VRRP Protocol: redundant backup with high reliability

VRRP applies to layer-3 switches and routers on the default gateway for static configuration, provides faster and more effective redundancy and fault tolerance capabilities for terminal systems that rely on the default gateway for Wan access or access to other LAN domains.

With the increasing popularity of the Internet and Intranet, people are increasingly dependent on the network. Restoration and redundancy have become a key feature in today's LAN. To this end, IETF has developed the Virtual Router Redundancy Protocol, VRRP), which is applied to layer-3 switches and routers on the default gateway for static configuration, provides faster and more effective redundancy and fault tolerance capabilities for terminal systems that rely on the default gateway for Wan access or access to other LAN domains.

VRRP introduced to the Exchange Network

In recent years, with the continuous improvement of switch layer-3 switching capabilities, VRRP technology is no longer limited to the use of routers, but is gradually introduced to large-scale exchange networks, in order to achieve high-reliability redundant backup for layer-3 switching.

VRRP combines multiple routers in the system into a VRRP group, which has a virtual default gateway address. But at any time, only the active router Master that controls the vro address in a group is used to forward data packets. If the active router fails, it will select a redundant Backup router with the highest priority) to replace the active router. The VRRP virtual gateway address is configured for the terminal in the network. When a fault occurs, the virtual router does not change and the host remains connected. The network will not be affected by single point of failure, in this way, the router switching problem in the network is better solved.

　
Topological Structure of VRRP Exchange Network

A typical vswitch network topology with VRRP technology is shown in Figure 1.

Generation of multicast traffic

In a large-scale exchange network, because multiple VLANs exist, multiple VRRP gateways need to be configured. These gateways will continuously send multicast packets with a default latency of one second. In addition, when the network topology in the Spanning Tree changes, the Bridge Protocol Data Unit BPDU with the topology changes will be activated. When applications such as on-demand video streaming and Multimedia conferences have a large volume of traffic, and the data generated by the network itself, it will cause a local network congestion at a certain time point ", it has even become the cause of a multicast storm.

In the network shown in, a large number of application data flows in the network preemptible the bandwidth on the link, increasing the processing pressure on the core switch, making it impossible for the trunk port to forward VRRP packets and BPDU packets in a timely manner, as A result, Switch-A does not receive the forwarding delay of the BPDU packet of Switch-B for 15 seconds. Switch-A Switch port enters the learning forwarding status from blocking. At the same time, Switch-A sends the topology change BPDU packet to Switch-B to notify Switch-B that the network topology changes. In addition, the VRRP status of Switch-A changes from Backup to Master, send VRRP multicast to all forwarding ports.

When the multicast arrives at the access layer Switch, the access layer Switch forwards data to Switch-B. After receiving the multicast packet, Switch-B forwards the packet to Switch-A through the interconnect port, forming A loop.

Achieve Network Optimization

Increase the bandwidth of the core switch interconnection link

The interconnection links between core switches have a large amount of broadcast and multicast traffic. If the link is blocked due to insufficient bandwidth, the VRRP backup switch will become the Master, in this way, there will be two masters in the network. Therefore, the normal link bandwidth is guaranteed by the normal VRRP network.

Modify VRRP notification interval

The default notification interval of VRRP is one second. If this time is prolonged, the broadcast volume of VRRP on each port is doubled. However, because the VRRP announcement interval is prolonged, the master-slave switchover time in the VRRP network will be extended, making it more difficult for the core switch to detect changes in the network topology. Therefore, you cannot simply extend the time, but make appropriate adjustments based on the actual situation of the network.

Set STP
The Spanning Tree Protocol uses priority settings to determine who is the root switch. Ensure the status of the root switch and Master route switch until the network topology changes. In addition, STP can adopt DUAL-LAY and Per-VLAN modes.

Trim Trunk ports
Trim the Trunk port, from binding all VLANs to binding only some VLANs that can be received by this port, reducing the calculation workload of the switch and the traffic flow on the port.

　
Multicast storm Formation Process

Access layer switch port Redundancy

The access layer Switches of Cisco, Avaya, and 3Com all have port redundancy, or are called elastic links. On the port of the switch, through proprietary technology of the manufacturer, implements hardware failover. When the main link is interrupted, the backup link will be 1 ~ Switch to the working status within 5 seconds. This technology does not require tree generation protocol operations, and does not have BPDU overhead. The switching time is much lower than the convergence time of the spanning tree. The default convergence time of the Spanning Tree is 50 seconds, as the network topology becomes more complex, the convergence time will also increase ). Therefore, the access layer switch is suitable for implementing redundant link backup in this simple and easy-to-use mode.

Setting VRRP virtual port priority

We recommend that you set the virtual port priority of VRRP on the core switch that generates the root bridge to the highest level to ensure the consistency between the VRRP master router and the generated root bridge, prevent VRRP status fluctuations due to network topology changes.

Disable unnecessary broadcast and Multicast Services

The network has a lot of broadcast and multicast business applications, some are meaningless or unnecessary, such as the internal network NTP service, ARP-TX services, etc, this can save more bandwidth.

Upgrade vswitch Software

The upgrade of the vswitch software version can add new functions to the vswitch and improve the original algorithm. Therefore, it is recommended that network engineering maintenance personnel upgrade the switch software frequently to speed up the switch and improve the performance of the switch.

Although VRRP networks cannot completely eliminate the existence of multicast traffic, network engineers control the traffic and customize their own solutions for different network environments, so that it does not affect the normal operation of the network, it will provide faster and more effective redundancy and fault tolerance capabilities for the network.