CN104636076B - A kind of distributed block device drives method and system for cloud storage - Google Patents

Abstract

The invention discloses a distributed block device driving method and system for cloud storage. The method includes: converting the I/O request from the upper file system into an access message of the virtual block device, and then calling the network card driver to send the data packet to the server, so as to access the real physical storage device on the server. The invention can support remote loading and running of multiple operating systems, support command interfaces of multiple storage devices, and can efficiently complete data request and processing.

Description

A distributed block device driver method and system for cloud storage

technical field

The invention relates to the field of cloud computing, in particular to a distributed block device driving method and system for cloud storage.

Background technique

With the development of network technology, users synchronize important personal data to network storage devices through terminal devices for storage and management. As a new type of network computing model, cloud computing provides users with familiar operating systems and applications as network storage content. The cloud computing terminal remotely loads the operating system through the network, adopts the block flow method to execute dynamically, and provides users with various cross-platform applications and services. Since the cloud computing terminal itself does not preset a physical storage device, it needs to use a virtual storage device driver to interact with the physical storage resources on the cloud computing server, so as to provide the user with a virtual disk view. The driving method needs to be able to support remote loading and running of multiple operating systems, support command interfaces of multiple storage devices, and be able to efficiently complete data requests and processing.

The traditional driving method has the following problems:

First, traditional driving methods such as the Windows driver model and the Linux driver model are device driver models for a specific operating system, and are only applicable to a stand-alone environment, and cannot meet the cross-platform and networked requirements of cloud computing virtual storage systems .

Second, the device model based on the virtual machine and the network storage protocol can realize the remote startup and operation of the operating system. The famous achievements include Collective, ISR and OSCircular. Among them, Collective and ISR are systems developed based on the VMWare device model, while OS Circular uses the Xen-HVM device model. However, when the virtual machine is running, it occupies more memory and extends the I/O access path, which has a great impact on data access performance.

Contents of the invention

In view of the above, the present invention proposes a distributed block device driving method and system for cloud storage.

The present invention proposes a distributed block device driving method for cloud storage, including:

The VBD of the terminal receives the I/O request from the OS and sends it to the VSP;

VSP extracts the virtual disk parameters in the I/O request and sends them to VHBA;

The VHBA encapsulates the virtual disk parameters into a network data message and sends it to the server;

The VIOP of the server parses the received network data message and sends it to the PSP;

According to the mapping relationship between the virtual disk and the physical disk, the PSP converts the virtual disk parameters into physical disk parameters and accesses the real storage device;

The VHBA returns the processing result to the corresponding virtual disk device object and activates the VSP;

VBD copies the returned data to the specified memory to complete the I/O request.

Further, it also includes:

PSP sets up the monitor program for VBus. When it detects that VSP enters the initialization state, PSP will detect the physical disk, allocate and initialize the block device interface, make the corresponding field point to the device number of the physical device, register with VBus and write the physical disk. Disk information, waiting for VBus to write virtual disk information, VBus is a bus abstraction between VSP and PSP, used to complete the interaction between the virtual disk interface and the physical storage interface;

VSP enters the initialization state, registers with VBus, writes the virtual disk request report, allocates and initializes the interface structure, makes the corresponding field point to the device number of the virtual disk device, and then enters the initialization completion state;

After the PSP detects that a new virtual device is registered in the VBus, it obtains a virtual device request report in the VBus, and creates a virtual disk according to the virtual disk information in the report;

After the creation is completed, when the PSP monitors that the VSP enters the initialization completion state, it updates the physical disk information of the device to the VBus, enters the connection state, and prepares to process the I/O request passed from the VSP;

The VSP detects that the PSP enters the connection state, establishes a connection with the PSP, and reads the physical disk information from the VBus to complete the mapping between the virtual disk and the physical disk.

