WO2022120992A1

WO2022120992A1 - Virtual-environment-based memory sharing system and method

Info

Publication number: WO2022120992A1
Application number: PCT/CN2020/139462
Authority: WO
Inventors: 李永康; 王洋; 须成忠
Original assignee: 中国科学院深圳先进技术研究院
Priority date: 2020-12-10
Filing date: 2020-12-25
Publication date: 2022-06-16
Also published as: CN112783667A; CN112783667B

Abstract

Disclosed are a virtual-environment-based memory sharing system and method. The system comprises at least two hosts, wherein each host comprises at least one virtual machine, a coordinator and a memory module; the virtual machine, the coordinator and the memory module are electrically connected to each other; the coordinator of a local host is in remote communication connection with the coordinator of a remote host by means of an RDMA module; the virtual machine is used for receiving an access request initiated by an application program, and sending the access request to the coordinator; and the coordinator queries whether memory cache data requested by the access request is present in the memory module, and if so, the coordinator returns the memory cache data to the application program, and if not, the coordinator sends, by means of the RDMA module, the access request to the coordinator of the remote host in which memory data is stored. By means of the method in the present invention, a memory can be shared between virtual machines of the same host and between virtual machines across hosts.

Description

Memory sharing system and method based on virtual environment

technical field

The invention relates to the technical field of memory sharing, in particular to a memory sharing system and method based on a virtual environment.

Background technique

As DRAM prices continue to drop, it is cost-effective to build commodity servers with hundreds of gigabytes of DRAM capacity. However, DRAM now faces the problem of low integration. Although large clusters can reach TB-level main memory, the maximum memory of a single node is only a few hundred GB, which is difficult to meet the storage requirements of large-scale applications. Driven by this background, the shared memory model of accessing remote data through memory semantics makes a unified global memory abstraction very attractive for distributed computing. However, when it comes to the communication between remote hosts, the traditional TCP/IP network will cause a very large overhead due to the processing of the kernel protocol stack, and it cannot well meet the low-latency requirements of this shared memory model. need.

The remote direct memory access (RDMA module) technology offloads the processing of the protocol stack to the network card, and the network card directly uses the DMA technology to access the memory data after parsing the data packet, avoiding multiple copies of the data. At the same time, through the offloading of data processing, the CPU is freed from the network data transmission, which improves the efficiency of CPU usage. Through this zero-copy, kernel-bypassing feature, low latency, high throughput, and high performance during data transfer are achieved.

There has been a lot of in-depth research on memory sharing between virtual machines on the same host. However, judging from the current development trend of cloud computing, it is an irresistible trend for applications to run in a virtual environment, so memory sharing on physical hosts is no longer attractive. We need to consider memory sharing in the entire cluster environment based on multiple physical hosts in virtual environments such as virtual machines and containers.

Researchers from the National University of Singapore and the University of California have proposed a distributed memory management system GAM using RDMA modules and data caching. GAM implements a memory sharing model on a physically distributed cluster and can manage free memory distributed among multiple nodes to provide a unified memory model. At the same time, GAM provides a set of user-friendly APIs for memory operations. In order to improve the performance of data reading, GAM implements the cache of data on each host, and deeply thinks about cache coherence, distinguishes the different states of memory pages when they are shared, and designs data when reading and writing data. And consistency handling when data fails. At the same time, GAM also adopts the method of mixing communication primitives in the use of the RDMA module, and only uses the unilateral primitive WRITE in data transmission.

Most of the existing memory sharing systems at home and abroad are dedicated to memory sharing between different virtual machines on the same host, or between different physical hosts. There is not much research on memory sharing across virtual clusters.

technical problem

The present invention provides a memory sharing system and method based on a virtual environment, so as to solve the problem that the existing memory sharing method cannot efficiently realize memory sharing on a virtual cluster.

technical solutions

In order to solve the above technical problems, the present invention provides a memory sharing system based on a virtual environment, which includes: at least two hosts, each host includes at least one virtual machine, a coordinator and a memory module, a virtual machine, a coordinator, The memory modules are electrically connected to each other, and the coordinator of the local host and the coordinator of the remote host are connected by remote communication through the RDMA module; the virtual machine is used to receive the access request initiated by the application and send the access request to the coordinator; the coordinator queries the memory Whether the memory data requested by the access request exists in the module; if so, return the memory data to the application; if not, the coordinator sends an access request to the coordinator of the remote host that stores the memory data through the RDMA module.

