Memory optimization method and device based on physical memory in virtual environment

Abstract

The invention provides a memory optimization scheme based on physical memory in a virtual environment. The Application sends the first request message to the MM_Agent_Client for querying the physical memory characteristics of the Host. MM_Agent_Client forwards the first request message to MM_Agent_Server. Based on the first request message, the MM_Agent_Server acquires the physical memory characteristics of the Host, and sends it to the MM_Agent_Client through the first response message. MM_Agent_Client forwards the first response message to Application. Application creates a page table and applies for physical memory pages according to the required memory size. The Application sends a second request message to the MM_Agent_Client for querying information about the requested physical memory page. MM_Agent_Client forwards the second request message to MM_Agent_Server. Based on the second request message, the MM_Agent_Server obtains the relevant information of the physical memory page requested by the Application, and sends it to the MM_Agent_Client through the second response message. MM_Agent_Client forwards the second response message to Application. Application creates a memory object pool, and performs memory optimization based on physical memory characteristics and information about physical memory pages.

Description

technical field [0001] The present application relates to a communication system, in particular to a physical memory-based memory optimization method and device in a virtual environment. Background technique [0002] Memory optimization is very important for traditional "IP network forwarding", which is closely related to specific hardware. Application developers need to be aware of the hardware configuration, distribution, and characteristics of physical memory, so as to optimize applications based on parameters such as the number of channels and ranks of physical memory. Memory optimization can greatly improve application performance. [0003] In a virtualized cloud environment, an application program runs on a guest operating system (Guest OS) in a virtual machine (virtual machine, VM). The real physical memory is invisible to the Guest OS. Application programs can only access virtual physical memory virtualized by a virtual machine monitor (virtual machine monitor, VM...

AI Technical Summary

Problems solved by technology

For example, continuous memory allocation in a virtual machine may become discontinuous after mapping, making the program run inefficiently

[0005] In addition, in a virtualized cloud environment, when a virtual machine is migrated offline, the physical memory structure of the host machine (host machine) may change, such as from NUMA to non-NUMA architecture. Will cause the memory optimization applied in the original virtual machine to fail

Images