You are on page 1of 8

Table of Contents

1 IRF2.0 ··························································································································································1-1
Overview ·················································································································································1-1
Introduction······································································································································1-1
Advantages······································································································································1-1
Application ·······································································································································1-1
Basic Concepts ·······································································································································1-2
IRF Working Process ······························································································································1-3
Physical Connections ······················································································································1-4
Topology Collection ·························································································································1-4
Role Election ···································································································································1-4
IRF Management and Maintenance ································································································1-6

i
1 IRF2.0

H3C has been dedicated to the research and development of the IRF technology. After IRF1.0, H3C
has developed IRF2.0, which features more functions. This manual will briefly introduce IRF2.0.

Overview
Introduction

Intelligent Resilient Framework (IRF) is a software virtualization technology developed by H3C. IRF
allows you to connect multiple devices through physical IRF ports to combine them into a logical device
after necessary configurations. With this virtual device, you can manage and maintain multiple devices.
To facilitate description, the virtual device is called IRF. Therefore, IRF in this documentation has two
meanings: the IRF technology and the IRF device.

Advantages

IRF features the following advantages:


z Streamlined management: After an IRF is formed, you can log in to the IRF system by connecting
to a port of any IRF member to uniformly manage all the members in the IRF.
z High reliability: An IRF system comprises multiple member devices: the master runs, manages and
maintains the IRF, whereas the slaves process services as well as function as the backups. As
soon as the master fails, the IRF system elects a new master immediately to prevent service
interruption and implement 1:N backup. Besides, not only the physical IRF ports of members can
be aggregated, but also the physical links between the IRF system and the upper or lower layer
devices can be aggregated, and thus the reliability of the IRF system is increased through the link
backup.
z Powerful network expansion capability: By adding member devices, you can increase the number
of IRF ports, expand network bandwidth, and improve processing capability of the IRF system.

Application

Typically, you can deploy an IRF at the core layer, distribution layer and access layer. A master and
several slaves form an IRF. It is a single device to upper and lower layer devices, as shown in Figure
1-1.

1-1
Figure 1-1 IRF networking

IP network IP network

Master Slave Slave

Equals IRF

IRF cable IRF cable

Basic Concepts
IRF involves these concepts:

Operation mode

The device can operate in either of the following two modes:


z Standalone mode: The device operates in a standalone manner. It does not form any IRF with
other devices.
z IRF mode: The device interconnects with other devices to form an IRF.
You can use commands to switch between the two modes.

Role

The devices that form an IRF are called IRF members. Each of them plays either of the following two
roles:
z Master: Manages the entire IRF.
z Slave: All member devices except for the master. When the master fails, the system automatically
elects a new master from among the slaves.
Master and slaves are elected. An IRF has only one master at a time. Other members are the slaves.
For the detailed introduction to role election, refer to Role Election.

IRF port

It is an IRF-dedicated logical port, which can be numbered as IRF-Port1 or IRF-Port2. An IRF port is
effective only after it is bound with a physical IRF port.

1-2
In standalone mode, IRF ports are numbered IRF-Port1 or IRF-Port2; in IRF port, IRF ports are
numbered IRF-Portn/1 and IRF-Portn/2, where n is the member ID. For simplicity, IRF-Port1 and
IRF-Port2 are used in this manual.

Physical IRF ports

Physical IRF ports are physical ports used for IRF connection. Physical IRF ports can be IRF dedicated
ports, Ethernet ports or optical ports.
Typically, Ethernet ports or optical ports forward service packet. When they are bound to IRF ports, they
are used as physical IRF ports to forward packets between member devices. Packets that can be
forwarded include IRF negotiation packets and service packets between member devices.

IRF merge

As shown in Figure 1-2, two IRFs operate normally. Through physical connections and some necessary
configurations, a new IRF is established. This process is called IRF merge.
Figure 1-2 IRF merge diagram

IRF split

As shown in Figure 1-3, after an IRF is established, the failure of IRF cables causes physical
disconnection between two devices. The process is called IRF split.
Figure 1-3 IRF split diagram

Member priority

Member priority determines the role of a member device. The greater the priority value, the higher the
priority.

IRF Working Process


The life cycle of an IRF involves four stages: Physical Connections, Topology Collection, Role Election,
and IRF Management and Maintenance. Member devices in an IRF need to establish physical IRF

1-3
connections, and then automatically perform topology collection and role election. After this, the IRF
system runs normally to enter the IRF management and maintenance stage.

