ISIS Routing Protocol Note - Intermediate Systems To Intermediate Systems
IS-IS is a link-state interior gateway protocol similar to OSPF. It is commonly used in service provider core networks due to its simple flat network design and high scalability. IS-IS forms both level 1 and level 2 adjacencies by default and uses two levels to separate intra-area and inter-area routing. Route leaking allows selective distribution of routes between levels to influence traffic engineering.
ISIS Routing Protocol Note - Intermediate Systems To Intermediate Systems
IS-IS is a link-state interior gateway protocol similar to OSPF. It is commonly used in service provider core networks due to its simple flat network design and high scalability. IS-IS forms both level 1 and level 2 adjacencies by default and uses two levels to separate intra-area and inter-area routing. Route leaking allows selective distribution of routes between levels to influence traffic engineering.
Link state IGP similiar to OSPF, OSPF for IP and ISIS for CLNS.
then developed to support IPv4 and IPv6.
Used in SP Core, less feature compared to EIGRP and OSPF but most suitable with SP due to : simply flat network design Carriers and ISPs run BGP-based networks, and their IGP is used mainly for finding BGP session endpoints. So they want their IGP to be as simple as possibleoften making their entire routing domain a single IGP area. IS-IS is certainly simpler than OSPF in many aspects, and many will argue that it is more scalable in this type of single large area application. So when you encounter IS-IS in a service provider network, what you will usually encounter is a single L2 area. Jeff Doyle (Routing TCP IP Volume 1) highly scalable Remaining lifetime, is an age of LSP. IS-IS Remaining Lifetime begins at MaxAge and decrement to Zerowhereas OSPF LSAs Age increment from zero to MaxAge. default IS-IS Maxages is 1200s (20Mnutes), and IS-IS refresh interval is 15 minutes.because IS-IS remaining lifetimes begins with non zero number and is decrement to zero. the beginning number can be set from its default 1200s (20minutes) , to maximum 65535 second (18.2Hours). this is major contributing factor to IS-IS scalability in very large areas. If the area links are reasonably stable, increasing the refresh time and the Remaining Lifetime can sharply reduce the flooding load caused by LSP refreshes. Jeff Doyle
support ipv4 and ipv6 routing
Enabling ISIS on IOS :
-enable process globally {router isis [process id]} -define NET address {net [nsap]} -enable in interface process {ip router isis [process id]} Enabling ISIS on IOS XR : -enable process globally {router isis [process id]} -define NET address {net [nsap]} -enable interfaces under IS-IS process {interface gig0/1/0/4 address-family ipv4 unicast}
NET (Network Entity Title)
isis router id use ISO NSAP addressing format, minimum 8 bytes maximum 20bytes NET Format : AA.AAAA.AAAA.AAAA.AAAA.AAAA.AAAA.SSSS.SSSS.SSSS.NN Area not like ospf area System ID Router ID Inside the Area N-Selector always zero. 49.xxxx (49) in NSAP is like private address
IS-IS Adjacency Levels
-ISIS uses two levels of adjacency Level 1 (L1) Level 2 (L2) -Process and Interface default to Level-1-2 Forms both L1 and L2 Separate LSP Database, Double the overhead -Commonly in Service provider core, we only use single level of adjacency or (ospf single area) because mpls ldp and traffic engineering limitations L2 Adjacency -Inter or Inra area adjacency -Like area 0 in OSPF -Must be Contiguous L1 Adjacecny -Intra area adjacency only -Like OSPF Not So Stubby Area Intra area routes Default route out redistribution allowed Router Roles
L1-L2 Routing -Like ABR in OSPF used as exit point from L1 to L2 -Inject default route to level-1 Sets the attached bit (only if its connected to different area, will not advertise default route even its L1L2 router if all of its interface are in same area) Level Manipulation used to utilize isis database global under isis process, affect all interfaces is-type under interface configuration, affect only that interfaces isis circuit-type ISIS Network type, path selection and Route leaking ISIS have only 2 network type : broadcast (ethernet) default on multipoint interfaaces uses DIS instead of DR/BDR can be seen from show isis neighbor the circuit id indicated the DIS on the segment LSP are multicast point to point default on point to point interfaces change network type isis network point-to-point point to point uses (serial IIH) and broadcast uses (L2/L1 LAN IIH), incompatible with one another. DIS election like ospf dr/bdr no backup DIS election is dynamic, preemption can occur separate elction for L1 and L2 occur by : highest prioirity (default 64) highest snpa (mac address) best practice any segment that only have 2 routers, we can just change the network to only point to point DIS only needed on segment with 3 or more router ISIS path selection isis doesnt use bandwidth metric as ospf does, all link defualt to cost 10 can be manually modified neighbor must agree on metric style control how the actual metric is encoded on the updates / TLV (type link value) isis support 2 different metric style : narrow (default), metric 1-63 wide (support more numerical value of cost, and also needed for enhance application like mpls te), metric 1-16777214 incompatible with one another, the adjacency is up but the routes will not installed on routing table level 1 paths are preffered over level 2 paths, like ospf intra area routes that preffered over inter area routes IS-IS route leaking selectively take level-2 routes and pass the into level-1 or filter level-1 routes as they move into level-2 level 2 domain know all prefixes level 1 domain only know L1 prefixes route leaking used for traffic engineering based on longest match prefix (i ia) configuration IOS : redistribute isis ip level-2 into level-1 (distribute-list/route-map) IOS XR :