Further, it also includes:

The operating system traverses all drivers existing on the PCI bus and loads the VSP driver;

VSP initializes the specific virtual block device object, registers the interrupt handler, and the virtual block device object will call the VHBA driver as a PCI device;

VHBA initializes the virtual bus adapter controller object, allocates hardware resources through information in the PCI configuration space, registers interrupt services, and expands the next-level virtual interface bus structure VBus;

VHBA invokes the network scanning interface and sends a connection command to the server to confirm whether the storage device is ready to receive virtual I/O requests;

After getting a successful confirmation reply, VHBA sends a virtual disk scan command to obtain the device name, device number and virtual disk information;

According to the above device configuration information, the VHBA registers a new virtual disk device object with the VBus and enters the stage of creating a virtual block device.

The present invention proposes a distributed block device driver system for cloud storage, including a terminal, the terminal includes: VBD, VSP and VHBA, wherein:

VBD, receives the I/O request from the OS and sends it to VSP, copies the returned data to the specified memory, and completes the I/O request;

VSP extracts the virtual disk parameters in the I/O request and sends them to VHBA;

VHBA encapsulates the virtual disk parameters into a network data message and sends it to the server, returns the server processing result to the corresponding virtual disk device object, and activates the VSP;

Wherein, the server parses the received network data message, converts the virtual disk parameters into physical disk parameters according to the mapping relationship between the virtual disk and the physical disk, and accesses the real storage device.

Further, VSP enters the initialization state, registers with VBus, writes the virtual disk request report, allocates and initializes the interface structure, makes the corresponding field point to the device number of the virtual disk device, and then enters the initialization completion state; VSP detects that the PSP enters the connection state , establish a connection with the PSP, and read in the physical disk information from the VBus to complete the mapping between the virtual disk and the physical disk.

Further, VSP initializes the specific virtual block device object, registers the interrupt handler, and the virtual block device object will call the VHBA driver as a PCI device;

VHBA initializes the virtual bus adapter controller object, allocates hardware resources through the information of the PCI configuration space, registers interrupt services, and expands the next-level virtual interface bus structure VBus; VHBA calls the network scanning interface and sends a connection command to the server to confirm whether the storage device is ready Receive the virtual I/O request; after obtaining a successful confirmation reply, VHBA sends a virtual disk scanning command to obtain the device name, device number and virtual disk information; according to the above device configuration information, VHBA registers a new virtual disk device object with VBus, and enters The creation phase of a virtual block device.

The present invention proposes a distributed block device driver system for cloud storage, including a server, the server includes VIOP and PSP, wherein:

VIOP, analyze the network data message received from the terminal, and send it to the PSP;

PSP, according to the mapping relationship between the virtual disk and the physical disk, converts the virtual disk parameters carried in the network data message into physical disk parameters, and accesses the real storage device;

Among them, the terminal receives the I/O request from the OS, extracts the virtual disk parameters in the I/O request, encapsulates the virtual disk parameters into a network data packet and sends it to the server, and returns the server result to the corresponding virtual disk device Object that completes the I/O request.

Further, the PSP sets up a monitor program for the VBus. When it detects that the VSP enters the initialization state, the PSP will detect the physical disk, allocate and initialize the block device interface, make the corresponding field point to the device number of the physical device, register and write to the VBus Enter the physical disk information, and wait for the VBus to write the virtual disk information. The VBus is a bus abstraction between the VSP and the PSP to complete the interaction between the virtual disk interface and the physical storage interface; the PSP detects that a new virtual disk has been registered in the VBus. After the device is installed, obtain the virtual device request report in the VBus, and create a virtual disk according to the virtual disk information in the report; after the creation is completed, when the PSP detects that the VSP enters the initialization completion state, update the physical disk information of the device to the VBus, and enter the connection state , ready to process the I/O request passed from the VSP.

The present invention proposes a distributed block device driver system for cloud storage, including any of the above-mentioned terminals and any of the above-mentioned servers.

Further, a data exchange channel is established between the terminal and the server through a standard communication transmission protocol.

The invention describes an open architecture for the storage device driver in the cloud computing environment. The architecture is independent of specific operating systems, hardware platforms, and system I/O buses, and can meet the cross-platform and networked requirements of cloud computing virtual storage systems.

The invention can support remote loading and running of various operating systems, support instruction interfaces of various storage devices, and can efficiently complete data request and processing.

