System and method for realizing swap-out and swap-in of memory

Abstract

The invention discloses a system for realizing swap-out and swap-in of a memory. The system is provided with a redirection layer; when a memory page needs to be swapped out in a process corresponding to a specific container in a system, the redirection layer provides a position, which is set in an external storage device and corresponds to a specified virtual memory space of the container, for the memory page, so that the memory page needing to be swapped out is stored in the external storage device of the specified virtual memory space; and when the process of the specific container needs to be swapped into the memory page, the redirection layer provides a position, which is set in an external storage device and corresponds to the specified virtual memory space of the container, for the specific container, so that the process in the container can be swapped into the memory page corresponding to the process of the specific container. By adopting the system, storage isolation according to the granularity of the container is realized.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a system for implementing memory swap-out-swap-in. The present application also relates to a method, device, electronic device and storage device for realizing swapping out and swapping in memory. Background technique [0002] At present, in the cloud computing environment, especially in the serverless (serverless) mode, capacity expansion and contraction are often required to cope with changes in service request traffic. The so-called expansion and reduction refers to the increase of the running instance of the application from one to multiple, or the reduction from multiple copies to fewer copies. The ability to scale up and down is called elasticity. The rapid expansion and contraction scenario of the serverless mode is an elastic scenario. In a serverless elastic scenario, in order to shorten the service response time, multiple running instances of applications...

AI Technical Summary

Problems solved by technology

However, the Serverless mode allocates resources and deploys at the container granularity. The implementation of container technology relies on CGroups. The actual resource allocation for running instances is allocated with CGroups as the granularity. , the storage of memory pages swapped out to the swap device will become randomized, and this randomization will affect the performance of the next swap in

Therefore, using the page as the granularity cannot meet the memory swap-out-swap-in requirements in the serverless mode. It is very important to provide a container-granularity isolated storage to achieve a larger-grained memory swap-out-swap-in solution.

Images