OSPF convergence process

OSPF convergence is called LSA flooding, also called LSDB synchronization;

1. OSPF interface network type-OSPF protocol works in different ways under different network type interfaces

Interface network type OSPF interface network type (ospf working mode)
LoopBack 0. Cisco-LoopBack does not have hello packets sent by the 32-bit host route Huawei-displayed as p2p type actually LoopBack working mode
Point-to-point (serial HDLC/PPP/GRE) p2p. hello time10s Automatically establish neighbors without selecting DR/BDR
BMA/Ethernet Broadcast hello time 10s automatically establish neighbor selection DR/BDR
NBMA/Frame Relay nbma hello time 30s Manually establish neighbors and select DR/BDR
MGRE p2p. hello time10s Automatically establish a neighbor without selecting DR/BDR-There can only be one neighbor in a network segment; when a Huawei device is in a MGRE network segment and the interface is in point-to-point mode, only the device that receives the hello first Establish neighbor relationships; Cisco will roll over neighbors in this situation;

    Note: In the MGRE environment, the default ospf working mode of the interface is point-to-point. This method cannot realize the full connection of neighbors in the NBMA network segment; therefore, you can only modify the working mode of the interface:
    Modify all interfaces on the MGRE network segment to the Broadcast working mode. Remember that if some of the interfaces are modified to Broadcast, the other part of the interfaces will still remain point-to-point. The neighbor relationship can be established because the hello time is the same, but the working mechanism is different in the DR/BDR election. In the end, the network segment cannot converge normally;
    [R1]int t0/0/0
    [R1-Tunnel0/0/0]ospf network-type broadcast

    [1] At the same time, we must pay attention to the network topology; if the network segment is a fully connected mesh structure, then the DR election will proceed normally; but if it is a partial mesh or center-to-site topology, DR location errors may occur;
    [ 2] If the network topology can only be partial mesh or center-to-site, you need to manually intervene in the DR location
                or modify the working mode of all interfaces to point-to-multipoint working mode;

    Ospf's point-to-multipoint working mode: It can only be configured manually and is suitable for some mesh structure topologies;
    Hello time 30s, DR/BDR is not selected, and neighbor relationships are automatically established;

2. Irregular areas of OSPF

    If an ABR device is not connected to the backbone area 0, it cannot share routes between areas by default.
    1. Non-backbone areas far away from the backbone
    2. Discontinuous backbone

    Solution:
    1) Establish a tunnel on the legal and illegal ABR, and then declare it to the ospf protocol
    Disadvantages: 1. Periodic update, keep-alive, trigger update to cause resource occupation for the intermediate crossing area
    2. Poor routing-when ospf When learning two identical goals but coming from different areas at the same time, the backbone area is preferred;

    2) Virtual link—Establish virtual links on legal and illegal ABRs, authorized by legal ABRs for illegal ABRs;
    enable illegal ABRs to share routes between areas;
    because no new paths are added, there is no routing Poor problem;
    [r2-ospf-1]area 1 The area where two ABRs are located together
    [r2-ospf-1-area-0.0.0.1]vlink-peer 4.4.4.4 RID of the peer ABR
    Cisco in order to avoid periodic information For the occupation of the intermediate area, cancel all periodic behaviors
    on the virtual link ; —Unreliable Huawei still maintains periodic keep-alive and update on the virtual link — Resource occupation of the intermediate area

    3)Multi-process two-way re-release (recommended)
    Multi-process—On the same device, different processes can work on different interfaces, establish their neighbor relationships, and generate their own databases (not shared); only the calculated routes are loaded into the same routing table; one interface Can only be announced by one process;
    two-way redistribution, ASBR (autonomous system border router, protocol border router), two-way sharing of routes generated by different processes or different protocols;
    [r2]ospf 1
    [r2-ospf-1]import -route ospf 2
    [r2-ospf-1]q
    [r2]ospf 2
    [r2-ospf-2]import-route ospf 1

3. Various LSAs of OSPF-database tables

    <r1>display ospf lsdb
    <r1>display ospf lsdb router 1.1.1.1
        category name link-id (number in the directory)
    All categories of LSAs carry the following information
    Type: Router category name here is Class 1
    Ls id: 1.1.1.1 The number of the link-id in the catalog
    Adv rtr: 1.1.1.1 Notifier—The RID of the source device for the LSA update
    Ls age: 1359 Aging time of 1800s, refresh cycle to 0, trigger update to 0, max 3609s
    Len: 48
    Options: E
    seq#: 8000000c serial number
    chksum: 0x818c checksum
    OSPF LSA is 1800 will be updated once, the serial number will be updated once,
    the new and old comparison of LSA will be added 1. The serial number will be compared
    first, the larger the serial number, the better,
    2. If the sequence is If the number is the same, the larger the checksum, the better
    3. If the checksum is also the same, it will compare the LSA Age time to see if it is equal to the MAX-age time (3600)
    4. If the age time is not equal to the max-age time, Will compare their differences, if the difference is greater than 15 minutes (900 seconds), the smaller is the best
    5. If the age time is not equal to the max-age time, the difference between them will be compared. If the difference is less than 15 minutes, it means the same LSA.
    Under what circumstances will the LSA be updated if one of them is ignored :
    1. 1800 will be updated periodically when it expires
    2. Trigger update (interface address change (add or delete), modify interface cost value, delete interface, or delete notification)

Category name Spread Notifier Information carried
Type 1 LSA-router In this area Every router in this area Direct connection topology of each device in the area
Type 2 LSA-Network In this area DR of this network segment The topology of the network segment
Type 3 LSA-summary The entire OSPF domain APR O IA inter-domain routing
Type 4 LSA-asbr The entire ospf domain except the area where the ASBR is located; used in the area where the ASBR is located; type 1 mark location APR ASBR location
5 types LSA-ase The entire OSPF domain ASBR OE external routing
Category name link-id Notifier
Type 1 LSA-router RID of the notifier Every router in this area
Type 2 LSA-Network IP address of the DR interface DR in each MA network segment
Type 3 LSA-summary Destination network number for inter-domain routing ABR, when passing the next ABR, it is modified to the new ABR
Type 4 LSA-asbr RID of ASBR ABR, when passing the next ABR, it is modified to the new ABR
5 types LSA-ase Destination network number of the external route ASBR does not change when passed inside ospf

Guess you like

Origin blog.csdn.net/Han_V_Qin/article/details/112709614