Confidential Information of Huawei.

No Spreading Without Permission

Confidential Information of Huawei. No Spreading Without Permission
Confidential Information of Huawei. No Spreading Without Permission
Confidential Information of Huawei. No Spreading Without Permission
Confidential Information of Huawei. No Spreading Without Permission
 IS-IS protocol can be used between routers to exchange the routing information.
Initially, IS-IS protocol is described in ISO10589 and it can only support CLNP
network. Later, RFC 1195 has defined Integrated IS-IS. Integrated IS-IS can be
used in a pure CLNP network, a pure IP network or a dual CLNP/IP network. This
implies that integrated IS-IS protocol can be used to exchange the CLNP and IP
routing information between the routers at the same time.
IS-IS protocol means Integrated IS-IS protocol in this lesson unless it is
specifically mentioned.

Confidential Information of Huawei. No Spreading Without Permission

IS-IS is configured in the core layer of bearer network as an important IGP, to make
the routers access each other.
The figure in the slide is a typical IS-IS network. IS-IS is run on all the routers in core
layer.

Confidential Information of Huawei. No Spreading Without Permission

 Different from other routing protocols, ISIS protocol run directly on top of the
data link layer. In order to synchronize the link-state database, IS-IS exchange the
link information by sending the Protocol Data Unit (PDU).
（ ）
The PDU Protocol Data Unit consists of the following 3 parts ：
1. The header of the data link layer: OSI family identifier 0xFEFE is marked to
indicate that the encapsulated is OSI packet structure.
2. IS-IS fixed packet header: The first byte is 0x83, indicates IS-IS packet. ( If
ES-IS protocol is adopted, the first byte of the fixed packet header is 0x82; if CLNP
protocol is adopted, the first byte of the fixed packet header is 0x81).
：
3. IS-IS TLVs Its content might vary according to packet type and includes
many TLV structure.
TLV means TYPE, LENGTH, and VALUE. In fact, TLV is a variable length data
structure that consists of type, length and value attribute. The type attribute is 1
byte and it describes the information it contains. Length attribute describes the total
length of the TLV. The value attribute is variable and it indicates the content of the
TLV. The content of the value attribute varies according to the type of TLV.
The advantages of using TLV structure to build the packet are high flexibility and
extensibility. By using the TLV, the entire packet structure of the packet is fixed
except the TLV part. This is because one packet can be embedded with many TLV
structures. The new TLV structure can be added rather than modify the whole
packet structure in order to support new technologies. This is the key strength of
the IS-IS protocol design which lies in the ease of extension through the
introduction of new TLV rather than new packet types.
Thus, IS-IS can easily support new technologies such as TE, IPV6 and etc
through the used of TLV.

Confidential Information of Huawei. No Spreading Without Permission

 This is a structure of the IS-IS Hello packet. The header of the IS-IS Hello packet
is a standard Ethernet header. It comprises of destination MAC address, source
MAC address, and the length of the field. At LLC field, DSAP and SSAP are both
marked by 0xFE. These indicate that it is an OSI data packet. The next part consists
of the IS-IS data, and the first byte is indicated by 0x83. The data portion of the IS-IS
consists of the fixed length header field and variable length TLV field.

Confidential Information of Huawei. No Spreading Without Permission

 ISIS protocol is a type of IGP routing protocol, its ultimate goal is to calculate the
routes. There are 3 working proccesses to achieve this goal:
1. Establish the adjacency relationship
Discover the neighbor and establish the adjacency relationship. In this way, the
link state information can be exchanged. Thus, adjacency relationship
establishment is the basic of IS-IS working proccess.
2.Synchronize the link state database
IS-IS is a link state routing protocol used to calculate the precise and integrated
route. Every router must be equipped with a set of accurate and identical LSDB. To
achieve this, the database must be synchonized first through the exchange of link
state information between the adjacent routers. The final result is that every router
is having a synchonized LSDB which is used to describe the whole topology of the
network
3.Calculate the route
Use the synchronized LSDB information to run SPF algorithm and calculate the
routes. The routing information is then used to direct the data transmission.
These 3 aspects of the IS-IS protocol will be discussed in the coming slides in
detail.

Confidential Information of Huawei. No Spreading Without Permission

Confidential Information of Huawei. No Spreading Without Permission
Confidential Information of Huawei. No Spreading Without Permission
 The prior to establish adjacencies relationship is to know the existence of each
other. Therefore, every IS must be identified first by using unique System ID.
The type of adjacency relationship to be established is dependent on the IS type
and the area that the IS resides. The area that the IS belongs to is identified by
using the AREA ID Identifier.
Next, we will discuss the concept related to the adjacency relationship
establishment.