Physical Connections

To establish an IRF, you need to connect physical IRF ports following this rule:
The physical IRF port bound to IRF-Port1 on an IRF member can only connect to the physical IRF port
bound to IRF-Port2 on a neighbor device, and the physical IRF port bound to IRF-Port2 on the IRF
member can only connect to the physical IRF port bound to IRF-Port1 on the neighbor device, as shown
in Figure 1-4; otherwise, the IRF cannot be established.
Figure 1-4 Physical IRF connections

Topology Collection

Each member in an IRF records its known topology information, and exchanges hello packets with the
directly connected neighbors to collect topology of the entire IRF.
1) At the initiation of the device startup, the member device records only its own topology information.
2) When an IRF port of the member device becomes up, the member device sends its known
topology information from this port periodically.
3) Upon receiving the topology information from the directly connected neighbor, the member device
updates the local topology information.
After topology collection lasts for a period of time, all members have obtained the complete topology
information (known as topology convergence), and the IRF will enter the next stage: role election.

Role Election

The process of defining the role (master or slave) of IRF members is role election.
Role election is held when the topology is instable, such as, forming an IRF, adding a new member, IRF
split, or IRF merge. The master is elected according to the following principles one by one, until the only
winner is found out:
1) The current master wins, even if a new member has a higher priority. (When an IRF is being formed,
there is no master, and all member devices consider themselves as the master, so this principle will
be skipped)
2) A member with a higher priority wins.
3) A member with the longest system up-time wins.
4) A member with the lowest CPU MAC address wins.
After role election, the IRF is established and enters the next stage: IRF management and
maintenance.

1-4
During IRF mergence, IRF election is held, and role election rules are followed. Members of the loser
side will reboot and join the winner side as slaves.

After role election, an IRF is established:


z An IRF that combines box-type devices operates like a chassis-type distributed device. The master
operates as the active main board (AMB) of the IRF, and the slaves operate as the standby main
boards (SMBs) (also operate as interface boards), as shown in Figure 1-5.
Figure 1-5 Box-type device virtualization schematic diagram

Device A Device B
(MemberID=1) IRF- Port2 IRF- Port1 (MemberID=2)

IRF link
Service XGE1/3/0/1 XGE2/3/0/1 Service
interface Physical IRF Physical IRF interface
port port

After an IRF is
formed.
Suppose Device
A is elected as
the master.

IRF
Master
(MemberID=1)
AMB of the IRF

Slave
(MemberID=2)
SMB of the IRF

z An IRF virtualized by chassis-type distributed devices also operates like one chassis-type
distributed device. It has more standby main boards and interface boards. The active main board of
the master is the active main board of the IRF, and the standby main boards of the master, the
active main board and standby main boards of the slaves are the standby main boards (also
operate as interface boards) of the IRF, as shown in Figure 1-6.

1-5
Figure 1-6 Chassis-type device virtualization schematic diagram

IRF Management and Maintenance

After role election, an IRF is established. All member devices operate as a virtual device on the network,
and all resources on the member devices are processed and managed by this virtual device.

Member ID

An IRF uses member IDs to uniquely identify and manage member devices. The member ID of all the
member devices must be unique.
If member IDs of member devices are not unique, an IRF cannot be established. If a member wants to
join an IRF, but its member ID is the same as an existing one, the device cannot join the IRF. You can
use the following two methods to ensure the uniqueness of member IDs:Before establishing an IRF,
plan and configure member IDs for IRF members.Use the member ID collision processing
mechanism.

IRF topology maintenance

In an IRF, if a member becomes down or the IRF link is down, its neighbor broadcasts the information to
all the other members immediately. The member devices that receive the message determine whether
it is the master or a slave is down according to the local IRF topology information table. If the master is
down, a new election is triggered and the local IRF topology is updated; if a slave is down, the local IRF
topology is updated to ensure fast convergence of the IRF topology.

1-6
IRF split and multi-active detection mechanism

As described earlier, IRF link failures may cause an integral IRF system to split into two or more IRFs
operating with the same global configuration. Because these IRFs look the same to other devices in the
network, problems such as route flapping and STP calculation will result. To offset the risk, the
multi-active detection (MAD) mechanism is introduced. The MAD mechanism detects the presence of
multiple identical IRFs split from an IRF system and handles the problem to make the network operate
normally. The MAD mechanism provides the following functions:
z Detection: Enabled in an IRF system, the MAD mechanism periodically detects the IRF system for
multiple active IRFs with the same global configuration. This is done with the Link Aggregation
Control Protocol (LACP), the Bidirectional Forwarding Detection (BFD) protocol, or gratuitous ARP.
z Collision handling: If multiple identical active IRFs are detected, the MAD mechanism keeps only
the one with the lowest master ID (the active ID) to operate normally. The state of all the other IRFs
will be set to recovery and all the ports in them will be shut down except for the IRF ports and ports
manually specified not to shut down.
z Failure recovery: The MAD mechanism prompts MAD to users by outputting log messages. In this
case, the device tries to automatically repair the failed IRF links. If the repair fails, users need to
manually repair them. After a failed IRF link is recovered, the member devices in the recovery IRF
reboot to join the active IRF and all disabled ports come up to forward traffic.

1-7

You might also like