GLBP: more balanced LAN

With the GLBP protocol, Server Load balancer for key applications and redundant routing backup can be achieved without changing the network structure without more configuration. This maximizes user investment, it can greatly improve network performance with minimal management costs.

GLBP: more balanced LAN

With the continuous development of network applications, users are increasingly demanding on network reliability. Once the routing between devices in two locations fails, the network will be interrupted if redundant devices are not enabled. In this case, packets hosted on the network will be lost. If the data sent at this time is important to the company's business, it will inevitably bring certain economic losses to the company. How can this problem be avoided?

If only a redundant router is physically connected, when a primary route fails, the host still sends packets to the default gateway configured in advance for the destination address to be packets from other networks, however, it is not possible for a router to automatically switch under a fault.

The selection of routes also directly affects the network throughput. If the primary route resources are exhausted, it cannot continue as the forwarding information of the relay node, resulting in the network failure. At this time, there may be a lot of remaining resources for backup routing, which leads to unequal network energy consumption.

This requires a technology that can achieve stable switching between the master and slave routes, and consider the load balancing technology. The Gateway SLB protocol GLBP came into being.

What is GLBP?

GLBP stands for Gateway Load Banancing Protocol, which is a proprietary Protocol of Cisco. GLBP can bind up to four MAC addresses to one virtual IP address, allowing the client to use the same virtual IP address as the gateway address. After the client sends an ARP request, it responds to different target MAC addresses and forwards packets through different routers. Therefore, it plays a role in load balancing to some extent. This is different from HRSP and VRRP. An active router must be selected for both HRSP and VRRP. The active router forwards data packets on behalf of the vro address, while the Standby Router is idle. While GLBP provides virtual routers, all the routers in the GLBP group can forward some data streams and participate in load balancing. It can be seen that GLBP can fully utilize network resources without excessive configuration and management.

Active virtual gateway election

The election of active virtual gateways uses an election mechanism similar to HRSP to elect active gateways. The routers with the highest priority become Active Routers. If the highest priority is the same, the routers with the highest IP address become active routers. Active Routers are called AVGAcitve Virtual gateways. Other non-AVG routers provide redundancy. If the AVG fails, a new election will take place. AVFActive Virtual Forwarder ). Both AVG and AVF form GLBP group members. Each GLBP group can have up to four members.

Virtual MAC Address Allocation

GLBP automatically manages virtual MAC Address allocation. If a vro is elected as AVG, AVG begins to distribute the virtual MAC address to AVF in sequence. AVF is divided into PVFPrimary Virtual Forwarder) and SVFSecondary Virtual Forwarder ). A vro directly assigned a virtual MAC address by AVG is called PVF. A group member who does not know the real IP address of AVG can only use hellos packets to identify the vro and then assign a virtual MAC address, this class is called SVF. PVF's virtual MAC consists of a fixed MAC prefix + group number. GLBP can have up to four routers as the default IP Gateway. The virtual MAC addresses of each gateway are PVF address numbers plus 1. For example, if the PVF virtual MAC address is 0007. b40020.a01, the first SVF is 0007. b40020.a02, and so on. After a virtual MAC address is assigned, all GLBP group members are involved in packet forwarding, but each member is only responsible for forwarding packets related to the virtual MAC address assigned to him.

GLBP configuration Verification

Configure the relevant GLBP commands on the specified ports of the GLBP group members and set different priorities for different members. A router with a higher priority is an active router. Its status value is active, and other routers are backup routers. Its status value is standby. The active vro automatically assigns a virtual MAC address to all group members. In this way, each group member obtains the virtual MAC address of all members, but the status of the virtual MAC address varies under each vro. If the status value is active, the MAC address is the active MAC address of the vro and forwards related packets. If the status value is listen, the MAC address is in the listening status, once a fault is detected on another vro and cannot be forwarded, The listen status is automatically changed to active, and the data forwarding function of the MAC address is taken over for redundancy. Under normal circumstances, each group member is only responsible for forwarding packets whose MAC address status is Active.

Use GLBP for Load Balancing

GLBP assigns multiple switches or routers to the same GLBP group. GLBP automatically manages and assigns different virtual MAC addresses to the selected AVG. Each GLBP group can have up to four virtual MAC addresses. When the client sends an ARP request to query the vgateway address, AVG responds to all ARP requests related to the vgateway address and determines which MAC address to return to the client based on the server Load balancer algorithm, therefore, the MAC addresses obtained by the client are different. GLBP implements network load balancing by using different virtual MAC addresses in ARP responses.

This method has two advantages: first, the client does not need to point to the physical MAC address of the redundant route, and the default gateways of all clients point to the unique virtual router IP address, implements load balancing. Second, even if a router fails, other routers in the GLBP group can immediately take over the virtual MAC address of the faulty router, without affecting the transmission of client data.

GLBP Enterprise Network practice

Application of the Server Load balancer Function

In a CEN, the core layer originally had only one Cisco 6509, and a large amount of data between networks is exchanged through the core device. It also undertakes the routing function between different VLANs. Once a device in the core layer fails, the entire network is paralyzed. To further improve the reliability of the core layer network and prevent the entire network from being paralyzed due to device faults at the core layer, we have added a Cisco 6509 core to form a dual-core network structure with the original core equipment, the GLBP protocol is run between core switches. nodes at the access layer, such as the office network, aggregation of secondary units, and finance, are connected to the core devices through redundant links, and routing redundancy is achieved through GLBP, ensures high reliability of the core layer of the network.

Based on the design topology, we divide different GLBP groups for different VLANs. The detailed plan is shown in the table:

Take VLAN2 as an example. The office network is connected to two Cisco 6509 VLAN2 ports, which are 10.192.2.252 and 10.192.2.253, respectively, in VLAN2, each host gateway is set to the VIP address 10.192.2.254 of the glbp 20 group. The host in VLAN2 successfully accesses the Headquarters and the Internet through this VIP address. In this case, GLBP takes effect. The host receives different response addresses through ARP requests. The core switch processes data based on their active MAC addresses, we observe the GLBP running status on two different hosts. By running the arp -- a command on the host, the two hosts direct to the same gateway IP address, but the MAC address is different, indicating that the two hosts reach the destination address through different routes. According to this, the enterprise network achieves Load Balancing through GLBP.

Implementation of the routing redundancy Function

Taking VLAN2 as an example, we dropped one of the core Cisco 6509 VLAN2 ports to simulate a fault in the core route. Then observe the two hosts, and find that both hosts can still access the external network, and the user does not feel any changes. Then use the arp -- a command to observe that there are still two different MAC addresses, why? Show gblp brief on the Cisco 6509 instance in normal operation. We can see that this switch has detected another switch failure, change the MAC address in the listening status to active. That is to say, the vswitch has two MAC addresses and forwards data for requests sent to these two MAC addresses, all these changes are transparent to users, so users do not feel the network interruption, so they can smoothly and quickly implement the GLBP routing redundancy function.

GLBP enterprise network planning table

Application of GLBP in Enterprise Networks

GLBP route Redundancy

The GLBP protocol can achieve load balancing of key applications and redundant routing without changing the network structure, this maximizes user investment and greatly improves network performance with the minimum management cost. In large enterprise networks, GLBP technology has great application value to achieve balanced load of key equipment.

1. Common lan network faults and troubleshooting strategies
2. Future trends: 802.11v network protocol analysis