Confidential Information of Huawei. No Spreading Without Permission

 CLNP is the layer 3 network layer protocol of the OSI protocol stack. Its NSAP
addresses are not fixed in length and can be up to 20 bytes long (160 bits).
The NSAP format consists of 2 components namely IDP (Initial Domain Part)
and DSP (Domain-specific Part). The IDP is similar to the primary network No. in
the TCP/IP address. The DSP is similar to the subnet No., host No. and port No. in
the TCP/IP address.
The IDP is further broken down into the AFI (Authority and Format Identifier) and
the IDI (Initial Domain Identifier). AFI is used to identify the address format and
address assignment authority. IDI is used to identify a domain. DSP Domain （
）
Specific Part consists of the HODSP (High Order DSP), System ID and NSEL.
The HODSP is used to partition areas, and is functionally similar to the subnet No.
in the TCP/IP address. The System ID is used to differentiate the host or router,
and is functionally similar to the host No. in the TCP/IP address. The NSEL is used
to indicate the selected service, and is functionally similar to the port No. in the
TCP/IP address.
The area ID consists of the IDP and the HODSP of DSP. The Area ID field has
variable length ranging from 1 byte to 13 bytes.
A SYSTEM ID is used to uniquely identify a host or a router in an area. The
length of the system ID is specified to be 1 to 8 bytes. The system ID of all nodes in
an IS-IS routing domain must be of the same length. Thus, VRP have set the
SYSTEM ID to 6 bytes. During the configuration, we can use the MAC address as
system ID. Besides, we can also convert the IP address from 4 bytes to 6 bytes so
that it can be used as SYSTEM ID. Usually, the Router ID is used for conversion.
Suppose a router takes the IP address 168.10.1.1 of the interface Loopback0 as
router ID. Then its system ID used in IS-IS can be transformed with the following
method:

Confidential Information of Huawei. No Spreading Without Permission

 1. Extend every part of the IP address 168.10.1.1 to three bits. Add 0 to the front of
the part that includes less than three bits.
2. Divide the extended address 168.010.001.001 into three parts, with each part
consisting of four decimal digits. The reconstructed 1680.1000.1001 is just the system
ID.
Actually, there are many ways to designate a system ID, as long as it can uniquely
identify an ES or a router.
NSEL is 1 byte long and its value helps the network layer in handling off the
datagram to the appropriate application or service user. NET is a special NSAP
address with N-SEL 0. Current VRP version supports at most 3 NETs.
Some requirements when configuring NET ：
1. An IS-IS router must have at least one NET configured (In fact, current VRP
version can support up to 3 NETs per router).
2. IS-IS routers that belong to the same area must have same AREA ID.
3. Each router in an IS-IS routing domain must have a unique system ID.
4. 2 Level-2 IS in one routing domain cannot have the same System ID.
5. The System ID of all nodes in an IS-IS routing domain must be of the same
length.
6. All the NETs of one IS must be having the same System ID.
7. Consider that the 6-byte length of the System ID field and the 1-byte NSEL field,
and the Area ID may vary between 1 and 13 bytes. Because only 1 byte is sufficient to
define the Area ID, the smallest length of an NET on a Huawei router is 8 bytes. The
largest length of an NET is therefore 20 bytes.

Confidential Information of Huawei. No Spreading Without Permission

Confidential Information of Huawei. No Spreading Without Permission
 IS-IS adopted the area division concept in its design. The advantages of area
division to the entire network are: reduce the number of the routing entries in the
router, make the route aggregation an easier task, and make the network easier to
expand and etc.
In fact, configuring a Network Entity Title (NET) means assign a system ID to the
router and defines the IS-IS area address. The area address is identified through
the AREA ID identifier of NET. IS-IS routers that are having the same area ID are
considered to be in the same area.
One IS-IS router can be configured with at most 3 NETs at the same time. The
Area ID of these 3 NETs are different but the System ID are the same. Configuring
3 NETs in a single IS-IS router is useful for merging areas, splitting area and
renumbering.
In the following slides, we will see that the type of adjacencies relationship to be
established is depending on the area that the routers belong to. In addition, we can
reduce the number of routing entries of some routers in the area via setting the
router type.

Confidential Information of Huawei. No Spreading Without Permission

 The size of the link state database increases according to network scale. It is a
great challenge to the routers which has low processing capability if all the routers
are required to maintain the same copies of LSDB information with regarding to the
whole routing domain. This is because the routers must be equipped with sufficient
space to store the huge database. Besides, it takes longer time to calculate route by
using the information in the large database. This can directly cause the network
takes longer time to converge.
On the other hand, any changes in the network will cause LSP flooding. The
increase of the network scale might bring the network to the instability state. This is
because any changes in the network might cause LSP flooding to every router in the
whole routing domain. This flooding process is therefore affecting every single router
in the network.
To solve these problems, we can use area division and hierarchical design
method. By using area division, the whole routing domain can be divided into many
smaller segments known as areas. As a result, the routers with low processing
capability inside an area are required to know the routing information of the local
area only. In addition, the effect of changes inside an area is limited to that area only.
After the hierarchical design, all the internal area routers belonged to Level-1
while the routers connecting to another area belong to level-2. The internal area
routers (Level-1 router) can directly accessing each other provided that they are in
the same area. To route a packet to another area, an Level-1 router must forward
the packet to an L1/L2 router.