As a further improvement of this application, the memory module includes shared memory, cache, and address space allocated by the operating system and application programs. The cache is used to cache the memory data of the remote host. After the coordinator receives the access request, it queries whether there is memory data in the cache. .

As a further improvement of the present application, an API interface, a library function and a virtual PCI device are provided in the virtual machine, and the application program calls the library function through the API interface to operate the virtual PCI device to access the memory module.

As a further improvement of the present application, the coordinator includes a communication module, a monitoring module and a message processing module; the communication module is used to implement communication between different virtual machines on the same host or between different hosts; the monitoring module is used to monitor the activities of the virtual machines, And by analyzing the access log of the memory module to distinguish the access type of each memory page, and configure the corresponding cache policy for the memory pages of different access types according to the preset policy rules; the message processing module is used for the message received by the communication module. Parse processing and distribute pre-configured message handlers for different types of messages.

As a further improvement of the present application, the communication modules between different hosts implement the data transmission process through the SEND/RECV bilateral primitive, the WRITE unilateral primitive or the WRITE_WITH_IMM primitive of the RDMA module.

As a further improvement of this application, when an access request or a memory page failure control message is transmitted between communication modules, data transmission is realized through the SEND/RECV bilateral primitive; when a large amount of data is transmitted between communication modules, the WRITE single The edge primitive realizes data transmission; when the response message is transmitted between communication modules, the data transmission is realized through the WRITE_WITH_IMM primitive.

As a further improvement of the present application, the monitoring module analyzes the access log of the memory module to distinguish the access type of each memory page, and configures corresponding cache policies for memory pages of different access types according to preset policy rules, including; When the number of accesses by the target remote host reaches the first preset condition, the access type of the memory page is divided into exclusive access, and the memory page is migrated to the memory module of the target remote host; when the record of read data on the memory page reaches the second In the preset condition, the access type of the memory page is divided into read sharing, and the memory page is stored on the remote host with the highest frequency of accessing the memory page, and the permission is set to read-only; when the record of written data on the memory page reaches In the third preset condition, the access type of the memory page is divided into write sharing, and synchronization of the memory page to other remote hosts is prohibited.

As a further improvement of the present application, the monitoring module generates an access log when monitoring that the remote host accesses the memory module of the local host, and the access log records the access host, the access address and the access type.

As a further improvement of the present application, the communication between the virtual machine and the coordinator is implemented by way of process communication.

In order to solve the above technical problems, the present invention also provides a virtual environment-based memory sharing method, which is applied to one of the above-mentioned virtual environment-based memory sharing systems. The method includes: when the virtual machine receives an access request initiated by an application program. , send the access request to the coordinator; the coordinator analyzes the access request and determines whether the access address of the access request is the local host or the remote host; if it is the local host, it obtains the data requested by the access request from the local host and feeds it back to the application ; If it is a remote host, judge whether the data requested by the access request is cached in the memory module; when the data requested by the access request is cached in the memory module, the data requested by the cached access request is fed back to the application; When the data requested by the access request does not exist in the module, the coordinator sends an access request to the coordinator of the remote host through the RDMA module, and feeds back the data obtained by the access to the application program.

beneficial effect

The beneficial effects of the present invention are: the memory sharing system based on the virtual environment of the present invention sets up a coordinator in the host, and realizes data communication between different virtual machines in the host through the coordinator to realize memory sharing, and aims at different hosts. Between them, it uses the RDMA module to realize the communication between the coordinators of different hosts, so as to realize the communication between different hosts, and finally realize the memory sharing of the entire virtual cluster, and the use of the RDMA module effectively reduces the data interaction between different hosts. Delay.

Description of drawings

1 is a schematic structural diagram of a memory sharing system based on a virtual environment according to an embodiment of the present invention;

2 is a schematic structural diagram of a memory module of a memory sharing system based on a virtual environment according to an embodiment of the present invention;

3 is a schematic structural diagram of a virtual machine of a memory sharing system based on a virtual environment according to an embodiment of the present invention;

4 is a schematic structural diagram of a coordinator of a virtual environment-based memory sharing system according to an embodiment of the present invention;

5 is a schematic diagram of an operation flow that needs to be performed when a cached memory page of a memory sharing system based on a virtual environment is modified according to an embodiment of the present invention;

FIG. 6 is a schematic flowchart of a memory sharing method based on a virtual environment according to an embodiment of the present invention.