Description of drawings

The accompanying drawings described here are used to provide a further understanding of the present invention, and constitute a part of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention, and do not constitute improper limitations to the present invention. In the attached picture:

FIG. 1 is a schematic diagram showing a comparison between the device driving method in the cloud computing environment of the present invention and the traditional device driving method.

FIG. 2 is a schematic diagram of the architecture of the distributed block device driver system in an embodiment of the present invention.

FIG. 3 is a schematic flowchart of a distributed block device driving method in an embodiment of the present invention.

detailed description

Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that the relative arrangements and numerical values of components and steps set forth in these embodiments do not limit the scope of the present invention unless specifically stated otherwise.

At the same time, it should be understood that, for the convenience of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship.

The following description of at least one exemplary embodiment is merely illustrative in nature and in no way taken as limiting the invention, its application or uses.

Techniques, methods and devices known to those of ordinary skill in the relevant art may not be discussed in detail, but where appropriate, such techniques, methods and devices should be considered part of the Authorized Specification.

In all examples shown and discussed herein, any specific values should be construed as exemplary only, and not as limitations. Therefore, other examples of the exemplary embodiment may have different values.

It should be noted that like numerals and letters denote like items in the following figures, therefore, once an item is defined in one figure, it does not require further discussion in subsequent figures.

The invention describes an open architecture for the storage device driver in the cloud computing environment. The architecture is independent of a specific operating system, hardware platform, and system I/O bus. FIG. 1 is a schematic diagram comparing the device driving method under the cloud computing environment of the present invention with the traditional device driving method.

The left side of the arrow in Figure 1 is the traditional device driving method, which does not require a server and only drives a local disk. Traditional methods are OS-specific and based on stand-alone systems. The storage device driver is loaded during the initialization process of the operating system kernel. After that, the I/O requests generated when the application program is running are all completed by the driver to access the local disk. The right side of the arrow in Fig. 1 is the driving method of the present invention. Among them, the terminal is on the left and the server is on the right. The invention is independent of specific operating system, hardware platform and system bus, and the invention extends the drive to the network environment.

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be described in further detail below in conjunction with specific embodiments and with reference to the accompanying drawings.

FIG. 2 is a schematic diagram of the architecture of the distributed block device driver system in an embodiment of the present invention. The system includes terminals and servers. Wherein, the terminal includes a front-end module, and the server includes a back-end module.

Both the front-end module and the back-end module are respectively loaded into the operating systems of the terminal and the server with kernel-state drivers, and the two have different structures and functions. According to different functional levels, the front-end module and the back-end module respectively establish their own driver stacks. The front-end module and the back-end module will be described respectively below with reference to the accompanying drawings.

1. Front-end module driver stack

The front-end module is responsible for obtaining the I/O request from the upper file system, converting the request into a virtual block device access message, and then calling the network card driver to send the data packet to the server. The front-end module driver protocol stack runs in the kernel mode of the terminal, and is divided into three parts according to the functional structure.

(1) Virtual block device driver layer (virtual block device, VBD).

The main task of VBD is to obtain requests from outside the I/O subsystem of the operating system, convert them into feasible virtual block device access requests, and send them down to the middle-level driver, that is, to the VSP, and copy the data returned by the VSP Go to the specified memory to complete the I/O request. The VBD driver provides a virtual block device interface, which has all the attributes of a block device, and can be specifically represented as a virtual disk device, a virtual tape device, a virtual CD-ROM device, and so on. In order to meet the open requirements of the system, the VBD component also provides the operating system with a PnP interface, a device power management interface and an I/O data access interface that comply with relevant standards. The PnP interface provides functions such as starting, configuring, stopping, and deleting virtual storage device entities. The device power management interface transmits the power signal of the virtual storage device to the server, and the server provides the power management strategy, and the terminal sets the corresponding power state according to the result returned by the server. In terms of I/O data interface, VBD provides comprehensive support for common data interaction interfaces ATA and SCSI.

(2) Virtual storage interface driver layer (virtual storage port, VSP).

