A method for selecting cloud service performance adaptive action types based on deep learning

Abstract

The present invention provides a method for selecting an adaptive action type of cloud service performance based on deep learning, including monitoring physical cluster data, virtual machine data, and service component data; combining constraint events agreed in SLA and real-time monitoring data to determine current service performance Whether optimization is needed: If the current data triggers a constraint event, the adaptive action type will be determined according to the adaptive method library, otherwise continue monitoring; cloud service performance self-optimization will be performed according to the determined adaptive action type; feedback learning, update the adaptive method library, Return to continue monitoring. The cloud environment itself has the characteristics of high scalability and dynamic reconstruction, so that the service performance of the cloud service is restricted by many factors in the actual operating environment. The best adaptive action is selected from the adaptive action set.

Description

technical field [0001] The invention belongs to the technical field of cloud services, and in particular relates to a method for selecting a cloud service performance adaptive action type based on deep learning. Background technique [0002] In the era of big data, the exponential growth of information has directly led to the explosion of data in many industries. As a result, the entire society has undergone tremendous changes in the way of life and production. The technology that supports this technological change The foundation is thanks to the various "cloud" services offered by current cloud providers and cloud service providers. Cloud computing technology (Cloud Computing) is the use of high-speed Internet transmission capabilities to transfer data processing from personal computers or servers to computer clusters on the Internet. Cloud computing is an emerging business computing model, which distributes computing tasks on a resource pool composed of a large number of ...

AI Technical Summary

Problems solved by technology

[0011] However, the solutions generated based on a single optimization technology are difficult to meet the diversification of service environments and service pressures faced by existing cloud service components

These research works are either based on the correspondence between performance and resources, using mathematical and information theory methods for online modeling, or using statistical learning or machine learning methods to analyze experimental data or historical logs and perform offline modeling. Research on the impact of resource change factors such as resource competition and performance interference among the service providers on service performance. The feature extraction of resource change indicators that affect service performance is not comprehensive enough, and the accuracy of some parameter values ​​​​in the performance model is not enough.

Although some scholars have also adopted neural network technology to deal with the problem of nonlinear diversification of cloud services, the efficiency of traditional methods is not ideal for multi-dimensional deep data structures.

[0012] Deep learning technology is one of the learning technologies with the highest learning efficiency in the current field of machine learning. However, this algorithm is currently mostly used in multi-dimensional and single-characteristic data fields such as image recognition and audio processing, and is rarely applied in multi-dimensional and complex data fields.

Images