BEST MODE FOR CARRYING OUT THE INVENTION

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

The terms "first", "second" and "third" in the present invention are only used for description purposes, and should not be understood as indicating or implying relative importance or implying the number of indicated technical features. Thus, a feature defined as "first", "second", "third" may expressly or implicitly include at least one of that feature. In the description of the present invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise expressly and specifically defined. All directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative positional relationship between various components under a certain posture (as shown in the accompanying drawings). , motion situation, etc., if the specific posture changes, the directional indication also changes accordingly. Furthermore, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units is not limited to the listed steps or units, but optionally also includes unlisted steps or units, or optionally also includes For other steps or units inherent to these processes, methods, products or devices.

Reference herein to an "embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present invention. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor a separate or alternative embodiment that is mutually exclusive of other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

FIG. 1 is a schematic structural diagram of a memory sharing system based on a virtual environment according to an embodiment of the present invention. As shown in FIG. 1, the virtual environment-based memory sharing system 100 includes: at least two hosts 1, each host 1 includes at least one virtual machine 2, a coordinator 3 and a memory module 4, the virtual machine 2, the coordinator 3 , the memory modules 4 are electrically connected to each other, and the coordinator 3 of the local host and the coordinator 3 of the remote host are connected by remote communication through the RDMA module 5; the virtual machine 2 is used to receive the access request initiated by the application program, and send the access request to the coordinator. Coordinator 3; Coordinator 3 inquires whether there is memory data requested by the access request in memory module 4; if so, returns the memory data to the application; The coordinator 3 of the remote host issues an access request.

It should be noted that the virtual environment-based memory sharing system 100 is developed based on KVM/QEMU, wherein the virtual machine 2 runs on KVM. In this embodiment, the last simulated virtual machine of QEMU is used to simulate the shared memory on the cluster. Multiple virtual machines 2 can be built on the same host 1. It should be noted that when an application initiates an access request to the virtual machine of host A, host A is the local host, and when the memory data accessed by the application is in the memory module of host B, then host B is the remote host.

Specifically, when the virtual machine 2 receives the access request initiated by the application, it sends the access request to the coordinator 3; the coordinator 3 analyzes the access request, and judges whether the access address of the access request is the local host or the remote host; if it is the local host, Then obtain the memory data requested by the access request from the local host and feed it back to the application; if it is a remote host, then determine whether the memory data requested by the access request is cached in the memory module 4; When the memory data requested, the memory data requested by the cached access request is fed back to the application program; when the memory data requested by the access request does not exist in the memory module 4, the coordinator 3 sends the data to the coordinator of the remote host through the RDMA module 5 3 Send an access request, and feed back the memory data obtained by the access to the application.

The memory sharing system based on the virtual environment of this embodiment sets the coordinator 3 in the host, and realizes the data communication between different virtual machines 2 in the host 1 through the coordinator 3 to realize memory sharing. During the time, it realizes the communication between the coordinators 3 of different hosts 1 by using the RDMA module 5, so as to realize the communication between different hosts 1, and finally realize the memory sharing of the whole virtual cluster, and the RDMA module 5 effectively reduces the different Latency of data exchange between hosts 1.

Further, as shown in FIG. 2, the memory module 4 includes a shared memory 41, a cache 42 and an address space 43 allocated by the operating system and application programs. The cache 42 is used to cache the memory data of the remote host. After the coordinator 3 receives the access request, , query whether the memory data exists in the cache 42 .

Specifically, the size of the shared memory 41 of each host 1 is indicated by the command line during system initialization, and when the shared memory 41 is actually allocated, a fixed-size cache 42 needs to be allocated, and the cache 42 is used to cache remote Host memory data. It should be noted that the memory data in the cache is invisible to the virtual machine 2. Therefore, when the virtual machine sends an access request to the coordinator 3, the coordinator 3 queries the cache 42 whether there is the memory data to be queried. , if it exists, directly return the memory data to the application, if not, send the access request to the remote host.

Further, as shown in FIG. 3 , the virtual machine 2 is provided with an API interface 21 , a library function 22 and a virtual PCI device 23 , and the application calls the library function 22 through the API interface 21 to operate the virtual PCI device 23 to access the memory module 4 .

