CN106598696B - Method and device for data interaction between virtual machines - Google Patents

Abstract

The invention relates to the field of computers, in particular to a method and a device for data interaction between virtual machines, which aim to solve the problem of processing bottleneck possibly occurring in data transmission between the virtual machines, and the method comprises the following steps: when the first virtual machine determines that the application in the first virtual machine needs to send data to the application in the second virtual machine, determining the initial address and the data length of the data to be sent; the first virtual machine determines the identification of the target memory chip according to the identification of the second virtual machine; and the first virtual machine determines the initial copying position of the data to be sent according to the write offset information in the target memory chip, copies the data to be sent to the target memory chip according to the initial address and the data length of the data to be sent, and updates the write offset information. Therefore, the method does not depend on the existing network protocol, and the service processing between the virtual machines is accelerated.

Description

A method and device for data interaction between virtual machines

technical field

The present invention relates to the field of computers, and in particular, to a method and device for data interaction between virtual machines.

Background technique

Existing virtual machine technologies are widely used in cluster services, and there will be a large amount of data interaction between virtual machines (virtual machines, VMs) in cluster services. The current common practice is based on the virtual machine as an independent unit, and data transmission between virtual machines is carried out through a standard transmission channel, which has the advantages of standardization, ease of use, and good compatibility. However, in a scenario where a single traffic interaction between virtual machines is large and multiple interactive processing is required between virtual machines, if data transmission is performed through a standard transmission channel, the application may be limited or unable to reach the processing bottleneck. application.

The following two methods for accelerating the processing of inter-VM services are proposed in the prior art:

Method 1: The high-performance network I/O framework (the fast packet I/O framework, netmap) maps the network cache (buffer) to the user space corresponding to each virtual machine through memory mapping, so that the virtual machine can directly operate the buffer . Specifically, when there is no memory mapping, the data of the upper-layer application needs to be copied several times from the user state to the kernel state, and then to the cache of the network card driver. After memory mapping, the data copy level is reduced. The number of data copies is reduced, the service processing between virtual machines is accelerated, and the performance of the TCP/IP protocol stack is improved. Netmap is a general practice for accelerated processing based on network card driver forwarding.

Method 2: Accelerate network transmission through Single-root I/O virtualization (SR-IOV), wherein SR-IOV is based on the Peripheral Component Interconnect Special Interest Group (PCI-IOV) SIG) for implementing standard terminology to create and manage a device for local shares. Like netmap, it is also based on network card acceleration to realize network packet transmission.

SUMMARY OF THE INVENTION

The purpose of the embodiments of the present invention is to provide a method and apparatus for data interaction between virtual machines, so as to solve the problem that the application is limited or cannot be applied due to reaching a processing bottleneck when data transmission between virtual machines may occur. .

The purpose of the embodiment of the present invention is achieved through the following technical solutions:

In a first aspect, a method for data interaction between virtual machines includes: when a first virtual machine determines that an application in the first virtual machine needs to send data to an application in a second virtual machine, determining a start address of the data to be sent and data length, the first virtual machine and the second virtual machine belong to the same host; the embodiment of the present invention is mainly applied to the scenario of communication between virtual machines in the same host. Further, the first virtual machine determines the identifier of the target memory slice according to the identifier of the second virtual machine, and the target memory slice refers to the data allocated by the host machine for the first virtual machine to store the data sent by the first virtual machine to the second virtual machine. A memory slice, wherein the host allocates a unique memory slice corresponding to each other virtual machine in the host for each virtual machine. The first virtual machine determines the initial copy position of the data to be sent according to the write offset information in the target memory slice, and copies the data to be sent to the target memory slice according to the start address and data length of the data to be sent, and updates the write offset information ; wherein, the write offset information is used to indicate the write offset of the data stored in the target memory slice, that is, the end position after the last copy of the data. Therefore, compared with the prior art, which implements network packet transmission based on network cards, the method provided by the embodiment of the present invention only needs to copy data once, and does not depend on the existing network protocol, and realizes virtualized virtualisation through the shared memory in the host machine. The interaction between virtual machines can be accelerated to speed up the business processing between virtual machines, so as to meet the requirements of scenarios that require high interaction performance between virtual machines.

In a possible implementation manner, the target memory slice includes a management control area and a service load area; the management control area is used to store the identifier of the target memory slice, the size of the remaining space in the service load area, and the the write offset information; the service load area is used to store the data sent by the first virtual machine to the second virtual machine; the first virtual machine is based on the starting address and data length of the data to be sent Copying the to-be-sent data to the memory slice includes: when the first virtual machine determines that the size of the remaining memory space in the service load area is less than or equal to the data length of the data, determining that the size of the remaining memory space before exceeding the preset duration is determined. When the size of the remaining memory space in the service load area is greater than the data length of the data, the data to be sent is copied to the service load area according to the start address and data length of the data to be sent. At this time, the first virtual machine waits for the size of the remaining memory space to be greater than the data length of the data to be copied, so as to avoid the situation of packet loss.

