Load sharing method and system of link aggregation

A link aggregation and load sharing technology, applied in transmission systems, digital transmission systems, electrical components, etc., can solve problems such as limited resource utilization, insufficient link utilization, and small number of sessions, to achieve intelligent dynamic load sharing , Improve the effect that the load varies greatly and reduce the computational burden

Active Publication Date: 2019-03-29
FENGHUO COMM SCI & TECH CO LTD
View PDF10 Cites 15 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, some data center network traffic measurement studies have shown that the number of sessions in the data center network is small, the rate difference is large, and the traffic changes rapidly.
[0006] If the link aggregation of data center switches performs load sharing according to the traditional hash algorithm, the traffic load will be unbalanced due to hash collisions, and the loads of different physical links will vary greatly. rate is not enough
The actual throughput is less than the sum of the bandwidth of all physical links, which limits resource utilization and seriously affects the performance of Ethernet in the data center network

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Load sharing method and system of link aggregation
  • Load sharing method and system of link aggregation
  • Load sharing method and system of link aggregation

Examples

Experimental program
Comparison scheme
Effect test

no. 1 example

[0038] This embodiment provides a load sharing method for link aggregation, including steps:

[0039] Allocate forwarding ports for large flow sessions, and adjust ports according to the average bandwidth ratio difference of large flow sessions on each port; at the same time, adjust port forwarding weights according to the bandwidth utilization of each port in link aggregation.

[0040] Distinguish the received data packets, forward large-flow sessions according to the allocated ports, and schedule other sessions according to the forwarding weight of each port, so that the session flow in each port reaches the forwarding weight, and realizes the load balancing of each port.

no. 2 example

[0042] Based on the first embodiment, set a large flow session forwarding table, periodically judge the forwarding traffic of the data packet, determine the data packet exceeding the large flow session threshold as a large flow session, add it to the large flow session forwarding table, and distribute forwarded port.

[0043] Further, the threshold of a large flow session can be 10MB, and every 5 seconds is used as a sampling period to judge the forwarding traffic of the data packet. If the forwarding traffic exceeds 10MB within 5 seconds, it is determined that the corresponding data flow is a large flow session. Create a flow table entry in the publication, and assign the corresponding forwarding port to the created entry, so as to facilitate the forwarding of the large flow session.

[0044] Preferably, the allocated ports can be adjusted according to the large flow session forwarding table. By periodically reading the forwarding information of the large-flow session in the l...

no. 3 example

[0047] Based on the second embodiment, the allocated ports are adjusted according to the difference in the average bandwidth ratio of the large flow sessions on each port. Calculate the average bandwidth ratio of large flow sessions in each port, and when the difference in the ratio of each port exceeds the bandwidth ratio threshold, adjust part of the data flow in the port with the largest bandwidth ratio to the port with the smallest bandwidth ratio.

[0048] Further, if the cycle is 1 second, the average bandwidth ratio of the large flow session in each port is obtained by calculating the forwarding information of the large flow session in the past 3 cycles.

[0049] The average bandwidth ratio of the large flow session on the port = the average number of bytes of data packets forwarded by the large flow session forwarded by each port in the past 3 cycles × 8 / port forwarding capability;

[0050] Taking the bandwidth ratio threshold as 10% as an example, if the difference ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a load sharing method and system of link aggregation, and relates to the link field of data communication technology. The system includes a control plane and a forwarding plane. Forwarding ports are assigned by the control plane for large-flow sessions, port adjustment is carried out according to average bandwidth ratio difference of the large-flow sessions of all the ports, and port forwarding weights are adjusted at the same time according to bandwidth utilization rates of all the ports of link aggregation. The received large-flow sessions are forwarded by the forwarding plane according to the assigned ports, and load balancing is carried out on remaining sessions according to the forwarding weights of all the ports. The system can avoid network congestion and load difference, improve resource utilization rates, and realize load balancing of a data center network.

Description

technical field [0001] The invention relates to the link field of data communication technology, in particular to a load sharing method and system for link aggregation. Background technique [0002] Compared with the traditional network, the east-west traffic in the data center is relatively large. With the current use of cloud computing and distributed technology, as well as the widespread deployment of virtualization, the interaction between servers in the data center is becoming more and more frequent. The traffic between the network is increasing, making the network one of the bottlenecks of data center performance. The flow of most flows in the data center is less than 10KB and the duration is very short, all in the range of hundreds of milliseconds, and the number of active flows in each rack of the entire data center is less than 10,000 per second; regardless of the specific type of application What is, in most data centers, the link utilization of every layer except...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): H04L12/709H04L12/803H04L45/243
CPCH04L45/245H04L47/125
Inventor 汪学舜董喜明
Owner FENGHUO COMM SCI & TECH CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap