No relevant resource is found in the selected language.

This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Read our privacy policy>Search

Reminder

To have a better experience, please upgrade your IE browser.

upgrade

AR500, AR510, and AR530 V200R007 CLI-based Configuration Guide - Ethernet Switching

This document describes the configuration of Ethernet services, including configuring transparent bridge, MAC table, link aggregation, VLANs, STP/RSTP/MSTP, and so on.The document provides the configuration procedures and configuration examples to illustrate the service configuration methods and application scenario.
Rate and give feedback:
Huawei uses machine translation combined with human proofreading to translate this document to different languages in order to help you better understand the content of this document. Note: Even the most advanced machine translation cannot match the quality of professional translators. Huawei shall not bear any responsibility for translation accuracy and it is recommended that you refer to the English document (a link for which has been provided).
SEP Implementation Mechanisms

SEP Implementation Mechanisms

Neighbor Negotiation Mechanism

After an interface is added to a SEP segment, neighbor negotiations start. The interface and its neighbor exchange Hello packets to establish a neighbor relationship. After neighbor negotiations succeed, the two interfaces continue to exchange Hello packets to detect their neighbor status.

Neighbor negotiations prevent unidirectional links because neighbor negotiations are bidirectional. Interfaces at both ends of a link, must send Hello packets to each other, as a means of status confirmation. If an interface does not receive a Hello packet from an interface at the other end of a link within a specified period, the interface considers the other to be Down.

Neighbor negotiations provide information required to obtain the SEP segment topology. Interfaces establish neighbor relationships through neighbor negotiations, forming a complete SEP segment. Therefore, the SEP segment topology can be obtained.

Synchronization of SEP LSA Databases and Topology Display

  • Synchronization of SEP link state advertisement (LSA) databases

    After neighbor negotiations are complete, devices in a SEP segment enter the LSA database synchronization phase and periodically send LSAs. After a device receives LSAs from other devices, the device updates its LSA database. This ensures that the LSA databases of all devices in the SEP segment are consistent.

    If a device does not receive LSAs from its peer device or other devices in the SEP segment within three LSA transmission intervals, the device will age the database that saves the LSAs of the other devices in the SEP segment.

    When a faulty device in a SEP segment recovers, the device needs to obtain topology information from the other devices in the SEP segment and sends LSA request packets to the other devices. After receiving LSA request packets from the device, neighboring interfaces reply with LSA ACK packets that contain the latest link state information.

  • SEP segment topology display

    The topology display function allows you to view the topology with the highest network connectivity on any device in a SEP segment. Link state synchronization ensures that all devices in a SEP segment display the same topology.

    Table 13-4 shows the types of SEP segment topologies.

    Table 13-4  Types of SEP segment topologies

    Topology Type

    Description

    Constraint

    Ring topology

    Each interface in a SEP segment has a neighboring interface in Up state and a brother interface, and each node has two interfaces in the SEP segment.

    • If the primary edge interface is elected on a ring, the primary edge interface is listed first in the topology information displayed on each interface.

    • If the primary edge interface is not elected but the secondary edge interface is elected, the secondary edge interface is listed first in the topology information displayed on each interface.

    Linear topology

    All topologies except ring topologies are linear topologies.

    For interfaces at both ends of a link:
    • If one interface functions as the primary edge interface, the primary edge interface is listed first in the topology information displayed on each interface.

    • If the primary edge interface is not elected but the secondary edge interface is elected, the secondary edge interface is listed first in the topology information displayed on each interface.

    NOTE:

    The constraints listed in Table 13-4 ensure that each node in a ring or linear topology displays the same topology information.

Primary Edge Interface Election

Only interfaces that are configured as no-neighbor edge interfaces, primary edge interfaces, and secondary edge interfaces can participate in primary edge interface election.

NOTE:

If only one interface on a node has SEP enabled, you must set the role of the interface to edge so that the interface can function as an edge interface.

As shown in Figure 13-3, if there is no faulty link on the network and SEP is enabled on the interfaces, the following situations occur:
  • Common interfaces do not participate in primary edge interface election. Only P1 on Router1 and P1 on Router5 participate in primary edge interface election.

  • If P1 on Router1 and P1 on Router5 have the same role, P1 with a higher MAC address is elected as the primary edge interface.

After the primary edge interface is selected, it periodically sends primary edge interface election packets without waiting for the success of neighbor negotiations. A primary edge interface election packet contains the interface role (primary edge interface, secondary edge interface, or common interface), bridge MAC address of the interface, interface ID, and integrity of the topology database.

Figure 13-3  Networking diagram of electing the primary edge interface

As shown in Figure 13-3, if a link fault occurs in the SEP segment, P1 on Router1 and P1 on Router5 receive fault notification packets or P1 on LSW5 does not receive primary edge interface election packets within a specified period. Then P1 on Router1 becomes the secondary edge interface. Consequently, two secondary edge interfaces exist in the SEP segment and periodically send primary edge interface election packets.