In a possible implementation manner, the method further includes: when the first virtual machine determines that the size of the remaining memory space in the service load area is less than or equal to the data length of the data, the first virtual machine sends the data to be sent. Failed to copy data to the service load area. At this time, the first virtual machine does not wait for the remaining space to change, and when the size of the remaining memory space is less than or equal to the data length of the data, the data is directly discarded to avoid subsequent data accumulation and speed up data processing.

In addition, in the circular cache management system, when the write offset reaches the end of the service load area, that is, the remaining space in the service load area is 0, the content in the service load area can be overwritten according to the read offset to realize the circular cache. use.

A second aspect provides a method for data interaction between virtual machines, comprising: a first virtual machine detects a set of target memory slices corresponding to the first virtual machine, and each target memory slice is used for storing data sent to the first virtual machine. When the first virtual machine detects that the write offset information in the first target memory slice is updated, the first virtual machine reads all the data according to the write offset information and the read offset information in the first target memory slice. data in the first target memory slice, and update the read offset information; wherein, the write offset information is used to indicate the write offset of the data stored in the first target memory slice, and the read offset information The shift information is used to indicate the read offset of the read data in the first target memory slice. Therefore, the virtual machine can detect which virtual machines send data to itself, and read the data in the memory slice according to the read offset information and write offset information of the corresponding memory slice in time, so as to speed up the business processing between the virtual machines, so as to meet the Scenarios that require high interaction performance between virtual machines.

In addition, the first virtual machine learns the read offset of the read data according to the read offset information in the first target memory slice, starts to read the data from the read position, and can read the data up to the amount indicated by the write offset information. The write offset of the stored data, or the first virtual machine configures a preset data read length for each read data, when the difference between the write offset and the read offset is greater than or equal to the preset data read length, Read the data of the preset data read length. When the difference between the write offset and the read offset is less than the preset data read length, the read can be performed next time. When the first virtual machine reads data in the first target memory slice, the read offset information needs to be updated.

In a possible implementation manner, the first target memory slice includes a management control area and a service load area; the management control area is used to store the identifier of the first target memory slice and the remainder of the service load area. space size, the write offset information and the read offset information; the service load area is used to store the data sent by the second virtual machine to the first virtual machine.

A third aspect, a method for data interaction between virtual machines, comprising: a host machine allocating memory space as the shared memory of the virtual machine to be created; the host machine creating N virtual machines, N≥2, and N is a positive integer; The host machine divides the shared memory into N*(N-1) memory slices, and allocates (N-1) memory slices for each virtual machine; -1) The memory slices are in one-to-one correspondence with other (N-1) virtual machines except the i-th virtual machine, i≥1, and i is a positive integer. Therefore, the host can allocate (N-1) memory slices to each virtual machine according to the number N of virtual machines created, so as to speed up the business processing between the virtual machines, so as to meet the higher requirements on the interaction performance between the virtual machines. scene requirements.

Specifically, the eshare module can be a collection of multiple logical processing units, providing functions such as recording the memory address mapping corresponding to (N-1) memory slices, standard read and write interfaces, read and write mutual exclusion management, and read and write interface encapsulation.

In a possible implementation manner, the host machine allocates (N-1) memory slices to each virtual machine, including: for each memory slice allocated to the ith virtual machine, the host machine executes: Converting the host machine physical address corresponding to the current memory slice to the host machine virtual address corresponding to the current memory slice; converting the host machine virtual address corresponding to the current memory slice to the client machine physical address corresponding to the current memory slice; Convert the client physical address corresponding to the current memory slice to the client virtual address corresponding to the current memory slice. Therefore, the shared memory in the host can be mapped to the memory space of the virtual machine.

A fourth aspect, an apparatus for data interaction between virtual machines, comprising: a determination unit configured to determine a start address of data to be sent when an application in the apparatus needs to send data to an application in a second virtual machine and data length, the device and the second virtual machine belong to the same host; the analysis unit is used to determine the identifier of the target memory slice according to the identifier of the second virtual machine, and the target memory slice refers to the a memory slice allocated by the host to the device for storing data sent by the device to the second virtual machine; a processing unit, configured to determine the to-be-sent data according to write offset information in the target memory slice The initial copy position of the data, and the data to be sent is copied to the target memory slice according to the starting address and data length of the data to be sent, and the write offset information is updated; wherein, the write offset information It is used to indicate the write offset of the stored data in the target memory slice.

