the address. These addresses are part of the NET Value.

I have just started learning about IS-IS protocols. I am bit confuse on above sentence.

Can anyone explain about it or share any good material to understand above mentioned concept ?

[balaji.bandi]

ISIS is used to be rarely used case i in the enterprise network (Maybe mostly used Service Provider Environment). Engineers most prefer to have EIGRP/OSPF / i(e)BGP in most cases i have come across.

SD-Access (DNAC) getting more popular in the market and the new Cisco world - ISIS used here for a better flat network.

This is one of the best presentations i come across to help you most use case described.

 https://www.ciscolive.com/c/dam/r/ciscolive/us/docs/2019/pdf/BRKRST-2315.pdf

[Georg Pauwen]

Hello,

maybe an IS-IS address example is the best way to explain the concept. I hope that makes sense.

Take the below IS-IS address:

49.0001.1131.7700.1009.00

The first three bytes (49.0001) are the area ID.

49, the first byte, is the address family identifier (AFI) of the authority, which is equivalent to the IP address space that is assigned to an autonomous system.
The AFI value 49 is what IS-IS uses for private addressing, which is the equivalent of RFC 1918 address space for IP protocols.

0001, the second two bytes of the area ID, is the IS-IS area number, 1 in this example.

--> When routing packets between areas, this area ID is used for addressing.

The next six bytes, 1131.7700.1009, are called the system identifier, and these are used to identify the the router node on the network. The system identifier is equivalent to the host or address portion on an IP address.

-> When routing packets within an area, routers use the System-ID of the router for delivery.

The last two bytes are the NET selector (NSEL). For IS-IS, this value is always 00, indicating “this system.”

[Richard Burts]

To add a little to the excellent explanations - when forwarding to another area in ISIS only the area bits are examined in making the forwarding decision (it is forwarding toward the area and not the specific host). The forwarding decision does not examine or consider the system identifier until the packet has reached the edge of the destination area. I would suggest that this is somewhat similar to forwarding decisions in IPv4. When making a forwarding decision in IPv4 the router considers the network/subnet bits of the address but does not consider the host bits of the address until the packet reaches the edge of the destination network.

[Giuseppe Larosa · Accepted solution]

Hello @Ratheesh mv ,

IS-IS had been introduced as link state routing protocol to route  OSI connectionless protocol CLNP   CLNS is the connectionless service that is equivalent to IPv4.

All the sentences referring to how IS-IS routes within an area or between areas refers to CLNS routing.

In CLNS a single address is given to the whole node instead of having different addresses for diffferent interfaces.

A CLNS address like the net is is formed by area : system-id   and SEL with SEL being made of 0x00 in a NET the SEL can be seen a roughly equivalent to a TCP/UDP port it identifies a service at upper layer.

So a CLNS end node or a CLNS intermediate node like a router can be compared to a router with a loopback address IPv4 and unnumbered interfaces referring to the loopback interface.

ISIS L1 routing is done by system-id looking for the IS-IS node that lists that system-id as directly connected to it.

A different protocol called ES-IS is spoken between normal hosts and IS to make IS aware of what hosts / system-ids are connected to their interfaces. Then the IS system creates the L1 LSP that is a data structure that contains among the others the lists of adjacent ISs and adjacent End Systems and their system-IDs.

ISIS L2 routing is done based on area id that is the leftmost part of address. Once reached the destination area L1 routing happens again as explained above.

To be noted the CLNS address is variable in size from a minimum of 7 bytes  to a max of 20 bytes and this of course caused implementation issues as it could not be handled in hardware in an efficient way.

Later IS-IS became able to route IPv4  and then IPv6.

However, from the point of view of IS-IS IPv4 prefixes are leafs connected to an IS-IS node.

All the explanations that introduce IS-IS refers to the context of CLNS routing.

Hope to help

Giuseppe