Summary of network protocol learning

• OSPF (Open Shortest Path First Protocol)

1. OSPF protocol devices transmit LSAs instead of routing tables.
2. Interior gateway routing protocol, based on link state, works in a single autonomous area.
3. The routing table is not updated periodically but only when the path changes, and only the changed routing information is sent.
4. OSPF collects link state information and stores it in the link state database (the information in the link state database of each router in the AS is the same), and then calculates the respective routes through the Dijkstra algorithm, so there is no loop.
5. Working process: neighbor discovery (send Hello message) -> route advertisement -> route calculation.
6. In order to solve the problem of confusion caused by sending LSAs between neighbors, resulting in low work efficiency and excessive resource consumption, DR/BDR (BDR is the backup of DR) is proposed. After the DR/BDR is elected, the router sends the LSA to the DR/BDR, and the DR multicasts it to all non-DR/BDR ROUTERs.
7. There are three types of routers: DR, BDR, and DR-Other.
8. OSPF controls the election of DR/BDR through priority (DR/BDR election is based on ports, and the OSPF router priority is 1 by default. If it is manually set to 0, the device will not participate in the election), and the one with the highest priority is DR. The second largest is BDR, (when the DR fails, the BDR acts as the DR, and then elects a new BDR). When the priorities are the same, compare the Rout-ID (the one with the larger Rout-ID is the DR, and the second largest is the BDR).
9. If the DR/BDR has been elected, if a new router is added, even if its priority is high, it will not become the DR of the network segment, so the DR is not necessarily the router with the highest priority.
10. The route calculation is based on the cost value (cost=ideal bandwidth of the port/actual bandwidth of the port). Cost is calculated by adding the cost value of the port when sending, and not adding it when receiving.

• The difference between VLSM and CIDR

VLSM (Variable Length Subnet Mask):

Process: On the basis of a class A/B/C IP address, divide the subnet.

Impact: The subnet mask becomes longer (the network prefix becomes longer, and the host bits become shorter)

Function: save IP address space and reduce the length of routing table.

CIDR (Classless Inter-Domain Addressing):

Process: Use the "/" method to cancel the classification structure of the IP address.

Impact: The longest match strategy is used for routing.

Function: Through route aggregation, multiple subnets are formed into a supernet, which reduces the length of the routing table.

Summary : VLSM makes the subnet mask longer, aiming to save IP, VLSM makes the subnet mask shorter, and cancels the IP address classification structure.

• IS-IS protocol (Intermediate System to Intermediate System Routing Protocol)

1. An interior gateway protocol based on the link state and using the shortest path first algorithm for routing, the protocol priority is 15.
2. ISIS calculates the shortest path based on data link layer packets, while OSPF calculates based on the IP layer.
3. There are three types of routers: L1 (Level1) routers, L2 (Level/2) routers, and L1/2 (Level1/2) routers.
4. In order to support a large-scale routing network, ISIS adopts a two-level hierarchical structure of backbone areas and non-backbone areas in the autonomous system.
5. Level-1 routers are deployed in non-backbone areas, and Level-2 routers are deployed in backbone areas. Each non-backbone area is connected to the backbone area through a Level-1-2 router.
6. ISIS route calculation process: establish neighbor relationship -> link information exchange -> route calculation

• Establish a neighbor relationship: Negotiate various parameters by sending Hello packets to each other to establish a 2Way relationship (Huawei's 3way establishes a reliable connection).
• Link information exchange: the carrier is LSP (in order to alleviate LSP flooding, a pseudo-node is introduced), there are two situations. 1. The A-end router is elected as a pseudo node. In addition to the general IS router function, the pseudo node will also send two types of messages: LSP_DIS (for notifying IS neighbor information) and CSNP (LSP summary). 2. The B-end router is recommended as a pseudo-node, and the A-end router requests a specific message by sending PSNP.
• Routing calculation: Execute SPF to calculate routing, and each node calculates the shortest path with itself as the root.

1. Differences of various router LSDBs:

• Level-1 can only form neighbor relationships with Level-1 and level-1-2 routers in the same area, and is only responsible for maintaining the LSDB of Level-1, so it only forms SPF routes in the area, and points the next hop to the router through the default route. Level-1-2 routers communicate in different areas at the same time.
• A Level-2 router can form a neighbor relationship with the same or different Level-2 routers or different Level-1-2 routers. The LSDB contains only Level-2 link information, and the routing table contains routes for the entire network.
• The Level-1-2 router has two LSDBs, and the routing table contains the routes of the entire network.