VSP mainly implements various functions of the front-end module, including bus scanning, error recovery, command queuing, communication with the kernel, and various other initialization functions. It receives I/O requests from VBD, extracts the virtual disk parameters in the I/O requests and sends them to the lower layer, namely VHBA. Until the processing of the I/O request is completed, the VSP returns the processing result to the VBD. The VSP also performs error handling such as timeouts or command errors. Therefore, VSP is the bus layer driver of the virtual device interface, which provides a unified interface for the upper and lower layer drivers and performs bus scanning to complete device enumeration, data transmission and error handling.

(3) Virtual host bus adapter driver layer (virtual host bus adapter, VHBA).

VHBA uses a suitable network transmission protocol to realize the virtual interface bus structure VBus, and completes the transmission of status, data, commands and messages between the front-end module and the back-end module by interacting with the server. VHBA is the driver of the virtual controller, which is related to the specific hardware adapter. For example, it can call the network card driver to send an I/O request message to the server. That is, the VHBA encapsulates the virtual disk parameters into a network data message and sends it to the server, returns the processing result of the server to the corresponding virtual disk device object, and activates the VSP.

2. Back-end module driver stack

The back-end module is mainly responsible for processing virtual I/O requests from the front-end modules of multiple terminals, parsing the access information, queuing and scheduling according to the request types, constructing real block device access instructions, and passing them to physical devices for processing. Finally, the organized data is sent back to the corresponding front-end module. The back-end module driver is located in the kernel state of the server, and is divided into two levels according to the functional structure.

(1) Virtual I/O processing driver layer (virtual I/O processor, VIOP).

VIOP analyzes, classifies, queues and schedules the virtual I/O requests from the front-end module, and transmits the physical device access instructions to the PSP layer for processing. VIOP encapsulates the data or response from the PSP layer and transmits it to the corresponding front-end module through the network. The function of the server-side transport protocol is completed in VIOP.

(2) Physical storage interface driver layer (physical storage port, PSP).

PSP has nothing to do with network transmission, and mainly deals with physical device access instructions from the VIOP device driver layer. According to the mapping relationship between the virtual disk and the physical disk, the PSP converts the virtual disk parameters into physical disk parameters to access real storage devices. It can classify and aggregate the received block interface access instructions according to the interface type, and call the corresponding physical storage device driver for the interface type to process. PSP is also responsible for error command processing and command state maintenance.

The invention can support remote loading and running of various operating systems, support instruction interfaces of various storage devices, and can efficiently complete data request and processing. In the present invention, the virtual block device driver interacts with the physical storage resources on the cloud computing server, thereby providing a view of a virtual disk for the user.

The present invention adopts block-level virtualization technology and provides corresponding virtual block device interface to process I/O requests. The driving method adopts a client/server structure, and the driver program is divided into two parts, a front-end module and a back-end module, according to code functions. The front-end module runs on the cloud computing terminal and provides a virtual device interface for the terminal OS without specific hardware code. The back-end module runs on the cloud computing server and provides the interface between the I/O adapter and its physical device without requiring terminal OS-related codes. A standard communication transmission protocol such as iSCSI or NBD is used between the two modules to establish a data exchange channel between the two physically separated modules. The workflow of the front-end module and the back-end module will be described in detail below with reference to the accompanying drawings.

The cloud computing terminal does not have an operating system locally, so after booting, the BIOS needs to call the service provided by the PXE protocol to remotely load the operating system kernel. After the OS kernel is loaded, the front-end module protocol stack is loaded during the kernel initialization process. The loading of the OS system program and the I/O requests generated by the running of the application program after the OS starts are all handled by the front-end module and the back-end module working in the server. The workflow mainly includes three processes: bus scanning, virtual block device creation and I/O request processing.

Bus scanning, virtual block device creation, and I/O request processing are sequential execution processes, that is, bus scanning prepares for virtual block creation, virtual block device creation establishes a mapping between virtual disks and physical disks, and then performs I/O O Request processing needs to be converted based on the established mapping relationship. The above three processes will be described in detail below.

(1) Bus scan

Step 1: During the startup process of the computer system, the operating system scans the default PCI (Peripheral Component Interconnect, external device interconnection bus) root node, traverses all the drivers existing on the PCI bus, and loads the VSP driver.

