WO2012159309A1

WO2012159309A1 - Method and device for processing data by utilizing threads

Info

Publication number: WO2012159309A1
Application number: PCT/CN2011/076552
Authority: WO
Inventors: 周毅
Original assignee: 华为技术有限公司
Priority date: 2011-06-29
Filing date: 2011-06-29
Publication date: 2012-11-29
Also published as: CN103608775A; CN103608775B

Abstract

A method and a device for processing data by utilizing threads are provided. The method comprises: when the data, to which given processing needs to be performed, are received by a first thread, whether the critical resource required by performing the given processing to the data has been occupied by another thread or not is detected, and the data are sent to a processing waiting queue of another thread when the detection result is that the critical resource required by performing the given processing to the data has been occupied by another thread. The performance of data processing is improved by using the method and the device.

Description

Technical field

The present invention relates to the field of communications technologies, and in particular, to a method and apparatus for processing data using threads. Background technique

Asynchronous Transfer Mode (ATM) technology is a connection-oriented communication technology developed on the basis of the principle of packet switching. It transmits data on a physical link in units of ATM cells.

In the ATM technology, a physical link is usually divided into multiple virtual paths (VPs, Virtual Paths). Each VP is identified by a virtual channel identifier (VPI), and each VP can carry multiple virtual paths (VC, Virtual). Channel), each VC is identified by a virtual path identifier (VCI). The VCs on different VPs constitute a VC connection (VCC) for transmitting ATM cells, which is uniquely identified by the VCI of the VC and the VPI of the VP carrying the VC. The header portion of the transport ATM cell includes an identifier for identifying the VC connection to which it is transmitted, gp VCI and VPI.

The ATM network is divided into three layers: ATM Adaptation Layer (AAL), ATM layer, and physical layer. The AAL layer is located at the upper layer of the ATM and is responsible for splitting the data transmitted from the upper layer into ATM cells and transmitting them to the ATM layer, and reassembling the ATM cells from the ATM layer into data and transmitting them to the upper layer. The AAL layer is further divided into two sub-layers, namely a convergence (CS, Convergence Sublayer) sub-layer and a disassembly (SAR, Segmentation And Reassembly) sub-layer, wherein the CS sub-layer mainly implements a convergence function for transforming data into CS data units. The SAR sublayer is mainly used to segment or recombine CS data in units of ATM cells.

The AAL layer has different AAL protocols, each of which is used to support different traffic or service types used on an ATM network. ATM Adaptation Layer 2 (AAL2) is one of the AAL protocols that supports connection-oriented, variable bit rate services. The AAL2 protocol is widely used, and the common services are compressed packet voice communication and compressed video transmission.

The AAL2 protocol processing is more complicated. Specifically, in the AAL2 protocol, when coming from the upper level (should When a service specific segmentation and reassembly (SSSAR) message (for example, a voice message) is to be transmitted via a VC connection, AAL2 fragmentation processing needs to be performed on the SSSAR message, gp: first The SSSAR message is split into a Common Part Sublayer (CPS) message, and then the split CPS message is filled into the ATM cell and sent to the ATM layer, where a CPS message is sent. A text can be filled across multiple ATM cells, and a single ATM cell can contain multiple CPS messages. When an ATM cell received from the ATM layer via the VC connection is to be transmitted to the upper layer, AAL2 reassembly processing needs to be performed on the ATM cell, gP: first extracting the CPS message from the ATM cell, and then extracting the extracted CPS message. Reconstitute the SSSAR message and send it to the upper layer. Due to the complexity of the AAL2 protocol, its data processing performance is low. Summary of the invention

In view of the above-discussed deficiencies of the prior art, embodiments of the present invention provide a method and apparatus for processing data using threads that use multiple threads to process data in parallel, thereby improving the processing performance of the data.

An embodiment of the present invention provides a method for processing data in a first thread, including the steps of: detecting, when the first thread receives data that needs to perform specified processing, detecting that the specified processing is performed on the data Whether the resource has been occupied by another thread; and, when the detection result is positive, the data is sent to the processing wait queue of the other thread.

An embodiment of the present invention provides an apparatus for processing data in a first thread, including: a detecting module, configured to: when the first thread receives data that needs to perform specified processing, detecting that performing the data on the data Specifying whether a critical resource required for processing has been occupied by another thread; and, a sending module, configured to send the data to a processing wait queue of the another thread when the detection result is positive.