In a possible implementation manner, the target memory slice includes a management control area and a service load area; the management control area is used to store the identifier of the target memory slice, the size of the remaining space in the service load area, and the the write offset information; the service load area is used to store the data sent by the device to the second virtual machine; according to the start address and data length of the data to be sent, the data to be sent is copied to the When the memory chip is used, the processing unit is specifically configured to: when it is determined that the size of the remaining memory space in the service load area is less than or equal to the data length of the data, determine the remaining memory in the service load area before the preset time length is exceeded When the space size is greater than the data length of the data, the data to be sent is copied to the service load area according to the start address and data length of the data to be sent.

In a possible implementation manner, the processing unit is further configured to: when it is determined that the size of the remaining memory space in the service load area is less than or equal to the data length of the data, copy the data to be sent to the service Load zone failed.

A fifth aspect, an apparatus for data interaction between virtual machines, comprising: a detection unit for detecting a set of target memory slices corresponding to the apparatus, each target memory slice being used to store data sent to the apparatus; processing The unit is configured to read the first target memory slice according to the write offset information and the read offset information in the first target memory slice when it is detected that the write offset information in the first target memory slice is updated and update the read offset information; wherein, the write offset information is used to indicate the write offset of the data stored in the first target memory slice, and the read offset information is used to indicate the The read offset of the read data in the first target memory slice is described.

In a possible implementation manner, the first target memory slice includes a management control area and a service load area; the management control area is used to store the identifier of the first target memory slice and the remainder of the service load area. space size, the write offset information and the read offset information; the service load area is used to store the data sent by the second virtual machine to the device.

A sixth aspect, an apparatus for data interaction between virtual machines, comprising: an allocation unit for allocating memory space as the shared memory of the virtual machine to be created; a creation unit for creating N virtual machines, N≥2, N is a positive integer; the processing unit is used to divide the shared memory into N*(N-1) memory slices, and allocate (N-1) memory slices to each virtual machine; wherein, is the ith virtual The (N-1) memory slices allocated by the machine are in one-to-one correspondence with other (N-1) virtual machines except the i-th virtual machine, i≥1, and i is a positive integer.

In a possible implementation manner, when (N-1) memory slices are allocated to each virtual machine, the processing unit is specifically configured to: for each memory slice allocated to the ith virtual machine, execute: Converting the host machine physical address corresponding to the current memory slice to the host machine virtual address corresponding to the current memory slice; converting the host machine virtual address corresponding to the current memory slice to the client machine physical address corresponding to the current memory slice; Convert the client physical address corresponding to the current memory slice to the client virtual address corresponding to the current memory slice.

In a seventh aspect, a device for data interaction between virtual machines, comprising: a communication interface 1001 and a processor 1002; the processor is configured to: determine that an application in the apparatus needs to send data to an application in a second virtual machine When data, determine the starting address and data length of the data to be sent, the device and the second virtual machine belong to the same host; determine the identifier of the target memory slice according to the identifier of the second virtual machine, the target A memory slice refers to a memory slice allocated by the host to the device for storing data sent by the device to the second virtual machine; the memory slice to be determined is determined according to the write offset information in the target memory slice. Send the initial copy position of the data, and copy the data to be sent to the target memory slice through the communication interface according to the starting address and data length of the data to be sent, and update the write offset information; wherein, The write offset information is used to indicate the write offset of the data stored in the target memory slice.

In an eighth aspect, a device for data interaction between virtual machines, comprising: a communication interface and a processor; the processor is configured to: detect a set of target memory slices corresponding to the device, and each target memory slice is used to store data sent to the device; when the update of the write offset information in the first target memory slice is detected, pass the communication interface according to the write offset information and the read offset information in the first target memory slice Read the data in the first target memory slice, and update the read offset information; wherein the write offset information is used to indicate the write offset of the data stored in the first target memory slice, so The read offset information is used to indicate the read offset of the read data in the first target memory chip.

In a ninth aspect, a device for data interaction between virtual machines, such as a host machine, includes: a communication interface, a processor, and a memory; the processor is configured to: allocate memory space in the memory as a virtual machine to be created The shared memory is used to create N virtual machines, N≥2, N is a positive integer; the shared memory is divided into N*(N-1) memory slices, and (N-1) is allocated to each virtual machine ) memory slices; wherein, the (N-1) memory slices allocated for the i-th virtual machine correspond one-to-one with other (N-1) virtual machines except the i-th virtual machine, and i≥1 , i is a positive integer; the communication interface is used to communicate with each virtual machine in the host.

