Analysis of link aggregation, Trunk, port binding and bundling

Source: Internet
Author: User
Tags switches

The three concepts of link aggregation (link Aggregation), trunk, and port binding/bundling are easy to confuse, with the main thrust being to provide reliability between servers, switches, and storage using link redundancy, or load balancing with redundant ports, often in the most scenarios, with indiscriminate mixing, Today we briefly analyze the meanings and differences of each.

Link aggregation is the use of different aggregation, merging methods to connect multiple network links in parallel to provide network bandwidth and link redundancy, to achieve network business continuity; Emphasis is placed on the end-to-end link of the switch, typically involving ports, network cards, and switches (such as LACP) bindings. Different switch vendors describe link aggregation in different terms, but the target is consistent, and two or more ports are bound together as a high-bandwidth logical port to increase link speed, redundancy, resiliency, and load balancing. Cisco is called the Ether Channel, brocade is called the Brocade LAG, and is based on the standard IEEE 802.3ad LACP (Link Aggregation Control Protocol). The protocol was transferred to IEEE 802.1ax as a standard in 2008, and LACP can dynamically configure port aggregation without relying on any vendor, so most Ethernet switches support the protocol.

Typically, a parallel aggregation/binding is created between switches using multiple ports, and the Spanning Tree Protocol (STP) treats it as a loop, and technically, the link aggregation technique solves the problem by enabling the aggregation port to generate a single logical link (logical link) without causing the loop.

Trunk technology is typically used between switches to achieve reliability through redundant links, increasing the number of ports through cascading, and providing ISL (inter-switch link) Link utilization through trunks. Generally divided into the following several scenarios.

ISL Trunk Technology:

refers to the link relay between switches, generally used to increase the number of link ports, cascade and long-distance transmission (generally referred to as FC switches, using single-mode fiber can achieve long-distance transmission).

VSAN Trunk Technology:

is a Cisco FC Switch-specific concept, meaning that multiple Vsan traffic can be hosted via an ISL link, and Vsan resembles a VLAN with an IP switch, enabling two-layer isolated secure access. Brocade FC switches are also known as zoning for similar implementations of isolation.

Vlan/ethernettrunk Technology:

Refers to the VLAN tag tag, to connect the host in the same VLAN on different switches, set the cascade port of the two switches to trunk port, when the switch sends the packet from the Cascade port, it will make a token in the packet, So that other switches can identify which VLAN the packet belongs to, when the other switch receives the packet, it forwards the packet to the VLAN specified in the token, thus completing the VLAN internal data transfer across the switch. Hosting multiple VLAN data on one inter-switch link ISL increases link utilization without having to provide a separate ISL link between the switches for each VLAN ID.

Commonly used protocols have the standard 802.1Q and Cisco ISL Protocol, ISL is a Cisco private protocol that can only be used between Cisco's own switches.

Vlan/ethernettrunk Simple Summary:
1.VLAN in the real world is not just a single switch, but multiple switches work together. Each switch divides multiple VLANs, and these VLANs may have the same VLAN ID on multiple switches.
2. Several VLANs from different switches connect to a physical link that only needs to use this trunk link to host all VLAN traffic.
3. The switch link type has access link and relay link points. A relay link can host more than one VLAN.

Link Bundle/Port bonding:

Generally refers to the server port, storage Ethernet port through the bonding technology to achieve link redundancy and load balancing, the current Linux operating system core itself has supported 7 kinds of binding mode, and some binding mode (such as LACP) recommended with the Ethernet switch LACP protocol used.

BALANCE-RR (Round-robin) Mode

Also known as Bound 0 mode, all network card ports of the same priority are maintained in a cyclic queue (slave device chain list), and the bonding device drivers are selected in turn in these network card ports. Once a network card port completes a data transfer task, the bonding driver automatically dispatches the next NIC Port in the queue to complete the data Transfer task for the next packet. Send the data in such a continuous, repetitive way. The newly added bonding NIC Port will be placed at the tail of the queue.

active-backup (AP) Mode

Also known as bound 1 mode, one port in multiple NIC ports is set to active, and the other ports are in standby state. Initiates an alternate link when the active port or active link fails.

Balance-xor (XOR) Mode

Also known as Bound 2 mode, the packet is transmitted based on the specified transfer hash policy. The default policy is: (Source MAC address XOR destination MAC address)% slave number. Other transport policies can be specified by the "xmit_hash_policy" option.

Broadcast ( broadcast) mode

Also known as Bound 3 mode, through the bonding device in all the network card port to send the data to be transmitted, the two networks are completely independent, if the need for certain one-way traffic can reach the two independent network at the same time, need to use "broadcast" mode.

802.3ad (LACP) Mode

Also known as Bound 4 mode, dynamic Link aggregation strategy. Allow multiple network card ports to work under the same active aggregate, sharing the same rate and duplex mode. If a peer-to-peer Ethernet switch supports the port aggregation of LACP aggregation mode, it is recommended to use "802.3AD" mode to configure the bonding feature.

balance-tlb (TLB) Mode

Also known as bound 5 mode, adaptive transport load balancing strategy. Based on the different processing power and current load state of each bonding network Kadangkou (slave), the bonding driver distributes the Network Service request reasonably, distributing the load of network transmission in all NIC ports.

balance-alb (ALB) Mode

Also known as bound 6 mode, adaptive load balancing strategy. The BALANCE-ALB mode increases receive load balancing for IPV4 traffic (receive load balance), based on the BALANCE-TLB mode providing the transport load balancing feature. Based on the different processing power and current load state of each bonding network Kadangkou (slave), the bonding driver distributes the Network Service request reasonably, distributing the load of network transmit and transmit in all NIC ports.


This article from "10985768" blog, declined reprint!

Analysis of link aggregation, Trunk, port binding and bundling

Contact Us

The content source of this page is from Internet, which doesn't represent Alibaba Cloud's opinion; products and services mentioned on that page don't have any relationship with Alibaba Cloud. If the content of the page makes you feel confusing, please write us an email, we will handle the problem within 5 days after receiving your email.

If you find any instances of plagiarism from the community, please send an email to: info-contact@alibabacloud.com and provide relevant evidence. A staff member will contact you within 5 working days.

A Free Trial That Lets You Build Big!

Start building with 50+ products and up to 12 months usage for Elastic Compute Service

  • Sales Support

    1 on 1 presale consultation

  • After-Sales Support

    24/7 Technical Support 6 Free Tickets per Quarter Faster Response

  • Alibaba Cloud offers highly flexible support services tailored to meet your exact needs.