Confidential Information of Huawei. No Spreading Without Permission

 The whole routing domain is divided into 3 areas by configuring the NET. Leve-1
router contains the level-1 link state information (only contain the local link state
information) in its Level-1 LSDB. The level-1 router need not care the route of the
whole routing domain. Level-2 routers are used to connect the different areas
together. The LSDB of the level-2 router contain the link state information of the
whole routing domain. Therefore, the level-2 router can direct the data packet of
level-1 router toward the destination. Apart from that, there are Level-1-2 routers
which are responsible to exchange the routing information between the local area
and the backbone area. Level-1-2 routers can form adjacencies with both level-1
and level-2 routers. Thus, level-1-2 routers contain both level-1 LSDB and level-2
LSDB.
Level-1 routers do not contain the routing information for the destination outside
the area. Level-1 routers therefore route the traffic to destination outside of their area
to level-1-2 router. Level-1-2 router announces the routing information of the internal
area to level-2 router. This means that level-1-2 and level-2 routers work together to
transmit the data packets between different areas. The level-1-2 and level-2 routers
together form the backbone area. For example, RTA, RTB, RTC and RTD together
form the backbone area.
Level-1-2 router will set the ATT bit in the level-1 LSP when advertise the link
state information to the level-1 router in the native area. This is to inform the level-1
router about the existence of the level-2 capable router in the area. Then, the level-1
router will generate a default route point to this level-1-2 router. When the level-1
router wants to access the destination outside the area, it will use this default route
and send the data to the level-1-2 router. Since the level-1-2 router contains the
routing information of other areas, the data will be directed to the correct destination.

Confidential Information of Huawei. No Spreading Without Permission

 RTA and RTE are Level-1routers; RTB and RTD are Level-1-2 routers; RTC is
Level-2 router. RTA and RTB belong to area 49.0001; RTD and RTE belong to area
49.0002. RTA advertises network 1.1.1.1 into IS-IS while RTE advertises network
5.5.5.5 into IS-IS.
Enable IS-IS:
[RTA] isis
Configure network-entity:
[RTA-isis] network-entity 49.0001.0000.0000.1001.00
Configure IS level:
[RTA-isis] is-level level-1

Confidential Information of Huawei. No Spreading Without Permission

 The L1 router can only forms adjacency with the L1 routers and hosts in local area,
and has the information of the local area. The LSDB corresponding with L1 router is
Level-1 LSDB, storing the topology structures of various systems (routers and host)
in the local area. Level-1 router does not contain the routing information of other area,
so it cannot directly access the network in other areas. To access the network in
other areas, Level-1 router needs to send data packets to the nearest level-1-2 router.
Then, the level-1-2 router will be responsible to transmit the data packets. The
“nearest” level-1-2 router refers to the lowest cost toward level-1-2 router. The
nearest level-1-2 router might not necessarily be on the best path to a specific
destination outside the area.
Level 1-2 routers set the ATT bit in the Level 1 LSPs they advertise to the level-1
router on their native areas. This ATT bit provides a clue to level-1 only router about
the level-2 capable router in the area. This means the level-1 router can access the
network outside the area and outside the autonomous system via level-1-2 router.
This can be achieved by checking the ATT bit set by level-1-2 router. Level-1 router
might receive multiple LSPs with ATT bit set from different level-1-2 routers. In this
case, the level-1 router will select the nearest level-1-2 router as the exit point to the
network outside the area.

Confidential Information of Huawei. No Spreading Without Permission

 Level-2 routers form the backbone area of the IS-IS routing domain with other
level-2 or level-1-2 routers. Level-2 routers maintain the level-2 LSDB. Therefore,
level-2 routers contain the topology information of the backbone area. Level-2
routers which are connected to the backbone contain the topology information of the
whole routing domain even though they do not have any level-1 link state database.

Confidential Information of Huawei. No Spreading Without Permission

 The L1 router can only form adjacencies with L1 routers while L2 router can only
form adjacencies with L2 routers. The router on the area border functions as both L1
and L2 is responsible to exchange the routing information between the local area
and the backbone area. This type of router is called L12 router, which is located at
both level-1 and level-2. Similarly, the L12 router contains both Level 1 and Level 2
LSDBs. The L12 router is located at the area border and its responsibility is to
convert routing information of L1 LSDB to it of L2 LSDB for transmission in the
backbone area. The L12 router also instructs the L1 routers in local area on how to
access external network. The L12 routers on the area border instructs the L1 routers
in local area to select the nearest L12 router as the egress point to access external
network by setting the ATT bit. At the same time, the level-1 will generate a default
route to the nearest L12 router. The “nearest” L12 router refers to the lowest cost
toward L12 router. The nearest L12 router might not necessarily be on the best path
to a specific destination outside the area. Therefore, this method for level-1 router to
access the destination in other area or other AS may lead to the existence of less
optimal path. Following slides will further discuss the solution for this problem.

Confidential Information of Huawei. No Spreading Without Permission

Confidential Information of Huawei. No Spreading Without Permission
 The basic principle that must be followed when IS-IS routing protocol establish
the neighbor relationship via IIH packet:
1. Only the same level neighboring routers can form adjacency
2. The Area ID must be the same for Level-1 router
3. Network segment checking to ensure the network segment is the same. This
function is provided on Huawei routers. Theoretically, the same network segment
checking is not necessary since IS-IS is run directly on top of the data link layer.
This feature is turned on to ensure neighbor relationship is established accurately.
4. In OSPF protocol, different interfaces of the same router may belong to
different areas. Thus, OSPF routers can establish neighbor relationship with routers
from different area. However, in IS-IS protocol, one router must belong to the same
area. Thus, Level-1 IS-IS routers can only establish Level-1 neighbor relationship
and Level-2 IS-IS routers can only establish Level-2 neighbor relationship.