The method and apparatus of the embodiments of the present invention use multiple threads to process data in parallel, and data that needs to be processed using the same critical resource will be processed by only one thread, and different threads process data requiring different critical resources, so Synchronization and order preservation are not required between threads, which improves data processing performance. DRAWINGS

The objects, features, features and advantages of the present invention will become apparent from the following detailed description It’s obvious. among them:

1 is a flow chart showing a method of processing data using a thread according to a first embodiment of the present invention;

FIG. 2 shows a specific example of performing AAL2 fragmentation processing on data by using the method of the first embodiment of the present invention;

FIG. 3 shows a specific example of performing AAL2 recombination processing on data by using the method of Embodiment 1 of the present invention;

4 illustrates an apparatus for processing data in a first thread in accordance with an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION Embodiments of the present invention provide a method and apparatus for processing data using threads that employ multiple threads to process data in parallel, and data that needs to be processed using the same critical resource will be processed by only one of the threads. Different threads process data that requires different critical resources, so synchronization and ordering are not required between threads, which speeds up data processing. Here, as understood by those skilled in the art, a critical resource refers to a resource that is only allowed to be occupied by one thread at a time, including but not limited to a printer, a tape drive, a message buffer queue, a variable, an array, and/or a buffer.

Various embodiments of the present invention will be described in detail below with reference to the drawings. 1 is a flow chart showing a method of processing data using a thread according to a first embodiment of the present invention. In this embodiment, multiple threads are used to process multiple data streams that need to perform specified processing in parallel, and each data of each data stream needs to be processed using the same critical resource, but each data stream needs to use different critical resources. Processing, each thread has a processing wait queue for saving pending data.

As shown in FIG. 1, in step S110, when any one of the plurality of threads receives the data D that needs to perform the specified processing, the thread P detects whether the critical resource R required to perform the specified processing on the data D has been Occupied by another thread. In step S120, when the detection result of step S110 is affirmative, that is, the critical resource R has been occupied by another thread, the thread P sends the data D to the processing wait queue of the other thread, to the data by the other thread. D is processed. Then, thread P continues to receive the next data 0 that needs to perform the specified processing.

In step S130, when the detection result of step S110 is negative, that is, the critical resource R has not been occupied by another thread, the thread P occupies the critical resource R.

In step S140, the thread P performs specified processing on the data D using the occupied critical resource R. In step S150, after the specified processing is performed on the data D by using the critical resource R, the thread P checks whether the data sent by the other thread is stored in the processing waiting queue. In this embodiment, after the thread P occupies the critical resource R, when other threads receive the data that needs to be processed by using the critical resource R, the data is sent to the processing wait queue of the thread P, to be performed by the thread P. deal with.

In step S160, when the result of the check in step S150 is affirmative, that is, the data sent by the other thread is stored in the processing waiting queue of the thread P, the thread P performs the designation of the data held in the processing queue of the thread P by using the critical resource R. deal with.

In step S170, when the check result of step S150 is negative, that is, the data sent by the other thread is not stored in the processing wait queue of the thread P, or the data held in the processing wait queue of the thread P has been processed, the thread P is released. Occupation of critical resource R. Then, thread P continues to receive other data that needs to perform the specified processing.

As can be seen from the above description, according to the solution of the first embodiment, multiple threads are used to process data in parallel, and data that needs to be processed using the same critical resource will be processed by only one thread, and different thread processing requires different criticalities. The data of the resource, so there is no need for synchronization and order preservation between threads, thus speeding up the processing of data. A specific example of processing data using the method disclosed in the first embodiment will be given below. FIG. 2 shows a specific example of performing AAL2 fragmentation processing on data using the method of the first embodiment of the present invention. In this example, multiple threads are used to perform AAL2 fragmentation processing in parallel on multiple data streams that need to be transmitted via different VCCs, where each data stream includes multiple SSSAR messages. As shown in FIG. 2, in step S210, when any one of the plurality of threads receives a SSSAR message SB from the upper layer that needs to perform AAL2 fragmentation processing, the thread C attempts to VC the message SB to be transmitted. The connection XT applies a mutex to detect whether the critical resource ZY required to perform AAL2 fragmentation processing on the message SB is already occupied by another thread. Here, the critical resource ZY may include, but is not limited to, a context in which AAL2 fragmentation processing is performed on SSSAR messages.

In step S220, when the thread C fails to apply the mutex lock to the VC connection XT, it indicates that the critical resource ZY has been occupied by another thread, and the thread C sends the message SB to the processing waiting queue of the other thread. Another thread performs AAL2 fragmentation on the message SB. Then, thread C continues to receive the next SSSAR message that needs to perform AAL2 fragmentation processing from the upper layer.

In step S230, when the thread C successfully applies the mutex lock to the VC connection XT, it is determined that the critical resource ZY has not been occupied by another thread, so that the thread C occupies the critical resource by applying a mutex lock to the VC connection XT. ZY.

In step S240, the thread C performs AAL2 fragmentation processing on the message SB by using the occupied critical resource ZY to divide the message SB into ATM cells, wherein each ATM cell includes a VPI for identifying the VC connection XT. And VCI.

