Data Packet Processing via Flow Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In data networks, the increasing scale and user demand for flexibility lead to network performance degradation due to processing bottlenecks at both switch and server sides, particularly caused by high CPU load from processing large data packets, known as elephant flows, while small flows, or mice flows, consume less CPU resources but are numerous, resulting in inefficient resource utilization.
Innovation Solution
A method and system that differentiate between elephant flows and mice flows based on a threshold of data packets, where mice flows are processed at a virtual switch and elephant flows are transmitted to a hardware switch for processing, reducing CPU requirements at the server side and avoiding forwarding delays.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If all flows are processed at the virtual switch on server side, then flexibility and user isolation are ensured, but CPU processing load increases causing network performance degradation
Solution Approach 1:
The patent segments flows into two categories: mice flows (small data packets) and elephant flows (large data packets). Mice flows are processed at the virtual switch on the server side to maintain flexibility and user isolation, while elephant flows are offloaded to hardware switches in the network for processing. This segmentation resolves the contradiction by distributing processing responsibilities according to flow characteristics, reducing server CPU load while preserving the benefits of virtualization.
2Reliability
If elephant flows are processed at the virtual switch, then user isolation and security are maintained, but CPU bottleneck occurs causing forwarding delay
Solution Approach 1:
The patent extracts elephant flows (large data packet flows) from the virtual switch processing queue and redirects them to hardware switches in the network for processing. This extraction removes the time-consuming processing of large flows from the server CPU, eliminating the bottleneck that causes forwarding delays, while mice flows continue to be processed at the virtual switch to maintain user isolation and security.
3Loss of time
If more CPU resources are allocated for processing large flows, then forwarding delay is reduced, but resource utilization efficiency decreases
Solution Approach 1:
The patent introduces hardware switches as intermediaries to handle elephant flows. Instead of allocating more CPU resources at the server, the system uses hardware switches in the network as mediators to process large data packet flows. This approach reduces forwarding delay by offloading processing to dedicated hardware while maintaining high resource utilization efficiency at the server, as CPU resources are reserved for control plane operations and mice flow processing.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method and system for processing data packets in a data network have been disclosed. The method comprising: dividing flows containing data packets into elephant flows and mice flows according to a particular division parameter, wherein, an elephant flow contains more data packets than a mice flow; transmitting the divided elephant flows and mice flows to a virtual switch in a server; processing the mice flows at the virtual switch, and transmitting the elephant flows and the processed mice flows to a hardware switch in network; and processing the elephant flows at the hardware switch, and forwarding the received mice flows and the processed elephant flows. With the method and system, network performance may be effectively improved.