Confidential Information of Huawei. No Spreading Without Permission

 By default, the type of the IS-IS router is level-1-2. The level-1-2 IS has been
equipped with the capability to establish level-1 adjacency and level-2 adjacency
(Of course, there are some criteria that must be satisfied and we will discuss it
later). By default, the interface circuit-level type is Level-1-2 for the Level-1-2 IS-IS
router.
Level-1 IS can form only Level 1 adjacencies on all interface on which IS-IS is
enabled. Similarly, Level-2 IS can form only Level 2 adjacencies with neighboring
IS.
For Level-1-2 IS, we can modify its interface circuit-level type. For example, we
can change the interface circuit-level to Level-1 so that only Level-1 adjacency can
be formed between the connected IS and this interface.
Please note that the change of interface circuit level is valid only for broadcast
network with the router type as Level-1-2. On the point-to-point links, the Level-1
and Level-2 routers use the same protocol packets. Thus, this configuration makes
no sense.

Confidential Information of Huawei. No Spreading Without Permission

 For the IS-IS routers belong to the same area (Area ID same), Level-1 and
Level-2 adjacencies can be established. Level-1 router can form only Level-1
adjacency. Level-2 router can form only Level-2 adjacency. Level-1-2 router can
form Level-1 and Level-2 adjacency.
A router can form both Level 1 and Level 2 adjacencies with neighbors in the
same area ( Area ID same) and form only Level 2 adjacencies with neighbors in
different areas.

Confidential Information of Huawei. No Spreading Without Permission

 Routers belong to different areas can form only Level-2 adjacency. If two of the
Level-1-2 routers belong to different area, only Level-2 adjacencies can be
established.

Confidential Information of Huawei. No Spreading Without Permission

 Adjacency relationship establishment mechanism varies according to the
network types. IS-IS routing protocol only supports 2 network types namely Point
to Point network and Broadcast Network.
Theoretically, IS-IS cannot support NBMA network and point to Multipoint
network unless we configure the NBMA network and Point to Multipoint network to
multiple Point to Point networks. This can be done by creating a sub-interface.
The network types that OSPF routing protocol support are Point to Point,
Broadcast network, NBMA, Point to Multipoint, and Virtual Link.
Although the number of network types supported by IS-IS are less compared to
OSPF protocol, this doesn’t mean that IS-IS cannot adapt to multiple network types.
In fact, implementation of IS-IS is simpler compared to OSPF routing protocol.

Confidential Information of Huawei. No Spreading Without Permission

 There are 3 kinds of IS-IS Hello packets namely L1 LAN IIH, L2 LAN IIH and
point-to-point IIH. IS-IS uses multicast address to send LAN IIH. The multicast
address for L1 LAN IIH is 0180-C200-0014 and the multicast address for L2 LAN
IIH is 0180-C200-0015. At point to point link, point-to-point IIH is sent by means of
unicast address.
At the broadcast network, the IIH packets that are used to establish the Level-1
adjacency relationship and Level-2 adjacency relationship are different. We can
configure the type of interface circuit-level by using the IS-IS circuit-level [ level-1 |
level-1-2 | level-2 ] command at the interface view to indicate the IIH packet to be
sent.
By default, the Level 1 LAN IS-IS Hello and Level 2 LAN IS-IS Hello are sent out
simultaneously in a broadcast network for Level-1-2 IS with the Level-1-2 interface
circuit-level. In this case, we can configure the type of interface circuit-level to allow
the sending of only one type of the IIH. For example, we can allow the sending of
either L1 LAN IS-IS Hello or L2 LAN IS-IS Hello. This can indirectly limit the
adjacency type that can be established in this interface.
The functions of HELLO packets ：
1. Discover the neighbor
2. Negotiate the parameters and establish the adjacency relationship
3. Acts as the Keepalive packet after the adjacency has been established
The format of LAN IIH and Point-to-Point IIH packets are different, and this
determines that the adjacency relationship establishment process on LAN and
point-to-point link are different.

Confidential Information of Huawei. No Spreading Without Permission

 Adjacency relationship on Point-to-Point links are initialized by the receipt of ISH
packets through the ES-IS protocol. ISH packet is HELLO packet sent by IS, this
type of packet is defined in ES-IS protocol and we will not discuss it in detail here.
When an ISH is received on a newly enabled point-to-point link, the router
verifies whether an adjacency already exists with the sender by checking the
source SysID in the ISH against its adjacency database. The ISH is ignored if an
adjacency exists. If not, the receiving router creates a new adjacency and sets its
state to "initializing" and the system type to "unknown“. The router then sends the
new neighbor an IIH in response. Upon receiving a subsequent IIH from the new
neighbor, the router will check the parameters inside the packet. If the checking is
passed, the router then moves the adjacency to an ‘up’ state and changes the
neighbor's system type to IS. Thus, Point to point adjacency relationship formation
is a “ 2 way handshake process”. In this process, the local router is unable to
determine whether its hellos reach the remote end. This might lead to situations
where one end of an adjacency is up but the other end is not. Later, RFC has
defined a more reliable way to form point-to-point IS-IS adjacencies - by using a
three-way handshake process. The three-way handshake process for reliably
forming point-to-point adjacencies introduces a new type length value field (Type
240), known as Point-to-Point Adjacency State TLV. It records the remote ID in
Point to Point IIH packet.
The IIH packet parameters that are required to be checked are as follows:
1. Maximum Area Addresses: The default value 0 implies a maximum of three
addresses per area. This parameter must be consistence.
2. The length of System ID: Default value is 0 (indicates the length of the System
ID is 6 bytes). This parameter must be consistence.
3. Area ID and IS-IS router type: Level-1 and Level-2 adjacencies can be formed

