A traffic collection method for network or application performance management

Abstract

The embodiment of the invention provides a flow collection method for network or application performance management. The flow collection method comprises the following steps: a collection unit carriesout network packet monitoring on a target virtual network card in a shared memory by calling a packet grabbing socket interface of an operating system kernel; If the collecting unit detects that there are newly added network packets in the shared memory, the collecting unit collects the new network packets and completes the flow collection; The shared memory is the memory shared by the collectionunit and the operating system kernel, and the newly added network packets are the network packets sent or received by the target virtual machine corresponding to the target virtual network card. A method for collecting traffic for network or application performance management provided by an embodiment of that present invention carries out network packet monitoring on a target virtual network cardin a shared memory by setting a collection unit, if a newly added network packet is detected, the newly added network packet is acquired, the high-efficient flow collection can be carried out, a reliable data source is provided for a back-end analysis tool, and the flow of mutual visit can be completely collected without changing the virtual machine.

Description

technical field [0001] Embodiments of the present invention relate to the field of computer technology, and more specifically, to a traffic collection method for network or application performance management. Background technique [0002] Network Performance Management (NPM, Network Performance Management) combines the visibility of end-to-end business and application delivery dependencies with network behavior analysis to solve the challenges of finding, locating, and diagnosing faults in network operation and maintenance management. NPM can analyze the behavior of TCP or higher layers on each hop path of the application or all application behaviors based on transport layer protocols such as TCP on a core node, and find and locate problems in time when a fault occurs, so as to provide IT's operation and maintenance guarantee provides an "ultimate means" that can discover network and application faults in real time, drill down to a single session, and access all raw data. A...

AI Technical Summary

Problems solved by technology

[0006] The first is the functional problem of collecting on the switch: when using the second method to collect traffic, because there is direct network access between the virtual machines on the virtual server, the way of mirroring and copying traffic on the switch cannot completely collect the internal traffic of the server traffic

[0007] The second is the intrusive problem of collecting traffic inside the virtual machine: when using the first method to collect traffic, installing sFlow or tcpdump inside the virtual machine can collect all the traffic, but this method needs to change all virtual machines, and there is a software A series of problems such as installed version conflicts, permission conflicts to access virtual machines, and large-scale deployments that are difficult to manage

Images