Description of drawings

1 is a schematic diagram of a framework based on KVM+qemu+guest os in an embodiment of the present invention;

2 is an overview flowchart of a method for data interaction between virtual machines on the host side in an embodiment of the present invention;

3 is a schematic structural diagram of a memory chip in an embodiment of the present invention;

4 is one of the overview flowcharts of the method for data interaction between virtual machines in an embodiment of the present invention;

5 is the second overview flow chart of the method for data interaction between virtual machines in an embodiment of the present invention;

6 is a schematic diagram of a specific interaction process between virtual machines in an embodiment of the present invention;

FIG. 7 is a schematic structural diagram of an apparatus for data interaction between virtual machines in an embodiment of the present invention;

8 is a second schematic structural diagram of an apparatus for data interaction between virtual machines in an embodiment of the present invention;

9 is a third schematic structural diagram of an apparatus for data interaction between virtual machines in an embodiment of the present invention;

10 is one of the schematic diagrams of device structures for data interaction between virtual machines in an embodiment of the present invention;

11 is a second schematic structural diagram of an apparatus for data interaction between virtual machines in an embodiment of the present invention;

FIG. 12 is a third schematic structural diagram of an apparatus for data interaction between virtual machines according to an embodiment of the present invention.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of the embodiments. Dependency For the virtual machine built in the host itself, if it can be processed without going through the network protocol stack, this part of the performance loss can be saved.

Linux kernel virtualization technology currently has two mainstreams: Open Source Virtual Machine Monitor (XEN) and Kernel-based Virtual Machine (KVM). The general framework based on KVM+qemu+guest os is shown in Figure 1. Show. Generally, the bottom layer of the system is hardware, such as various devices, such as cpu and mem (memory stick). The layer above it is an operating system (OS: operate system), in which KVM is a module of the OS kernel (kernel) that provides virtual machine management. Quick Emulator (qemu) is an open source emulator software for a complete personal computer (PC) system. Each isolated guest os in the upper layer (each guest os is actually a process) can be simulated by qemu. The implementation of the embodiments of the present invention is mainly aimed at the above-mentioned KVM+qemu+guestos (Linux system) virtual machine system.

The preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

Referring to FIG. 2 , an embodiment of the present invention provides a method for data interaction between virtual machines, and the method includes:

Step 200: The host allocates memory space as the shared memory of the virtual machine to be created.

Step 210: The host creates N virtual machines, where N≥2, and N is a positive integer.

Step 220: The host divides the shared memory into N*(N-1) memory slices, and allocates (N-1) memory slices to each virtual machine.

Among them, the (N-1) memory slices allocated for the ith virtual machine are in one-to-one correspondence with other (N-1) virtual machines except the ith virtual machine, i≥1, and i is a positive integer.

Specifically, the host (Host) allocates memory space as the shared memory of the virtual machine to be created. Assuming that Host creates N virtual machines through qemu, the shared memory is divided into N*(N-1) memory slices. Wherein, each virtual machine corresponds to (N-1) memory slices, which are used to respectively store data sent by the virtual machine to other (N-1) virtual machines.

The structure of each memory slice is described in Figure 3, including the management control area and the service load area. Among them, the first section is the management control area, which occupies less memory space and is used to store the identification of the memory slice, the remaining memory space size of the service load area, read offset information and write offset information. The latter part is the service load area, which is used to store the interactive data between nodes.

Optionally, when the Host creates a virtual machine through qemu, qemu allocates (N-1) memory slices in the shared memory for the created virtual machine, that is, the shared memory in the Host is mapped to the memory space of the virtual machine.

Specifically, when allocating each memory slice to the ith virtual machine, the Host executes:

Convert the host physical address corresponding to the current memory slice to the host virtual address corresponding to the current memory slice, convert the host virtual address corresponding to the current memory slice to the client physical address corresponding to the current memory slice, and convert the current memory slice corresponding to the guest physical address. The guest physical address is converted to the guest virtual address corresponding to the current memory slice.

In addition, the memory slice identifier of each memory slice needs to be determined. For example, if the memory slice identifier allocated to the i-th virtual machine is ij, the service load area of the memory slice is used to store the i-th virtual machine and send it to the j-th virtual machine. data of a virtual machine. The memory slice is identified as ji, and the service load area of the memory slice is used to store the data sent by the jth virtual machine to the ith virtual machine, which is represented as the memory slice designated by the jth virtual machine. The method for determining the memory slice identifier here is just an example, and no specific limitation is made.

After the host machine completes the allocation of memory slices, the following describes the specific process of exchanging data between virtual machines with reference to the accompanying drawings.

Referring to FIG. 4 , an embodiment of the present invention provides a method for data interaction between virtual machines. On the sending side, the method includes:

Step 400: When the first virtual machine determines that the application in the first virtual machine needs to send data to the application in the second virtual machine, it determines the start address and data length of the data to be sent, and the first virtual machine and the second virtual machine belong to the same host.

It should be understood that the method provided by the embodiment of the present invention optimizes the interaction process between virtual machines belonging to the same host, and the method in the prior art is still used for the interaction between virtual machines in different hosts.