Confidential Information of Huawei. No Spreading Without Permission

if the Area IDs are the same. Only Level-2 adjacency can be formed if the Area ID is
different.
4. Authentication: If authentication has been configured, its parameters must be
matched with each other.
5. IP network segment: Network segment checking to ensure the network segment
is the same before neighbor relationship can be established, this function is turned on
by default on Huawei routers. Theoretically, the same network segment checking is not
necessary since IS-IS is run directly on top of the data link layer. This feature is turned
on to ensure neighbor relationship is established accurately. Recently a new command
ignore-ppp-ip-check is added in Huawei VRP. By executing this command, you can
cancel the network segment check on the PPP link. This command only applies to
PPP links.

Confidential Information of Huawei. No Spreading Without Permission

 At broadcast network, the adjacency relationship is formed via 3 ways
handshake process. IIH packets need to carry the neighbor identifier in order to
accomplish the 3 ways handshake process. In LAN IIH packets, TLV 6 is used to
carry neighbor identifier information.
In Point to Point IIH, TLV 6 doesn’t exist. The absence of information on
neighbors identifier in point-to-point IIHs as specified in the original hello format
caused reliability issues in forming point-to-point adjacencies. This is due to 3 ways
handshake can not be realized. Later, TLV Type 240 is proposed to address this
problem by recording the remote ID in Point to Point IIH packet. Obviously, 3 ways
handshake is much more reliable compared to 2 ways handshake. The situations
will not happen where one end of an adjacency is up but the other end is not.
In broadcast network, LAN IIH packet is used to form the adjacency relationship.
There are 2 types of LAN IIH namely L1 LAN IIH ( with multicast MAC address 01-
80-C2-00-00-14) and L2 LAN IIH ( with multicast MAC address 01-80-C2-00-00-
15). Level-1 ISIS routers exchange L1 IIH with each other in order to form
adjacency relationship. Level-2 ISIS routers exchange Level-2 IIH with each other
to form adjacency relationship.Level-1-2 ISIS routers exchange Level-1 LAN IIH
and Level-2 LAN IIH simultaneously to form adjacency relationship.
The process to establish adjacency relationship is as below:
1. A router sends IIHs on broadcast interfaces as soon as the interface is
enabled with IS-IS.
2. At the beginning, the router has not yet discovered the neighbor. Thus, the
neighbor information carried in TLV 6 is null.
3. When a router has received the IIH packet from the other side, the parameters
checking is proceeding (similar to the parameter checking in the Point to point link).
The neighbor existence is confirmed when the checking is passed. (It can not

Confidential Information of Huawei. No Spreading Without Permission

guarantee the other side aware of the existence of the local router, because the
neighbor information in the IIH packet received previously is null)
4. As a reply, the router will send an IIH packet with neighbor ID included.
5. The other side receives the IIH packet with its own ID included. At this moment,
the other side can confirm its own IIH has been received successfully. The adjacency
state become “ up” now.
3 ways handshake is as below:
1.Receive IIH packet, but its own ID is not included in TLV 6.
2.Send a IIH packet with its own ID included as a reply.
3.Receive IIH packet from the other side with its own ID included

Confidential Information of Huawei. No Spreading Without Permission

Confidential Information of Huawei. No Spreading Without Permission
 There are 2 types of LSP namely Level-1 LSP and Level-2 LSP. The link state
information is encapsulated into the LSP.
There are 2 types of CSNPs namely Level-1 CSNP and Level-2 CSNP. CSNP
advertised by a router contains summaries of all the known LSPs in its database.
There are 2 types of PSNPs namely Level-1 PSNP and Level-2 PSNP. PSNPs
complement CSNPs in the database synchronization process.

Confidential Information of Huawei. No Spreading Without Permission

 In IS-IS routing protocol, LSP ID is used to distinguish LSPs from each other and
to identify the originating routers. The LSP ID consists of three components namely
Source ID (system ID) , Pseudonode ID and LSP number.
Source ID (System ID): 6 bytes, indicates the system ID of the originating router
that generate the LSP.
Pseudonode ID: 1 byte, Pseudonode LSP is a special LSP generated by the
DIS on that link. Pseudonode LSP has non-zero pseudonode ID value. A router’s
ordinary LSP has pseudonode ID value of zero.
LSP Number: 1 byte, The LSP number refers to fragments of an LSP (regular or
pseudonode). The first fragment of an LSP is number zero. When any fragment of a
large LSP is lost in transmission, the receiver drops all the other fragments and the
whole set must be retransmitted
P field is used to establish level-1 adjacency relationship through level-2 routers,
just like OSPF's virtual link, but only take effect in level-2 layer. VRP doesn't
support this feature.
ATT: When the LSP generated by level-1-2 router advertised to the level-1 router
in the same area, this field should be set, and use this to tell the level-1 router that
through this level-1-2 router has an exit to the other areas. Then the level-1 router in
this area will generate a defaulf route, choose the nearest level-1-2 router as the
exit.
OL: indicates LSDB Overload. LSPs with the overload bit are still flooded in the
network, but the LSPs are not used when the route that passes a router configured
with the overload bit is calculated. That is, after a router is configured with the
overload bit, other routers ignore the router when perform SPF calculation. When
the memory of a router is not enough, the system automatically sends LSP packets