Specifically, the virtual PCI device 23 is established based on QEMU, and the number of registers and message processing functions are expanded. Among them, the library function 22 includes a library function provided by QEMU and a preconfigured library function for the response message.

Further, as shown in FIG. 4 , the coordinator 3 includes a communication module 31, a monitoring module 32 and a message processing module 33; the communication module 31 is used to implement communication between different virtual machines 2 on the same host 1 or between different hosts; monitoring The module 32 is used to monitor the activity of the virtual machine 2, and to distinguish the access type of each memory page by analyzing the access log of the memory module 4, and configure the corresponding cache policy for the memory pages of different access types according to preset policy rules; message The processing module 33 is configured to parse and process the messages received by the communication module, and distribute pre-configured message processing functions for different types of messages.

Specifically, the coordinator 2 is a loadable module responsible for tasks such as maintaining the virtual machine information of the host 1, communicating with the virtual machine 2, maintaining the shared memory on the host 1, and communicating with the remote host.

In this embodiment, the communication module 31 is responsible for completing the communication between the virtual machines 2. Since the same host and different hosts have great differences in communication, in this embodiment, the communication is divided into two levels, one is the same host 1 The communication between different virtual machines 2 and the virtual machine 2 and the coordinator 3, and the second is the communication between different hosts 1. The communication between the virtual machine 2 and the coordinator 3 is implemented in a process manner, and does not involve data transmission on the network, so the transmission efficiency is higher.

Further, the communication module 31 between different hosts 1 implements the data transmission process through the SEND/RECV bilateral primitive, the WRITE unilateral primitive or the WRITE_WITH_IMM primitive of the RDMA module 5 . Specifically, when the communication module 31 transmits an access request or a control message of memory page failure, data transmission is realized through the SEND/RECV bilateral primitive; when a large amount of data is transmitted between the communication modules 31, the WRITE unilateral primitive is used. The data transmission is realized by the WRITE_WITH_IMM primitive; when the response message is transmitted between the communication modules 31, the data transmission is realized through the WRITE_WITH_IMM primitive. The memory page invalidation refers to the memory page of the remote host stored in the cache in the local host, and the data transmission change of the memory page of the remote host is that the memory page stored in the cache is invalidated.

Specifically, the connections between different hosts 1 are based on RDMA. In the process of data transmission, the idea of software-defined networking is used to separate the control path and the data path, and different data transmission sources are used on these two paths. In addition, in order to avoid the interference of the two, the transmission of the control path and the data path is divided into two connections, specifically:

1. For access request messages and control messages such as some memory page failures, the data they carry is usually small. At the same time, in order to ensure the effective transmission of control commands, ensure that the other end can receive them, and has certain reliability. , the SEND/RECV bilateral primitive in RDMA module 5 can be used. In this way, the participation of the CPU is required. The receiver needs to use RECV to prepare for receiving data, and then the sender uses SEND to send data to ensure the reception. Fang received the message.

2. For general simple data transmission, the amount of data is large, and the requirements for delay are relatively high. For this, we use the WRITE unilateral primitive of RDMA module 5. In this way, as long as the two ends establish a connection, the sender can directly write the data to the receiver's memory without the participation of the receiver's CPU. , which further reduces system latency in the form of direct access to remote memory.

3. Part of the data in the data transmission is that the host 1 needs to return a response packet after receiving the control message to determine the received message. For this type of message, it is usually only necessary to carry the data in the message header. Therefore, the WRITE_WITH_IMM primitive is used To transmit such data, the WRITE_WITH_IMM primitive does not carry payload during data transmission, reducing the resources consumed by data transmission.

The monitoring module 32 is responsible for monitoring the opening, closing and other activities of the virtual machine 2, and is responsible for analyzing the access log of the memory module 4, thereby distinguishing the access type of each memory page in the memory module 4, and according to the preset policy rules for different access types The corresponding cache policy is configured for the memory page.

It should be noted that the monitoring module 32 generates an access log when monitoring that the remote host accesses the memory module 4 of the local host, and the access log records the contents of the access host, the access address, and the access type.

The message processing module 33 is responsible for analyzing and processing the messages received by the communication module 31. For different types of messages, the message types are confirmed according to the message structure, and then the messages are distributed to different message processing functions for processing. Among them, the type of the message can be divided according to the access object, such as the message of accessing the local host or the message of accessing the remote host, and can also broadcast the message of the newly created virtual machine 2 to the virtual cluster when the newly created virtual machine 2 is initialized, or It is a message or the like of the size of the shared memory of each host 1 . The message processing function is preset.