Step 2: VSP initializes a specific virtual block device object and registers an interrupt handler. The virtual block device object will call the VHBA driver as a PCI device.

Step 3: VHBA initializes the virtual bus adapter controller object, allocates hardware resources through information in the PCI configuration space, registers interrupt services, and expands the next-level virtual interface bus structure VBus. VBus is a bus abstraction between VSP and PSP, which is used to complete the interaction between the virtual disk interface and the physical storage interface.

Step 4: The VHBA calls the network scanning interface, and sends a connection command to the back-end module to confirm whether the storage device is ready to receive the virtual I/O request.

Step 5: After getting a successful confirmation reply, the VHBA sends a virtual disk scanning command to obtain the device name, device number and configuration information of the virtual disk (including the number of virtual disks and the size of each disk, etc.).

Step 6: According to the above device configuration information, the VHBA registers a new virtual disk device object with the VBus. Start the creation phase of the virtual block device.

(2) Virtual block device creation

VSP and PSP communicate and negotiate through the VBus bus abstraction provided by the bottom layer. The two work together to create a virtual block device. The process is as follows:

Step 1: PSP first sets up the monitor program for VBus. When it detects that VSP enters the initialization state, PSP will detect the physical disk, allocate and initialize the block device interface and other structures, make the corresponding fields point to the device number of the physical device, and send data to the VBus Register and write physical disk information, and wait for VBus to write virtual disk information.

Step 2: VSP enters the initialization state, registers with VBus, writes the virtual disk request report, allocates and initializes the interface structure, makes the corresponding field point to the device number of the virtual disk device, and then enters the initialization completion state.

Step 3: After the PSP monitoring function detects that a new virtual device is registered in the VBus, it obtains a virtual device request report in the VBus, and starts creating a virtual disk according to the virtual disk configuration information in the report.

Step 4: After the creation is completed, when the PSP monitors that the VSP enters the initialization completion state, update the physical sector number and sector size of the device to the VBus, and then enter the connection state, ready to process the I/O passed from the VSP ask.

Step 5: VSP detects that the PSP is in the connection state, immediately establishes a connection with the PSP, and reads information such as the number of physical sectors and sector size from the VBus, and then completes the mapping relationship table between the virtual disk and the physical disk. According to the mapping relationship table, VBD starts to register, allocate and configure common disk objects, and notify the OS that a new block device has been created. Thereafter, the OS can use the virtual disk as if it were a local disk.

(3) I/O request processing

FIG. 3 is a schematic flowchart of a distributed block device driving method in an embodiment of the present invention. This flow shows the I/O request processing flow. Include the following steps:

Step 1: After the VSP receives the I/O request from the OS from the VBD, it first adds the request to the access request queue, and adopts an appropriate queue management algorithm (such as first-in-first-out queue and weighted fair queue, etc.), and selects from the queue Appropriate I/O requests for further processing.

Step 2: VSP takes out one I/O request from the queue each time, that is to say, the next I/O request in the queue will be processed only after the processing of the current I/O request is completed.

Step 3: VSP extracts the relevant parameters in the I/O request data structure (including the start address of the requested disk block, block size and buffer pointer), calls the read/write interface of VHBA, and encapsulates the relevant parameters into a disk access read /write command, and then encapsulate it as a network data packet and send it to the server VIOP.

Step 4: VIOP parses the received request message, and classifies virtual I/O requests according to classification rules (by user classification or I/O type classification). The scheduling policy assigns different priorities to different types of I/O queues, for example, the highest priority may be assigned to system-type I/O requests.

Step 5: After the PSP aggregates the received I/O requests, it calls the physical disk interface to access the real storage device. After the server returns the processing result, the VHBA returns the result to the corresponding virtual disk device object and activates the VSP.

Step 6: VBD copies the returned data to the specified memory, and notifies the OS to complete an I/O request.

The present invention is not limited to computing terminals in the cloud computing environment, but can also be applied to various computing devices, such as various application servers.

So far, the present invention has been described in detail. Certain details well known in the art have not been described in order to avoid obscuring the inventive concept. Based on the above description, those skilled in the art can fully understand how to implement the technical solutions disclosed herein.

