A real-time identification method for mobile applications based on cloud computing

Abstract

The invention discloses a real-time identification method for mobile applications based on cloud computing. The method builds a cluster composed of multiple computers, expands computing resources, and multiple classifiers identify in parallel without causing the entire cluster to collapse due to the collapse of one node in the cluster. The system crashes to improve the stability of the system; the captured data is forwarded in units of data packets instead of data streams, reducing the waste of system resources caused by waiting to receive data, and achieving the effect of real-time identification; adding new applications to be identified In the case of different types, there is no need to retrain the model, and the original recognition model can be used as usual. It only needs to collect data related to the new application, train the new recognition model, and run in parallel with the old model to obtain the recognition effect of the new application.

Description

technical field [0001] The invention relates to a method for real-time identification of mobile applications based on cloud computing. Background technique [0002] The traditional application identification technology mainly analyzes the load information and traffic statistics characteristics of data packets, and uses deep packet inspection and deep flow inspection technology for application identification. As the network environment becomes increasingly complex, these traffic identification technologies gradually fail. Machine learning methods have been used in network traffic classification in recent years, and their performance has been improved compared with traditional methods. However, due to the characteristics of shallow network structures, fine-grained application identification cannot be achieved. At the same time, machine learning is very dependent on the characteristics of traffic data. It consumes a lot of manpower and time for feature extraction. Therefore, ...

AI Technical Summary

Problems solved by technology

When all the current TCP connections are not over, no data flow is submitted to the data processing module, so the classifier module process is suspended, resulting in a waste of system computing resources, especially GPU resources

[0005] Second, a wireless network may have hundreds of terminals connected, and each terminal may use an application that generates a large amount of instantaneous traffic. At this time, the traffic flowing through the router may reach the maximum bandwidth, and the number of instantaneously generated data streams is large. When the classifier module currently has a recognition task, the subsequent received data stream must wait

When the waiting queue is long, the real-time performance cannot be guaranteed, and the computer memory will be exhausted and the classifier module will crash.

[0006] Existing technologies can only identify applications within a specified range. If a new application type needs to be added, the model must be retrained, resulting in unnecessary waste of time

Images