Further, in order to speed up the access speed to the remote data and reduce the delay of the system, in this embodiment, the monitoring module 31 analyzes the access log of the memory module 4 to distinguish the access type of each memory page, and according to the preset policy rule is: The cache policies corresponding to the memory page configurations of different access types are as follows:

1. When the number of times the memory page is accessed by the target remote host reaches the first preset condition, the access type of the memory page is divided into exclusive access, and the memory page is migrated to the memory module 4 of the target remote host.

It should be noted that the first preset condition is preset.

Specifically, when the number of accesses of the memory page of the local host by a remote host accounts for more than 90% of the total number of accesses, the access type at this time is divided into exclusive access. At this time, considering that the delay of data transmission is too high, In order to reduce the delay, the memory page is migrated to the remote host, and the coordinator 3 of the remote host confirms the actual storage address in the memory module 4. If other remote hosts occasionally access the memory page, a notification is sent to the remote host. to confirm that the remote host can access the memory page.

2. When the record of read data on the memory page reaches the second preset condition, the access type of the memory page is divided into read sharing, and the memory page is stored on the remote host with the highest frequency of accessing the memory page, and the permissions are set. is read-only.

It should be noted that the second preset condition is preset.

Specifically, when it is detected that more than 80% of the records of the memory page are read data, the access type of the memory page is divided into read sharing. At this time, the memory page needs to be cached to the remote host with the highest access frequency to the memory page. On the remote host, the coordinator 3 places the data in the reserved cache space for use, and sets the permission to read-only. For example, Figure 5 shows the operation process that needs to be done when the cached memory page is modified. Among them, node0 issues a write request, node1 is the node that retains the original memory page, node2 is the node that has cached data, and when node0 sends a write request When writing a request, node1 will reply to the request to node0 and process data access. At the same time, node1 will send a notification of memory page invalidation to node2 and other nodes that have memory page cache, and node2, etc. will modify the flag of the cached page and reply the response packet. When subsequent applications need to access again, the data needs to be re-cached to the local host.

3. When the record of written data on the memory page reaches the third preset condition, the access type of the memory page is divided into write sharing, and synchronization of the memory page to other remote hosts is prohibited.

It should be noted that the third preset condition is preset.

Specifically, considering the particularity of write data, when more than 50% of the records in the memory page are written data, the access type of the memory page is divided into write sharing, because the data of the memory page keeps changing when the write operation occurs. , if the memory pages are cached on multiple remote hosts, considering the maintenance of data consistency, information needs to be exchanged frequently between multiple remote hosts, requiring a lot of synchronization operations, and even involving a lot of lock operations, which greatly reduces the system performance. performance, resulting in a significant increase in the delay of data transmission. Therefore, the shared memory page is not considered, and the application program still sends a request, and the coordinator 3 takes over the operation and accesses data to the remote host.

6 shows the virtual environment-based memory sharing method of the present invention, which is applied to one of the above-mentioned virtual environment-based memory sharing systems, and the method includes:

Step S1: when the virtual machine receives the access request initiated by the application, it sends the access request to the coordinator;

Step S2: The coordinator analyzes the access request, and determines whether the access address of the access request is a local host or a remote host; if it is a local host, go to step S3; if it is a remote host, go to step S4.

Step S3: obtain the data requested by the access request from the local host and feed it back to the application;

Step S4: judging whether the data requested by the access request is cached in the memory module; when the data requested by the access request is cached in the memory module, execute step S5; when the data requested by the access request does not exist in the memory module, execute Step S6.

Step S5: feedback the data requested by the cached access request to the application;

Step S6: The coordinator sends an access request to the coordinator of the remote host through the RDMA module, and feeds back the data obtained by the access to the application program.

In this embodiment, a coordinator is set in the host, and the coordinator realizes the data communication of different virtual machines in the host to realize memory sharing, and for different hosts, the RDMA module is used to realize the communication between the coordinators of different hosts. Therefore, the communication between different hosts is realized, and the memory sharing of the entire virtual cluster is finally realized, and the RDMA module is used to effectively reduce the delay of data interaction between different hosts.