Step 410: The first virtual machine determines the identifier of the target memory slice according to the identifier of the second virtual machine, and the target memory slice refers to the memory allocated by the host machine for the first virtual machine to store the data sent by the first virtual machine to the second virtual machine. memory chips.

From the above process of allocating memory slices for virtual machines by the host, when the total number of virtual machines in the host is N, each virtual machine corresponds to (N-1) memory slices, which are used to store the virtual machine and send it to other Data for (N-1) virtual machines. Therefore, when the application in the first virtual machine determines to send data to the application in the second virtual machine, the target memory slice can be determined from the currently allocated memory slice according to the identifier of the second virtual machine, that is, the target memory slice is used to store data sent to the second virtual machine. Two memory slices for virtual machine data. For example, an index table of the correspondence between the virtual machine identifier and the memory slice may be stored in the virtual machine, so as to determine the target memory slice corresponding to each virtual machine.

Step 420: The first virtual machine determines the initial copy position of the data to be sent according to the write offset information in the target memory slice, and copies the data to be sent to the target memory slice according to the start address and data length of the data to be sent, and updates the write offset information.

The write offset information is used to indicate the write offset of the data stored in the target memory slice.

It should be understood that, in the circular cache management system, the write offset information is used to indicate the write offset of the data stored in the memory slice, so as to determine the initial write position when data is written next time. Therefore, the first virtual machine can know the initial write position of the data to be written this time, that is, the initial copy position of the data to be sent, according to the write offset information, and then according to the start position and data length of the data to be sent, the data to be sent Copy to the target memory slice. It can be seen from this that, compared with the prior art, which implements network message transmission based on network cards, the method provided by the embodiment of the present invention only needs to copy data once, does not depend on the existing network protocol, and uses the shared memory in the host machine. Realize the interaction between virtual machines and speed up the business processing between virtual machines, so as to meet the scenarios that require high interaction performance between virtual machines.

For the above-mentioned target memory slice (the structures of other memory slices are similar), the memory slice includes a management control area and a service load area, and the management control area is used to store the identifier of the target memory slice, the remaining space size of the service load area, and write offset information; The service load area is used to store data sent by the first virtual machine to the second virtual machine;

Optionally, when the first virtual machine copies the data to be sent to the memory slice according to the start address and data length of the data to be sent, the following situations may occur, including:

Scenario 1: When the first virtual machine determines that the size of the remaining memory space in the service load area is less than or equal to the data length of the data, it determines that the size of the remaining memory space in the service load area is greater than the data length of the data before the preset time period exceeds the data length. The starting address and data length of the data to be sent are copied to the service load area.

The second situation: when the first virtual machine determines that the size of the remaining memory space in the service load area is less than or equal to the data length of the data, the first virtual machine fails to copy the data to be sent to the service load area. In this case, after a preset time interval, it is possible to retry to copy the data to be sent to the service load area.

The first virtual machine may call the sending interface to complete the above-mentioned copying of the data to be sent. When the sending interface is a blocking interface, it corresponds to the first situation; when the sending interface is a non-blocking interface, it corresponds to the second situation.

It should be understood that in the circular cache management system, when the write offset reaches the end of the service load area, that is, the remaining space in the service load area is 0, the content in the service load area can be overwritten according to the read offset to achieve Use of circular cache.

Referring to FIG. 5, an embodiment of the present invention provides a method for data interaction between virtual machines. On the receiving side, the method includes:

Step 500: The first virtual machine detects a set of target memory slices corresponding to the first virtual machine, and each target memory slice is used to store data sent to the first virtual machine.

The target memory slice set corresponding to the first virtual machine here refers to a set of memory slices used to store data sent by other virtual machines to the first virtual machine. For example, when the host machine includes N virtual machines, the target memory slice set corresponding to the first virtual machine includes (N-1) memory slices. The (N-1) memory slices are in one-to-one correspondence with other (N-1) virtual machines except the first virtual machine.

Step 510: When the first virtual machine detects that the write offset information in the first target memory slice is updated, it reads the data in the first target memory slice according to the write offset information and the read offset information in the first target memory slice , and update the read offset information.

The write offset information is used to indicate the write offset of the data stored in the first target memory slice, and the read offset information is used to indicate the read offset of the read data in the first target memory slice.

Specifically, when the first virtual machine detects that the write offset information in the first target memory slice is updated, it indicates that the virtual machine corresponding to the first target memory slice sends data to the first virtual machine, so that the Write offset information update.