Confidential Information of Huawei. No Spreading Without Permission

with the overload bit.
From the diagram above, we can observe that there are 4 LSPs. LSP ID
0000.0000.0003.02-00 has non-zero pseudonode ID. This indicates that a
pseudonode LSP is generated by the pseudonode. Then, what is the difference
between the pseudonode LSP and ordinary LSP? What is the function of it?

Confidential Information of Huawei. No Spreading Without Permission

 ：
DIS Designated IS in the broadcast network, similar to the DR in OSPF.
Pseudonode: Pseudonode is not an actual router, but it occupies an extra LSP.

In IS-IS protocol, a DIS will be elected on the broadcast network. The DIS
generates a pseudonode to interact with other routers. A pseudonode is not a real
router, but it occupies an extra LSP. Actually, the pseudonode LSP is created by the
DIS.

Confidential Information of Huawei. No Spreading Without Permission

 In LAN, the DIS is elected based on the interface priority. The priority value
ranges from 0 to 127 and its default interface priority is 64. The higher the value, the
higher the priority is. Separate DISs are elected for level 1 and level 2 routing. The
highest subnetwork point of attachment (SNPA) is used as tie-breaker when the
interfaces priorities are the same. In LAN, SNPA is referred to MAC address. In
frame relay network, SNPA is referred to local data link connection identifier (DLCI).
If DLCI is used as SNPA in frame relay network, the SNPA might be the same for 2
ends of the link. In this scenario, the router with the highest System ID is elected to
be the DIS. The interface of every IS-IS routers can be configured with either Level
1 or Level 2 priorities ranging from 0 to 127.
Unlike OSPF, the IS-IS DIS can be preempted at any time by any eligible router
connecting to the LAN. The newly elected DIS is responsible for purging the old
pseudonode LSP and flooding the network with new LSP. The DIS concept is
applicable to broadcast network only. The DIS election is not required on point-to-
point network. Separate DISs are elected for level 1 and level 2 routing on
broadcast network. No backup DIS is elected for either level-1 or level-2. This
doesn’t turn out to be a problem because DIS transmits hello packets 3 times faster
compare to the other routers on the LAN. This allow for fast detection of DIS failure
and immediate replacement.

Confidential Information of Huawei. No Spreading Without Permission

 After the adjacency relationship has been formed on the point-to-point link,
routers will exchange the CSNPs that describe the contents of the local link state
database. Next, routers will compare the received CSNP with the content of local
link state database. At the same time, the routers will send LSP to the neighbor that
send out PSNP packet to request the transmission of current or missing LSPs. The
neighbor that has received the LSP will send a reply by using a PSNP packet for
acknowledgement purpose.
If the acknowledgement is not received within a specified period, referred to as
retransmission interval, the LSP is assumed lost during the transmission and is
retransmitted. This retransmission process is repeated on the point-to-point link until
the PSNP acknowledgement is received. This can ensure the integrity of the LSDB.
CSNP contains only the summaries of the known LSP in the local database and
this can simplify the process of database synchronization. Furthermore, a router will
proactively send a copy of the LSP to its neighbor when it discovers that its
neighbor doesn’t have any of the LSP’s in its local database. This can accelerate
the database synchronization process. Please notice that CSNPs are sent only
once when the IS-IS adjacency is initialized preceding the exchange of LSPs over
the link on point-to-point link. After the adjacency has been formed, only PSNP
packets are used to either request LSP or to acknowledge LSP. If there are any
changes in the LSDB (for example the interface cost has been changed), router will
directly send out a LSP packet regarding the changes to the neighbor. The neighbor
will then check the sequence number of the received LSP to determine whether
need renew the LSP. Then, the PSNP is used for acknowledgement. At the same
time, retransmission mechanism is used to ensure the integrity of the database.

Confidential Information of Huawei. No Spreading Without Permission

 In the diagram shown, RTC is the last router to join the broadcast network. RTA
and RTB are already connected to the broadcast network and RTB is selected as
the DIS. After establishing the adjacencies with RTA and RTB, RTC creates an
LSP, RTC.00-00. RTB (DIS) advertises a CSNP by multicast over the link. RTC
receives a copy of the CSNP, checks it against the local LSDB, and found 3
missing LSPs: RTA.00-00, RTB.00-00, and RTB.01-00. At this moment, RTC has
only its own LSP, RTC.00-00. in its local link state database. RTC then sends out a
PSNP to request the complete copies of RTA.00-00, RTB.00-00, and RTB.01-00.
RTB floods RTA.00-00, RTB.00-00 and pseudonode LSP RTB.01-00 through
multicast. And then RTC receives the copies.
Please note that RTB is the DIS. Thus RTB will generate a router LSP and a
pseudonode LSP.
On broadcast network, no retransmission mechanism is adopted and the
flooding over broadcast link is unreliable. IS-IS routers rely on periodic multicast of
CSNPs from the DIS to ensure the database synchronization over the broadcast
links.