When all link faults in the SEP segment are rectified, the two secondary edge interfaces can receive primary edge interface election packets and elect a new primary edge interface within a configured interval (1s by default).

Specifying an Interface to Block

Normally, a blocked interface is one of the two interfaces that complete neighbor negotiations last. In some cases, however, the negotiated blocked interface may not be the required one. You can specify an interface to block according to network requirements. The specified interface preempts to be the blocked interface only after the preemption mechanism takes effect.

  • Interface blocking mode

    You can configure the interface blocking mode to specify a blocked interface. Table 13-5 lists interface blocking modes.

    Table 13-5  Interface blocking mode

    Interface Blocking Mode

    Description

    Specify the interface with the highest priority as the blocked interface.

    SEP compares interface priorities as follows:
    1. Compares configured interface priority values. A larger value indicates a higher priority.

    2. Compares bridge MAC addresses of interfaces with same priority values. A smaller bridge MAC address indicates a higher priority.

    3. Compares interface numbers of interfaces with identical bridge MAC addresses. A smaller interface number indicates a higher priority.

    Specify the interface in the middle of a SEP segment as the blocked interface.

    -

    Specify a blocked interface based on the configured hop count.

    SEP sets the hop count of the primary edge interface to 1 and the hop count of the neighboring interface of the primary interface to 2. Hop counts of other interfaces increase by steps of 1 in the downstream direction of the primary edge interface.

    Specify a blocked interface based on the device and interface names.

    After SEP is configured, the interface to be blocked is determined by the device and interface names. Before specifying an interface to block, run the display command to view the current ring topology and all interfaces, and then specify the device and interface names.

    If multiple interfaces on the ring have the same device and interface names, SEP blocks the interface nearest to the primary edge interface in the topology.

    NOTE:

    If you change the device name or interface name after specifying the interface to block, the interface cannot preempt to be the blocked interface.

  • Preemption

    After the interface blocking mode is specified, whether a specified interface will be blocked is determined by the preemption mode. Table 13-6 lists the preemption modes.
    Table 13-6  Preemption mode

    Preemption Mode

    Description

    Non-preemption mode

    When all link faults are rectified or the last two interfaces enabled with SEP complete neighbor negotiations, interfaces send blocking status packets to each other. The interface with the highest priority is then blocked, and the other interfaces enter the Forwarding state.

    Preemption Mode

    NOTE:

    Preemption can only be implemented on the device where the primary edge interface or no-neighbor primary edge interface resides.

    Preemption is classified into delayed preemption and manual preemption.

    • Delayed preemption

      After all the faulty interfaces recover, the edge interfaces no longer receive fault notification packets. If the primary edge interface does not receive fault advertisement packets within 3 seconds, it starts the delay timer. After the delay timer expires, nodes in the SEP segment start blocked interface preemption.

    • Manual preemption

      When the link status databases of the primary edge interface and secondary edge interface are complete, the primary edge interface or brother interface of the no-neighbor primary edge interface sends preemption packets to block a specified interface. The specified interface then sends blocking status packets to request the previously blocked interface to transition to the Forwarding state.

      NOTE:

      Only two interfaces on a device can be added to the same SEP segment. If one interface is the no-neighbor primary edge interface, the other interface is the brother interface of the no-neighbor primary edge interface.

      Whether the brother interface of the no-neighbor primary edge interface needs to send preemption packets depends on whether the brother interface is blocked.

      • If the brother interface is blocked, it does not need to send preemption packets.
      • If the brother interface is unblocked, it needs to send preemption packets.

SEP Topology Change Notification

SEP considers that the topology of a SEP-enabled network changes in either of the following situations described in Table 13-7.

Table 13-7  SEP topology change notification

SEP Topology Change Notification

Description

An interface fault occurs.

Figure 13-4 shows an interface fault in a SEP segment.

An interface fault can be a link fault or neighboring interface fault.

If a device having an interface in Forwarding state in the SEP segment receives a fault advertisement packet, the device needs to send a Flush-Forwarding Database (Flush-FDB) packet through the interface to notify other nodes in the SEP segment that there is a change in topology.

The fault is rectified and the preemption function takes effect.

After faults occur in the SEP segment and the last faulty interface recovers, the blocked interface is preempted and the topology is considered changed.

Preemption is triggered by the primary edge interface. When an interface in a SEP segment receives a preemption packet from the primary edge interface, the interface needs to send Flush-FDB packets to notify other nodes in the SEP segment that there is a change in topology.

Figure 13-4  Networking diagram for SEP topology change notification

NOTE:

The topology change notification function is configured on devices that connect an upper-layer network and a lower-layer network. If the topology of one network changes, devices affected inform the other network of the change.

Table 13-8 lists the scenarios in which topology changes are reported.