Specifically, the first virtual machine learns the read offset of the read data according to the read offset information in the first target memory slice, starts to read the data from the read position, and can read up to the write offset information. The indicated write offset of the stored data, or the first virtual machine configures the preset data read length for each read data, when the difference between the write offset and the read offset is greater than or equal to the preset data read length , read the data of the preset data read length. When the difference between the write offset and the read offset is less than the preset data read length, the read can be performed next time. When the first virtual machine reads data in the first target memory slice, the read offset information needs to be updated.

The following takes FIG. 6 as an example to specifically describe the interaction process between virtual machines, wherein the eshare module is used to execute the interaction process between virtual machines. Each virtual machine can send data to other hosts in the host through the shared memory (eshare) module, and read the data sent by the virtual machines in other hosts to itself. Specifically, the eshare module can be a collection of multiple logical processing units, providing functions such as recording the memory address mapping corresponding to (N-1) memory slices, standard read and write interfaces, read and write mutual exclusion management, and read and write interface encapsulation.

First, when Host creates VM_0 through qemu, it allocates (N-1) memory slices to it.

For the (N-1) memory slices, the host physical address corresponding to each memory slice is converted into the corresponding host virtual address, the host virtual address is further converted into the guest physical address, and finally the guest physical address is converted is the virtual address of the guest.

When app0 in VM_0 needs to send data to appN in VM_N, the start address and data length of the data to be sent are determined. VM_0 determines the identifier of the target memory slice according to the identifier of VM_N, where the target memory slice is allocated by the Host for VM_0 to store the data sent by VM_0 to VM_N.

Further, VM_0 determines the initial copy position of the data to be sent according to the write offset information in the target memory slice, and starts the data to be sent from the initial copy position according to the start address and data length of the data to be sent, and stores the data to be sent from the initial copy position. Send data copy to the service load area of the target memory slice. After the data copy is completed, the write offset information is updated.

When VM_N detects that the write offset information in the memory slice used to store the data sent by VM_0 to VM_N is updated, it reads the data in the first target memory slice according to the write offset information and the read offset information in the memory slice, And update the read offset information.

Referring to FIG. 7 , an embodiment of the present invention provides an apparatus for data interaction between virtual machines, including:

The determining unit 70 is configured to determine the start address and data length of the data to be sent when the application in the device needs to send data to the application in the second virtual machine, the device and the second virtual machine belong to the same a host;

The analysis unit 71 is configured to determine the identifier of the target memory slice according to the identifier of the second virtual machine, where the target memory slice refers to the memory allocated by the host for the device to store the device and send it to the first virtual machine. Two memory slices of virtual machine data;

The processing unit 72 is configured to determine the initial copy position of the data to be sent according to the write offset information in the target memory chip, and copy the data to be sent according to the start address and data length of the data to be sent To the target memory slice, update the write offset information;

The write offset information is used to indicate the write offset of the data stored in the target memory slice.

In a possible implementation manner, the target memory slice includes a management control area and a service load area;

The management control area is used to store the identifier of the target memory slice, the remaining space size of the service load area, and the write offset information; the service load area is used to store the data sent by the device to the second virtual machine data;

When copying the data to be sent to the memory chip according to the start address and data length of the data to be sent, the processing unit 72 is specifically configured to:

When it is determined that the size of the remaining memory space of the service load area is less than or equal to the data length of the data, when it is determined that the size of the remaining memory space of the service load area is greater than the data length of the data before exceeding the preset time length, according to the The starting address and data length of the data to be sent are copied to the service load area.

In a possible implementation manner, the processing unit 72 is further configured to:

When it is determined that the size of the remaining memory space in the service load area is less than or equal to the data length of the data, it fails to copy the data to be sent to the service load area.

Referring to FIG. 8, an embodiment of the present invention provides an apparatus for data interaction between virtual machines, including:

A detection unit 80, configured to detect a set of target memory chips corresponding to the device, and each target memory chip is used to store data sent to the device;

The processing unit 81 is configured to read the first target memory according to the write offset information and the read offset information in the first target memory slice when detecting that the write offset information in the first target memory slice is updated data in the memory slice, and update the read offset information;

The write offset information is used to indicate the write offset of the data stored in the first target memory slice, and the read offset information is used to indicate the read offset of the read data in the first target memory slice offset.

In a possible implementation manner, the first target memory slice includes a management control area and a service load area;

The management control area is used to store the identifier of the first target memory slice, the remaining space size of the service load area, the write offset information and the read offset information;

The service load area is used for storing data sent by the second virtual machine to the device.

Referring to FIG. 9, an embodiment of the present invention provides an apparatus for data interaction between virtual machines, including:

an allocation unit 90 for allocating memory space as the shared memory of the virtual machine to be created;

A creation unit 91 is used to create N virtual machines, where N≥2, and N is a positive integer;

a processing unit 92, configured to divide the shared memory into N*(N-1) memory slices, and allocate (N-1) memory slices to each virtual machine;

Among them, the (N-1) memory slices allocated for the i-th virtual machine are in one-to-one correspondence with other (N-1) virtual machines except the i-th virtual machine, i≥1, and i is a positive integer .