Confidential Information of Huawei. No Spreading Without Permission

 1.Level-1-2 routers contain the Level-1 LSDB and Level-2 LSDB.
2.Level-1 LSDB contains the LSP of the local area while Level-2 LSDB contains
the LSP of the backbone area.
3.Level 1-2 routers set the ATT bit in the Level 1 LSPs originated by itself.
4.Its routing table contains the routing information of the whole network.

Confidential Information of Huawei. No Spreading Without Permission

 Level-1-2 routers which are connected to the backbone area set the ATT bit in the
level-1 LSPs they advertise into their native area. The level-1 router in the local area
might receive many copies of the Level-1 LSP with ATT bit set. In this case, the
level-1 router will select the nearest level-1-2 router. At the same time, the level-1
router will generate a default route point to the selected nearest level-1-2 router.
After that, the level-1 router will send the data packets to the selected level-1-2
router by using the default route. In this example, RTA is level-1 router and it does
not contain the routing information on network 4.4.4.4. If RTA wants to access area
49.0002, it can only transmit the data packets to the nearest level-1-2 router. From
the diagram above, we observe that the cost from RTA to RTB is 10 while the cost
from RTA to RTC is 20. Therefore, RTA will select RTB as the nearest level-1-2
router and generate a default route point to RTB

Confidential Information of Huawei. No Spreading Without Permission

 The routing table of RTA does not contain the routing information on network
4.4.4.4.The cost from RTA to RTB is 10 while the cost from RTA to RTC is 20.
Therefore, RTA will select RTB as the nearest level-1-2 router and generate a default
route point to RTB.
When RTA wants to access network 4.4.4.4, it will first sends the data to the
nearest level-1-2 router. In the diagram above, the nearest level-1-2 router is RTB.
As a result, the less optimal path has been selected. This is because the cost from
RTA to RTD via RTB is 40. However, the cost from RTA to RTD via RTC is only 30.

Confidential Information of Huawei. No Spreading Without Permission

 To eliminate the problem of suboptimal routing presented in the previous slide, we
can configure route leaking on RTB and RTC. The routing table of RTA contains the
routing information towards network 4.4.4.4 after the configuration of route leaking on
RTB and RTC. At this moment, RTA no longer uses the default route to transmit the
data to network 4.4.4.4. RTA will choose RTA->RTC->RTD as the best path toward
network 4.4.4.4.

Confidential Information of Huawei. No Spreading Without Permission

Confidential Information of Huawei. No Spreading Without Permission
 We will learn the process of SPF calculation by using this example.
In this diagram, RTA is the root node ( this indicates that SPF computation is run on
RTA), the numerical number beside the link represents the cost of the link.

Confidential Information of Huawei. No Spreading Without Permission

 The RTA router generates the SPT as shown in the figure with itself as the root and
other routers as the leaf nodes through the SPF algorithm. Then it calculates the route
from root node to each destination network segment based on the overhead of the
SPT. The generated SPT is unidirectional and irreversible, so the root loop can be
avoided in the area through the algorithm.
This diagram shows the SPT in which the A is functioned as the Root.

Confidential Information of Huawei. No Spreading Without Permission

If B is functioned as root, the result of the SPF algorithm is shown as above.

Confidential Information of Huawei. No Spreading Without Permission

Confidential Information of Huawei. No Spreading Without Permission
 Currently, both IS-IS and OSPF are widely used IGP protocols. They both describe
the network topology by using the link state information. Initially, IS-IS was designed
only for CLNP environment. Later, RFC1195 has defined integrated IS-IS which
enable it to run on both CLNP and IP environment. In contrast, OSPF can only run in
IP environment but not in CLNP environment. Both IS-IS and OSPF adopts
hierarchical design and area division. IS-IS divides the network into level-1 and level-
2 (backbone area). OSPF divides the network into backbone area (area 0) and non
backbone area. The backbone of both IS-IS and OSPF must be contiguous. OSPF
supports the use of virtual link to connect physically separate backbone areas to
maintain the contiguity of the backbone but IS-IS not support virtual link. OSPF can
provide direct support to many network types compared with IS-IS. But IS-IS can also
provide support to various network types via proper planning. For example, NBMA
media can be configured as broadcast mode if all nodes are fully meshed in IS-IS
network.

Confidential Information of Huawei. No Spreading Without Permission

 OSPF is encapsulated in IP packet while IS-IS packet is encapsulated in data link
layer.
Both OSPF and IS-IS adopt area division but their implementation mechanisms
are different. In IS-IS, the routers are associated with a single area and the whole
router belongs to the area. In OSPF, the interfaces of the router could be assigned to
different areas. Therefore, the routers can be attached to multiple areas

Confidential Information of Huawei. No Spreading Without Permission

 Both IS-IS and OSPF use Hello protocol to discover neighbors and form
adjacencies relationship. However, the conditions for adjacencies formation are
different. The Hello protocol adopted by the IS-IS protocol is quite simple, while that
adopted by the OSPF protocol is more complicated. In addition, IS-IS protocol does
not require the hello interval and dead interval to match between the adjacent
neighbors. In contrast, OSPF protocol requires the hello interval and dead interval to
match on all routers on the same subnet. On PPP links, IS-IS protocol adopts 2 ways
handshake to form adjacencies relationship and its reliability is not as good as OSPF.
However, there exists 3 ways handshake on PPP link of the IS-IS protocol. Both
OSPF and IS-IS protocol adopt DR concept on multi-access network but the
implementation mechanism of these 2 protocols are different.