The method and apparatus of the invention may be implemented in many ways. For example, the method and device of the present invention can be realized by software, hardware, firmware or any combination of software, hardware, and firmware. The above sequence of steps used in the method is for illustration only, and the steps of the method of the present invention are not limited to the sequence specifically described above unless specifically stated otherwise. Furthermore, in some embodiments, the present invention can also be implemented as programs recorded in recording media, the programs including machine-readable instructions for realizing the method according to the present invention. Thus, the present invention also covers a recording medium storing a program for executing the method according to the present invention.

Although some specific embodiments of the present invention have been described in detail through examples, those skilled in the art should understand that the above examples are for illustration only, rather than limiting the scope of the present invention. Those skilled in the art will appreciate that modifications can be made to the above embodiments without departing from the scope and spirit of the invention. The scope of the invention is defined by the appended claims.

Claims (10)

1. A kind of 1. distributed block device drives method for cloud storage, it is characterised in that：

   The Virtual Block Device driving layer VBD of terminal receives the I/O from operating system and asks and send the drive of virtual memory interface to Dynamic layer VSP；

   VSP extracts the virtual disk parameter in I/O requests and sends virtual bus adapter driving layer VHBA to；

   The virtual disk parameter is encapsulated as network data message and is sent to server by VHBA；

   The network data message that the virtual i/o processing driving layer VIOP parsings of server receive, sends physical store interface drive to Dynamic layer PSP；

   PSP joins the virtual disk Parameter Switch for physical disk according to the mapping relations between virtual disk and physical disk Number, accesses real storage device；

   VIOP returns to PSP to the result of network data message the VHBA of terminal；

   The result is returned to corresponding virtual disk device object by VHBA, activates VSP；

   The result is returned to VBD by VSP；

   The result that VBD returns to VSP is copied in the internal memory specified, and completes I/O requests；

   Wherein, methods described also includes：PSP sets the oracle listener to VBus, after detecting that VSP enters init state, PSP will detect physical disk, distribute simultaneously INIT block equipment interface, the equipment for making respective field therein point to physical equipment Number, registered to VBus and write physical disk information, wait VBus write-in virtual disk informations, VBus is between VSP and PSP A kind of bus abstract, to complete the interaction between virtual disk interface and physical store interface.
2. 2. it is used for the distributed block device drives method of cloud storage according to claim 1, it is characterised in that also include：

   VSP enters init state, is registered to VBus, writes virtual disk request report, distributes simultaneously initialization interface knot Structure, respective field therein is set to enter initialization completion status after pointing to the device number of virtual disk equipment；

   PSP is detected new virtual unit is have registered in VBus after, obtain VBus in virtual unit request report, according in report Virtual disk information create virtual disk；

   After the completion of establishment, when PSP monitors that VSP enters initialization completion status, the physical disk of more new equipment into VBus Information, into connection status, prepare the I/O requests that processing passes over from VSP；

   VSP detects that PSP enters connection status, establishes the connection with PSP, and physical disk information is read in from VBus, completes Mapping between virtual disk and physical disk.
3. 3. the distributed block device drives method according to claim 1 or claim 2 for cloud storage, it is characterised in that also include：

   All drivers present on operating system traversal pci bus, loading VSP drivings；

   VSP initializes specific Virtual Block Device object, registers interrupt handling routine, Virtual Block Device object is as a PCI Equipment will call VHBA drivings；

   VHBA initialization of virtual bus adapter controller objects, hardware resource, registration are distributed by the information of pci configuration space Service is interrupted, expands next stage virtual interface bus structures VBus；

   VHBA calls network sweep interface, and sending link order to server confirms whether storage device prepares to receive virtual i/o Request；

   Obtain one confirms successfully to reply after, VHBA sends the implementor name of virtual disk scan instruction acquisition virtual disk equipment Title, device number and virtual disk information；

   According to storage device title, storage device number and the virtual disk information of virtual disk equipment, VHBA registers new to VBus Virtual disk device object, into the establishing stage of Virtual Block Device.
4. 4. a kind of distributed block device driver system for cloud storage, including terminal, the terminal include：Virtual Block Device drives Layer VBD, virtual memory interface driver layer VSP and virtual bus adapter driving layer VHBA, it is characterised in that：