In a possible implementation manner, when (N-1) memory slices are allocated to each virtual machine, the processing unit 92 is specifically used for:

For each memory slice allocated for the ith virtual machine, execute:

Converting the physical address of the host machine corresponding to the current memory slice into the virtual address of the host machine corresponding to the current memory slice;

converting the virtual address of the host machine corresponding to the current memory slice to the physical address of the client computer corresponding to the current memory slice;

Convert the client physical address corresponding to the current memory slice to the client virtual address corresponding to the current memory slice.

Referring to FIG. 10, an embodiment of the present invention provides a device for data interaction between virtual machines, including: a communication interface 1001 and a processor 1002;

the processor for:

When determining that the application in the device needs to send data to the application in the second virtual machine, determine the start address and data length of the data to be sent, the device and the second virtual machine belong to the same host; The identifier of the second virtual machine determines the identifier of the target memory slice, and the target memory slice refers to the memory slice allocated by the host for the device to store the data sent by the device to the second virtual machine ; Determine the initial copy position of the data to be sent according to the write offset information in the target memory chip, and copy the data to be sent through the communication interface according to the start address and data length of the data to be sent to the target memory slice, and update the write offset information; wherein, the write offset information is used to indicate the write offset of the data stored in the target memory slice.

Referring to FIG. 11, an embodiment of the present invention provides a device for data interaction between virtual machines, including: a communication interface 1101 and a processor 1102;

The processor 1101 is used for:

Detecting a set of target memory slices corresponding to the device, each target memory slice is used to store data sent to the device; when it is detected that the write offset information in the first target memory slice is updated, according to the write offset information and the read offset information in the first target memory slice through the communication interface to read the data in the first target memory slice, and update the read offset information; wherein, the write offset information uses In order to indicate the write offset of the data stored in the first target memory slice, the read offset information is used to indicate the read offset of the read data in the first target memory slice.

Referring to FIG. 12, an embodiment of the present invention provides a device for data interaction between virtual machines, such as a host, including: a communication interface 1201, a processor 1202, and a memory 1203;

The processor 1202 is used for:

Allocate the memory space in the storage 1203 as the shared memory of the virtual machine to be created; for creating N virtual machines, N≥2, N is a positive integer; the shared memory is divided into N*(N-1) memory slices, and allocate (N-1) memory slices for each virtual machine; among them, the (N-1) memory slices allocated for the i-th virtual machine are different from other (N-1) memory slices except the i-th virtual machine ( N-1) virtual machines are in one-to-one correspondence, i≥1, i is a positive integer;

The communication interface 1201 is used to communicate with each virtual machine in the host.

To sum up, the embodiments of the present invention provide a method and apparatus for data interaction between virtual machines, so as to solve the processing bottleneck problem that may occur in data transmission between virtual machines. The method includes: the first virtual machine determines the first When the application in the virtual machine needs to send data to the application in the second virtual machine, it determines the starting address and data length of the data to be sent; the first virtual machine determines the identifier of the target memory slice according to the identifier of the second virtual machine; The virtual machine determines the initial copy position of the data to be sent according to the write offset information in the target memory slice, and copies the data to be sent to the target memory slice according to the start address and data length of the data to be sent, and updates the write offset information. Therefore, compared with the prior art, which implements network packet transmission based on network cards, the method provided by the embodiment of the present invention only needs to copy data once, and does not depend on the existing network protocol, and realizes virtualized virtualisation through the shared memory in the host machine. The interaction between virtual machines can be accelerated to speed up the business processing between virtual machines, so as to meet the requirements of scenarios that require high interaction performance between virtual machines.

As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block in the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to the processor of a general purpose computer, special purpose computer, embedded processor or other programmable data processing device to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing device produce Means for implementing the functions specified in a flow or flow of a flowchart and/or a block or blocks of a block diagram.

These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory result in an article of manufacture comprising instruction means, the instructions The apparatus implements the functions specified in the flow or flow of the flowcharts and/or the block or blocks of the block diagrams.

These computer program instructions can also be loaded on a computer or other programmable data processing device to cause a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process such that The instructions provide steps for implementing the functions specified in the flow or blocks of the flowcharts and/or the block or blocks of the block diagrams.

Although preferred embodiments of the present invention have been described, additional changes and modifications to these embodiments may occur to those skilled in the art once the basic inventive concepts are known. Therefore, the appended claims are intended to be construed to include the preferred embodiment and all changes and modifications that fall within the scope of the present invention.

Obviously, those skilled in the art can make various changes and modifications to the embodiments of the present invention without departing from the spirit and scope of the embodiments of the present invention. Thus, provided that these modifications and variations of the embodiments of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.