The specific embodiments of the invention have been described in detail above, but they are only used as examples, and the present invention is not limited to the specific embodiments described above. For those skilled in the art, any equivalent modification or substitution of the invention is also within the scope of the present invention. Therefore, equivalent changes, modifications and improvements made without departing from the spirit and principle scope of the present invention etc., should be included in the scope of the present invention.

Claims

A memory sharing system based on a virtual environment, characterized in that it includes:

At least two hosts, each of which includes at least one virtual machine, a coordinator, and a memory module, the virtual machine, the coordinator, and the memory module are electrically connected to each other, and the coordinator of the local host is connected to the remote host. The coordinator is connected by remote communication through the RDMA module;

The virtual machine is configured to receive an access request initiated by an application, and send the access request to the coordinator;

The coordinator inquires whether the memory data of the access request exists in the memory module; if so, returns the memory data to the application; if not, the coordinator passes the RDMA module The access request is issued to the coordinator of the remote host where the memory data is stored.
The memory sharing system based on a virtual environment according to claim 1, wherein the memory module comprises a shared memory, a cache and an address space allocated by an operating system and an application program, and the cache is used to cache memory data of a remote host , after receiving the access request, the coordinator queries whether the memory data exists in the cache.
The memory sharing system based on a virtual environment according to claim 1, wherein the virtual machine is provided with an API interface, a library function and a virtual PCI device, and the application calls the library function through the API interface The memory module is accessed by operating the virtual PCI device.
The memory sharing system based on a virtual environment according to claim 1, wherein the coordinator comprises a communication module, a monitoring module and a message processing module; the communication module is used to realize the communication between different virtual machines on the same host. Communication between different hosts; the monitoring module is used to monitor the activity of the virtual machine, and to distinguish the access type of each memory page by analyzing the access log of the memory module, and according to the preset policy rules as different The memory page of the access type is configured with a corresponding cache policy; the message processing module is configured to parse and process the message received by the communication module, and distribute pre-configured message processing functions for different types of messages.
The virtual environment-based memory sharing system according to claim 4, wherein the communication modules between different hosts use the SEND/RECV bilateral primitive, WRITE unilateral primitive or WRITE_WITH_IMM primitive of the RDMA module Implement the data transfer process.
The memory sharing system based on a virtual environment according to claim 5, wherein when the access request or the control message of memory page failure is transmitted between the communication modules, the SEND/RECV bilateral primitive is used. Realize data transmission; when a large amount of data is transmitted between the communication modules, data transmission is realized through the WRITE unilateral primitive; when a response message is transmitted between the communication modules, the data is realized through the WRITE_WITH_IMM primitive transmission.
The memory sharing system based on a virtual environment according to claim 4, wherein the monitoring module analyzes the access log of the memory module to distinguish the access type of each memory page, and assigns different access types according to preset policy rules. The cache policy corresponding to the type of memory page configuration, including;

When the number of times the memory page is accessed by the target remote host reaches a first preset condition, dividing the access type of the memory page into exclusive access, and migrating the memory page to the memory module of the target remote host;

When the record of read data on the memory page reaches the second preset condition, the access type of the memory page is divided into read sharing, and the memory page is stored to the remote host with the highest frequency of accessing the memory page , and set the permission to read-only;

When the record of the written data on the memory page reaches the third preset condition, the access type of the memory page is divided into write sharing, and synchronization of the memory page to other remote hosts is prohibited.
The memory sharing system based on a virtual environment according to claim 4, wherein the monitoring module generates the access log when monitoring that the remote host accesses the memory module of the local host, and the access log The access host, access address and access type are recorded.
The memory sharing system based on a virtual environment according to claim 1, wherein the communication between the virtual machine and the coordinator is implemented in a process mode.
A memory sharing method based on a virtual environment, characterized in that it is applied to the memory sharing system based on a virtual environment according to one of claims 1-9, the method comprising:

When the virtual machine receives the access request initiated by the application, it sends the access request to the coordinator;

The coordinator analyzes the access request, and determines whether the access address of the access request is a local host or a remote host;

If it is a local host, obtain the data requested by the access request from the local host and feed it back to the application;

If it is a remote host, then determine whether the data requested by the access request is cached in the memory module;

When the data requested by the access request is cached in the memory module, the cached data requested by the access request is fed back to the application;

When the data requested by the access request does not exist in the memory module, the coordinator sends the access request to the coordinator of the remote host through the RDMA module, and feeds back the data obtained by the access to the application program.