Delayed copy incremental container check point processing method based on dirty page prediction

Abstract

The invention discloses a delayed copy incremental container check point processing method based on dirty page prediction which comprises the steps of predicting a next round of dirty page modification probability in advance according to a historical record of a container check point execution process, and performing progressive increase sorting on the dirty page probability; carrying out delayedcopying on pages with the dirty page degree values larger than a set threshold value in each check point operation, and the pages with the low dirty page degree values are copied preferentially; and designing a multi-level check point storage strategy, and performing corresponding check point storage operation according to the check point requirement of each level. The method aims at solving the problem of cloud computing container check point processing performance. A delayed copy increment check point algorithm based on dirty page degree prediction is designed. Delay copying is carried out according to the dirty page degree ascending order, frequent copying and transmission of dirty pages can be effectively avoided, the size of a container check point file can be effectively reduced, application downtime is shortened, resource expenditure is saved, and the low service level agreement violation rate is guaranteed.

Description

technical field [0001] The invention belongs to the technical field of computers, and more specifically relates to a delayed copy incremental container checkpoint processing method based on dirty page prediction. Background technique [0002] Cloud computing has the characteristics of elastic computing and pooled resources. It can provide basic resources such as computing, network, and storage for applications on demand, and supports rapid expansion of resources, which has been recognized by the majority of users. Compared with traditional virtual machines, containers have the advantages of fast startup speed, light weight, good isolation, convenient migration and flexible deployment, which are helpful for business automation of cloud computing data centers, shortened deployment cycles, reduced costs and rapid resource utilization. Provide cloud service applications with support services such as rapid resource elastic scaling and performance optimization. [0003] In a larg...

AI Technical Summary

Problems solved by technology

[0006] According to the above analysis, it can be seen that although the existing container checkpoint method can solve many fault problems, it still has the following deficiencies: 1) when performing the checkpoint operation, the application has a long downtime due to the frequent copying of the checkpoint page; 2) The local storage method adopted by the single-level checkpoint has few types of tolerant faults and a single recovery path, resulting in unguaranteed application reliability; 3) Although the multi-level checkpoint can tolerate multiple types of faults, it has high reliability , but due to the large checkpoint file, a large additional overhead is generated during the transmission process

Images