Claims (8)

1. A method for data interaction between virtual machines, comprising: When the first virtual machine determines that the application in the first virtual machine needs to send data to the application in the second virtual machine, it determines the starting address and data length of the data to be sent, and the first virtual machine and the second virtual machine The virtual machines belong to the same host; The first virtual machine determines the identifier of the target memory slice according to the identifier of the second virtual machine, and the target memory slice refers to the memory allocated by the host machine for the first virtual machine to store the first virtual machine. a memory slice of data sent by the machine to the second virtual machine; The first virtual machine determines the initial copy position of the data to be sent according to the write offset information in the target memory slice, and copies the data to be sent according to the start address and data length of the data to be sent To the target memory slice, update the write offset information; Wherein, the write offset information is used to indicate the write offset of the stored data in the target memory slice; The target memory slice includes a management control area and a service load area; the management control area is used to store the identifier of the target memory slice, the size of the remaining space in the service load area, and the write offset information; the service The load area is used for storing data sent by the first virtual machine to the second virtual machine. 2. The method of claim 1, wherein the first virtual machine copies the to-be-sent data to the memory slice according to the start address and data length of the to-be-sent data, comprising: When the first virtual machine determines that the size of the remaining memory space in the service load area is less than or equal to the data length of the data, it is determined that the size of the remaining memory space in the service load area is larger than the data of the data before exceeding the preset time length When the length is set, the data to be sent is copied to the service load area according to the start address and data length of the data to be sent. 3. The method of claim 1, further comprising: When the first virtual machine determines that the size of the remaining memory space in the service load area is less than or equal to the data length of the data, the first virtual machine fails to copy the data to be sent to the service load area. 4. A method for data interaction between virtual machines, comprising: The first virtual machine detects a set of target memory slices corresponding to the first virtual machine, and each target memory slice is used to store data sent to the first virtual machine; When the first virtual machine detects that the write offset information in the first target memory slice is updated, it reads the first target memory according to the write offset information and the read offset information in the first target memory slice data in the memory slice, and update the read offset information; The write offset information is used to indicate the write offset of the data stored in the first target memory slice, and the read offset information is used to indicate the read offset of the read data in the first target memory slice offset; The first target memory slice includes a management control area and a service load area; the management control area is used to store the identifier of the first target memory slice, the remaining space size of the service load area, and the write offset information. and the read offset information; the service load area is used to store the data sent by the second virtual machine to the first virtual machine. 5. A device for data interaction between virtual machines, comprising: a determining unit, configured to determine the start address and data length of the data to be sent when the application in the device needs to send data to the application in the second virtual machine, the device and the second virtual machine belong to the same host machine; an analysis unit, configured to determine an identifier of a target memory slice according to an identifier of the second virtual machine, where the target memory slice refers to a location allocated by the host for the device to store the device and send it to the second virtual machine. The memory slice of the data of the virtual machine; The processing unit is configured to determine the initial copy position of the data to be sent according to the write offset information in the target memory chip, and copy the data to be sent to For the target memory slice, update the write offset information; The write offset information is used to indicate the write offset of the stored data in the target memory slice; the target memory slice includes a management control area and a service load area; the management control area is used to store the target memory slice The identifier of the memory slice, the remaining space size of the service load area, and the write offset information; the service load area is used to store the data sent by the device to the second virtual machine. 6. The apparatus according to claim 5, wherein, when the data to be sent is copied to the memory chip according to the start address and data length of the data to be sent, the processing unit is specifically used for : When it is determined that the size of the remaining memory space of the service load area is less than or equal to the data length of the data, when it is determined that the size of the remaining memory space of the service load area is greater than the data length of the data before exceeding the preset time length, according to the The starting address and data length of the data to be sent are copied to the service load area. 7. The apparatus of claim 5, wherein the processing unit is further configured to: When it is determined that the size of the remaining memory space in the service load area is less than or equal to the data length of the data, it fails to copy the data to be sent to the service load area. 8. A device for data interaction between virtual machines, comprising: a detection unit, configured to detect a set of target memory chips corresponding to the device, and each target memory chip is used to store data sent to the device; A processing unit, configured to read the first target memory according to the write offset information and the read offset information in the first target memory slice when detecting that the write offset information in the first target memory slice is updated data in the slice, and update the read offset information; The write offset information is used to indicate the write offset of the data stored in the first target memory slice, and the read offset information is used to indicate the read offset of the read data in the first target memory slice offset; The first target memory slice includes a management control area and a service load area; the management control area is used to store the identifier of the first target memory slice, the remaining space size of the service load area, and the write offset information. and the read offset information; the service load area is used to store the data sent by the second virtual machine to the device.

Images