   VBD, receive the I/O from operating system and ask and send VSP to, the place to network data message that server is returned Reason result is copied in the internal memory specified, and completes I/O requests；

   VSP, extract the virtual disk parameter in I/O requests and send VHBA to, by processing of the server to network data message As a result VBD is returned to；

   VHBA, the virtual disk parameter is encapsulated as network data message and is sent to server, by server to network data The result of message returns to corresponding virtual disk device object, activates VSP；

   Wherein, the network data message that server parsing receives, will according to the mapping relations between virtual disk and physical disk The virtual disk Parameter Switch is physical disk parameter, accesses real storage device；

   VSP enters init state, is registered to VBus, writes virtual disk request report, distributes simultaneously initialization interface knot Structure, respective field therein is set to enter initialization completion status after pointing to the device number of virtual disk equipment, wherein VBus is A kind of bus abstract between VSP and PSP.
5. 5. it is used for the distributed block device driver system of cloud storage according to claim 4, it is characterised in that：

   VSP detects that physical store interface driving layer PSP enters connection status, establishes the connection with PSP, and read in from VBus Physical disk information, complete the mapping between virtual disk and physical disk.
6. 6. it is used for the distributed block device driver system of cloud storage according to claim 5, it is characterised in that：

   VSP initializes specific Virtual Block Device object, registers interrupt handling routine, Virtual Block Device object is as a PCI Equipment will call VHBA drivings；

   VHBA initialization of virtual bus adapter controller objects, hardware resource, registration are distributed by the information of pci configuration space Service is interrupted, expands next stage virtual interface bus structures VBus；VHBA calls network sweep interface, sends and connects to server Connect whether confirmation command storage device prepares to receive virtual i/o request；Obtain one confirms successfully reply after, VHBA send void Intend device name, device number and virtual disk information that disk scanning instruction obtains virtual disk equipment；Set according to virtual disk Standby storage device title, storage device number and virtual disk information, VHBA register new virtual disk device object to VBus, Into the establishing stage of Virtual Block Device.
7. 7. a kind of distributed block device driver system for cloud storage, including server, the server includes virtual i/o processing Drive layer VIOP and physical store interface driving layer PSP, it is characterised in that：

   VIOP, the network data message received from terminal is parsed, sends PSP to；

   PSP, according to the mapping relations between virtual disk and physical disk, the virtual disk carried in network data message is joined Number is converted to physical disk parameter, accesses real storage device；

   Wherein, terminal receives the I/O requests from operating system, the virtual disk parameter in extraction I/O requests, by the virtual magnetic Disk parameter is encapsulated as network data message and is sent to server, and server results are returned into corresponding virtual disk and set Standby object, complete I/O requests；

   PSP sets the oracle listener to VBus, and after detecting that VSP enters init state, PSP will detect physical disk, point With simultaneously INIT block equipment interface, respective field therein is pointed to the device number of physical equipment, registered and write to VBus Enter physical disk information, wait VBus write-in virtual disk informations, VBus is between virtual memory interface driver layer VSP and PSP A kind of bus abstract, to complete the interaction between virtual disk interface and physical store interface.
8. 8. it is used for the distributed block device driver system of cloud storage according to claim 7, it is characterised in that：

   PSP is detected new virtual unit is have registered in VBus after, obtain VBus in virtual unit request report, according in report Virtual disk information create virtual disk；After the completion of establishment, when PSP monitors that VSP enters initialization completion status, to The physical disk information of more new equipment in VBus, into connection status, prepare the I/O requests that processing passes over from VSP.
9. A kind of 9. distributed block device driver system for cloud storage, it is characterised in that：Including any drive of claim 4 to 6 The server that the terminal and any drive system of claim 7 or 8 that dynamic system includes include.
10. 10. it is used for the distributed block device driver system of cloud storage according to claim 9, it is characterised in that：Terminal kimonos Between business device data exchange channels are established by the communication transport protocols of standard.