Differences between OSPF and IS-IS:

1. ISIS calculates the shortest path based on data link layer packets, while OSPF calculates based on the IP layer.
2. In the IS-IS protocol, each router belongs to one area, but in OSPF, different ports of a router belong to different areas.
3. There is no distinction between backbone area and non-backbone area in a single area, and Area0 is defined as the backbone area in OSPF.
4. ISIS calculates the shortest path in each area to form a route, while in OSPF, the shortest path is used to form a route in the entire AS.

• MPLS (Multiprotocol Label Switching Protocol)

1. A tunneling technology that combines the third-layer routing technology and the second-layer switching technology to give full play to the flexibility of IP routing and the simplicity of layer-2 switching.
2. The network devices that perform MLPS label switching and message forwarding become LSRs (Label Switching Routers), and the network area formed by LSRs becomes an LSR domain. Routers in an MLPS domain are divided into two types: core LSRs and border routers LERs.
3. A label is a short and fixed-length identifier with local significance, which is used to uniquely identify a forwarding equivalence class (MPLS classifies packets with the same characteristics into one class, and the same equivalence class has the same processing strategy in the same MLPS domain ,).
4. When the IP message enters the MPLS network, the label is added at the MPLS ingress, and when the IP message leaves the MPLS network, the label is ejected by the egress LER. The label is sent by border router A to border router B, then A is called an upstream node, and B is a downstream node.
5. The path that an IP packet passes on an MPLS network is called a label switching path (LSP).
6. There are two ways to establish an LSP:

• Static LSP establishment: The user manually assigns labels to each forwarding equivalence class. The rule to be followed is that the value of the outgoing label of the previous node is equal to the value of the incoming label of the next node.
• Dynamic LSP establishment: dynamically established through the transition release protocol. Labels are allocated by downstream LSRs in the direction from downstream to upstream. In order to reduce the burden of the last hop, the label is popped up at the penultimate hop, so the label value of the penultimate hop is usually set to 3 on devices that support PHP (popup feature).

1. The transmission of IP packets on the MPLS network is the process of pushing labels -> label replacement -> label popping.

MPLS TE (MPLS Traffic Engineering)

1. MPLS TE combines MPLS technology and traffic engineering, and reserves resources by establishing an LSP tunnel to a specified path, so that network traffic bypasses congested nodes and achieves the purpose of balancing network traffic.
2. In the case of tight resources, MPLS TE can seize the bandwidth resources of low-priority LSP tunnels to meet the needs of large-bandwidth LSPs or important users.
3. LSP tunnel: For an LSP, once the source node assigns a label to the message, the forwarding of the traffic is completely determined by the label, which is transparent to the intermediate node. In this sense, an LSP can be regarded as a tunnel.
4. MPLS TE tunnel: When rerouting is deployed or traffic needs to be transmitted through multiple paths, multiple LSP tunnels may be used. In TE, such a group of LSP tunnels is called a TE tunnel.
5. CR-LSP: An LSP established based on certain constraints, and certain conditions such as specified bandwidth or QOS parameters are added to the original basis.

• Switching

In order to isolate the broadcast domain, VLAN is set up. In order to realize the communication between VLANs, there are two ways: 1. Through the configuration of different types of ports on the switch, the conversion of tag and untag 2. Layer 3 switch configuration IP. In order to further divide VLANs, PVLANs are generated, and VLANs are divided into two layers. The upper layer VLANs are globally visible, and the lower layer VLANs are invisible to each other. And in order to improve the reliability of the link, the link redundancy technology is proposed, and it will be formed at the same time, so the STP protocol should be adopted.

• IPv6

The expansion of IPV4, on the one hand, is to alleviate the problem of insufficient IPV4, and on the other hand, it is for high security.

IPV4 is 32 bits, and IPV4 is 128 bits. IPV6 is divided into 8 segments by ":", and each segment is represented by 4 hexadecimal numbers. In order to simplify the writing, the first 0 of each paragraph can be omitted, and it can also be represented by "::" (all the 0s in between are omitted). In addition, the first 92 digits can be expressed in hexadecimal, and the last 32 digits can be expressed in decimal.