Table 13-8  SEP topology change notification

SEP Topology Change Notification

Scenario

Description

Solution

Topology change notification from a lower-layer network to an upper-layer network

A SEP network is connected to an upper-layer network running other features such as SEP and STP.

  • If the blocked interface on a lower-layer SEP network is manually changed, the topology of the SEP segment changes. Because the upper-layer network is unable to detect the change in topology, traffic is interrupted.
  • If an interface on a lower-layer SEP network becomes faulty, the topology of the SEP segment changes but the upper-layer network is unable to detect the change. As a result, traffic is interrupted.

Configure the SEP topology change notification function.

Suppression of SEP TC Notification Packets

Topology changes of a SEP segment are advertised to other SEP segments or upper-layer networks. A large number of topology change (TC) notification packets are generated in the following cases:
  • A link becomes disconnected transiently.
  • A SEP segment is attacked by invalid TC notification packets.
  • There are multiple SEP ring networks.

    Figure 13-5 shows a networking scenario with three SEP ring networks. If the topology of SEP segment 3 changes, the number of TC notification packets doubles and SEP segment 2 is flooded with these packets. Each time TC notification packets pass through a SEP segment, the number of TC notification packets doubles.

    Figure 13-5  Networking diagram for multiple SEP ring networks

Sending a large number of TC notification packets reduces the CPU capability to quickly process other types of packets. In addition, devices in SEP segments frequently update MAC address entries, heavily consuming bandwidth resources. To solve such problems, the following measures can be taken to suppress TC notification packets:
  • Configure a device to process only one of the TC notification packets carrying the same source address.

  • Configure a device to process a specified number of TC notification packets within a specified period. By default, three TC notification packets with different source addresses are processed in 2s.

  • Avoid the networking scenario having more than three SEP ring networks.

SEP Multi-Instance

In common SEP networking shown in Figure 13-6, a physical ring network can be configured with only one SEP segment in which only one interface can be blocked.

If an interface in a complete SEP segment is blocked, all service data is transmitted only along the path where the primary edge interface is located. The path where the secondary edge interface is located remains idle, wasting bandwidth.

Figure 13-6  Networking diagram for SEP

SEP multi-instance allows two SEP segments to be configured on a physical ring. Each SEP segment independently detects the completeness of the physical ring, blocks or unblocks interfaces without affecting the other.

A physical ring may contain one or two SEP segments. Each SEP segment needs to be configured with a protected instance, each protected instance indicating a VLAN range. The topology calculated by a SEP segment is only valid for that SEP segment.

After different protected instances are configured for SEP segments and the mapping between protected instances and VLANs is set, a blocked interface is only valid for the VLANs protected by the SEP segment where the blocked interface resides. Data traffic for different VLANs can be transmitted along different paths. This implements traffic load balancing and link backup.

Figure 13-7  Networking diagram for SEP multi-instance

As shown in Figure 13-7, the SEP multi-instance ring network that consists of Router1 to Router4 has two SEP segments. P1 is the blocked interface in SEP segment 1, and P2 is the blocked interface in SEP segment 2.
  • Protected instance 1 is configured in SEP segment 1 to protect the data from VLAN 100 to VLAN 200. The data is transmitted along path Router1->Router2. As the blocked interface in SEP segment 2, P2 blocks only the data from VLAN 201 to VLAN 400.
  • Protected instance 2 is configured in SEP segment 2 to protect the data from VLAN 201 to VLAN 400. The data is transmitted along path Router3->Router4. As the blocked interface in SEP segment 1, P1 blocks only the data from VLAN 100 to VLAN 200.

When a node fault or link fault occurs, each SEP segment calculates its own topology independently, and the nodes in each SEP segment update their own LSA databases.

As shown in Figure 13-8, a fault occurs on the link between LSW3 and LSW4. The link fault does not affect the transmission path for the data from VLAN 100 to VLAN 200 in SEP segment 1, but blocks the transmission path for the data from VLAN 201 to VLAN 400 in SEP segment 2.

Figure 13-8  Networking diagram for a link fault on a SEP multi-instance network

After the link between Router3 and Router4 becomes faulty, Router3 starts to send LSAs to instruct the other devices in SEP segment 2 to update their LSA databases, and the blocked interface enters the Forwarding state. After the topology of SEP segment 2 is recalculated, the data from VLAN 201 to VLAN 400 is transmitted along path Router3->Router1->Router2.

After the link between Router3 and Router4 recovers, the devices in SEP segment 2 perform delayed preemption. After the preemption delay expires, P1 becomes the blocked interface again, and sends LSAs to instruct the other devices in SEP segment 2 to update their LSA databases. After the topology of SEP segment 2 is recalculated, the data from VLAN 201 to VLAN 400 is transmitted along path Router3->Router4.

Translation
Download
Updated: 2019-05-25

Document ID: EDOC1000097279

Views: 20090

Downloads: 64

Average rating:
This Document Applies to these Products
Related Documents
Related Version
Share
Previous Next