Confidential Information of Huawei. No Spreading Without Permission

The DIS election of IS-IS protocol is simpler. IS-IS elects only one DIS without backup.
This doesn’t introduce any problems as DIS advertises hello 3 times faster than other
nodes on the LAN. Therefore, the DIS failure detection is fast and new DIS is elected
when current DIS failed. Besides, the IS-IS DIS can be preempted at any time by a
router with a higher priority. IS-IS router with priority 0 can also take part in the DIS
election.
In contrast, the DR election mechanism on OSPF is more complicated. DR and BDR
are elected on OSPF protocol. The DR cannot be preempted and the BDR will take
over when the DR failed. Routers with priority 0 are not eligible to take part in the DR
election process. Besides, router with highest priority in the network might not
necessarily be elected as the DR. In OSPF, all nodes on the LAN can form
adjacencies only with DR and BDR. This is necessary because OSPF requires
databases to be reliably synchronized before adjacencies are established.

In conclusion, IS-IS solves the NxN adjacency problem by choosing a simple process
that uses periodic synchronization to achieve reliability and a deterministic DIS
election process in which the DIS is preemptive. OSPF solves the same problem but
uses a more complex process.

Confidential Information of Huawei. No Spreading Without Permission

 Both IS-IS and OSPF are link state protocols and they must form the consistent
link state database before calculate the route. OSPF uses LSA to describe the state
of the link while IS-IS uses LSP to describe the state of the link.

Confidential Information of Huawei. No Spreading Without Permission

 OSPF uses various types of LSA and the structure of its database is very
complicated. Therefore, the troubleshooting of OSPF protocol is more difficult. In
contrast, the structure of IS-IS database is simpler and therefore is easier for
troubleshooting. The process of link state database synchronization of OSPF and IS-
IS are different. The remaining time of the LSP in the IS-IS protocol is counted down
to 0 from 15 minutes (configurable) to clear the old LSP, while the remaining time in
the OSPF protocol is counted up to 60 minutes (not configurable) from 0 to clear the
old LSA.

Confidential Information of Huawei. No Spreading Without Permission

 Both IS-IS and OSPF use the same SPF algorithm for route calculation. However,
OSPF treats the IP prefixes as the nodes of the SPT while IS-IS treats the IP prefixes
as the leaves of the SPT. Therefore, the route calculation in IS-IS protocols is much
simpler. IS-IS performs only the partial route calculation (PRC) when the changes do
not affect the basic topology of the SPT but only IP prefixes. In this case, the full SPF
calculation is not necessary.
In addition, the interface cost of IS-IS protocol is ranging from 0 to 63. The total
metric for a path is the sum metric on the outgoing interfaces of all links between the
source and the destination. ISO 10589 specifies the maximum metric value for a
complete path to be only 1023. Therefore, value 1024 is unreachable for in IS-IS
network. In contrast, the interface cost of OSPF is ranging from 0 to 1024. Value
65535 is unreachable in OSPF network.
IETF has expanded the range of the IS-IS cost metric and path metric in
rfc3784.txt. Therefore, the metric type of IS-IS is more complicated compared with
OSPF. The metric of IS-IS can be divided into narrow, narrow-compatible, wide,
wide-compatible and compatible.

Confidential Information of Huawei. No Spreading Without Permission

Confidential Information of Huawei. No Spreading Without Permission
Q: What are the packets used by IS-IS protocol in link state database synchronization?
Describe the function of those packets.
A: The packets used by IS-IS protocol in link state database synchronization are
CSNP, PSNP and LSP. The function of CSNP is to describe all the LSPs in its own
LSDB. The function of PSNP is to request or acknowledge certain LSP. The function
of the LSP is to carry the state information.
Q: Please describe the main working processes of IS-IS.
A:There are 3 working proccesses to achieve this goal:
1. Establish the adjacency relationship
Discover the neighbor and establish the adjacency relationship. In this way, the link
state information can be exchanged. Thus, adjacency relationship establishment is
the basic of IS-IS working proccess.
2.Synchronize the link state database
IS-IS is a link state routing protocol used to calculate the precise and integrated route.
Every router must be equipped with a set of accurate and identical LSDB. To achieve
this, the database must be synchonized first through the exchange of link state
information between the adjacent routers. The final result is that every router is having
a synchonized LSDB which is used to describe the whole topology of the network
3.Calculate the route
Use the synchronized LSDB information to run SPF algorithm and calculate the
routes. The routing information is then used to direct the data transmission.

Confidential Information of Huawei. No Spreading Without Permission

Confidential Information of Huawei. No Spreading Without Permission
Confidential Information of Huawei. No Spreading Without Permission