In step S250, after performing AAL2 fragmentation processing on the message SB by using the critical resource ZY, the thread C checks whether the SSSAR message sent by another thread is stored in the processing waiting queue. In this example, after thread C occupies critical resource ZY, when other threads receive SSSAR messages that need to be processed using critical resource ZY, the SSSAR message is sent to thread C's processing wait queue.

In step S260, when the check result of step S250 is affirmative, that is, when the processing wait queue of the thread C stores the SSSAR message sent by another thread, the thread C uses the occupied critical resource ZY to process the thread C in the waiting queue. The saved SSSAR message performs AAL2 fragmentation processing to be segmented into ATM cells, where each ATM cell includes a VPI and a VCI for identifying the VC connection XT.

In step S270, when the check result of step S250 is negative, that is, the SSSAR message sent by another thread is not stored in the processing waiting queue of the thread C, or the thread C processes the SSSAR message stored in the processing waiting queue, Thread C releases the occupancy of critical resource ZY. Then, thread C continues to receive other SSSAR messages from the upper layer that need to perform AAL2 fragmentation processing. FIG. 3 shows a specific example of AAL2 recombination of data using the method of Embodiment 1 of the present invention. In this example, multiple threads are used to perform AAL2 reassembly processing in parallel on multiple data streams received from different VC connections, where each data stream includes multiple ATM cells.

As shown in FIG. 3, in step S310, when any one of the plurality of threads receives an ATM cell XY that needs to perform AAL2 reassembly processing from the ATM layer, the thread K is included according to the ATM cell XY. The VPI and the VCI recognize the VC connection ST receiving the ATM cell XY, and attempt to apply a mutex lock to the VC connection ST to detect whether the critical resource U required for performing the AAL2 reassembly processing on the ATM cell XY has been used by another thread. Occupied. Here, the critical resource LJ may include, but is not limited to, a context in which AAL2 reassembly processing is performed on ATM cells.

In step S320, when the thread K fails to apply the mutex lock to the VC connection ST, it indicates that the critical resource U has been occupied by another thread, and the thread K sends the ATM cell XY to the processing waiting queue of the other thread. The other thread performs AAL2 reassembly processing on the ATM cell XY. Then, thread K continues to receive the next ATM cell that needs to perform AAL2 reassembly processing from the ATM layer.

In step S330, when the thread K successfully applies the mutex lock to the VC connection ST, it indicates that the critical resource U has not been occupied by another thread, so that the thread K occupies the critical resource by applying a mutex lock to the VC connection ST. U.

In step S340, the thread K performs AAL2 reassembly processing on the ATM cell XY by using the occupied critical resource U to obtain an SSSAR message.

In step S350, after the AAL2 reorganization processing is performed on the ATM cell XY by using the critical resource U, the thread K checks whether the ATM cell transmitted by another thread is stored in the processing waiting queue. In this example, after the thread K occupies the critical resource U, when other threads receive the ATM cell that needs to be processed by the critical resource U, the ATM cell is sent to the processing wait queue of the thread K, Thread K is processed.

In step S360, when the check result of step S350 is affirmative, that is, when the processing wait queue of the thread K stores ATM cells transmitted by other threads, the thread K uses the occupied critical resource U to process the thread K in the waiting queue. The saved ATM cell performs AAL2 reassembly processing to obtain an SSSAR message.

In step S370, when the check result of step S350 is negative, that is, the ATM cell sent by another thread is not stored in the processing waiting queue of the thread K, or the thread K processes the ATM cell saved in the processing waiting queue, Thread K releases the occupation of critical resource U. Then, thread K continues to receive other ATM cells from the ATM layer that need to perform AAL2 reassembly processing. It will be understood by those skilled in the art that although in the above specific example description, the mutex is applied to the VC connection to detect whether the corresponding critical resource has been occupied by another thread, the present invention is not limited thereto. In other embodiments of the present invention, for example, a mutex may also be directly applied to the critical resource to detect whether the corresponding critical resource has been occupied by another thread.

It should be understood by those skilled in the art that the processing wait queue of each thread may include a plurality of processing wait queues having different priorities, when one thread sends the received data to the processing wait queue of another thread for execution by the other thread. In processing, the priority of the received data is first identified, and then the received data is sent to a processing wait queue in the other thread corresponding to the priority of the received data. When a thread processes the data held by its processing wait queues, it processes the data held by each processing wait queue according to the priority of its processing wait queues from high to low.

Those skilled in the art should understand that the method disclosed in the embodiments of the present invention can be used not only to perform AAL2 fragmentation processing and AAL2 reassembly processing on data, but also to perform other types of processing on data. Referring now to Figure 4, there is shown an apparatus for processing data in a first thread in accordance with an embodiment of the present invention, wherein the first thread is any of a plurality of threads.

As shown in FIG. 4, the device 400 includes a detection module 410 and a sending module 420, where the detecting module 410 is configured to perform a specified processing on the data received by the first thread when the first thread receives the data that needs to perform the specified processing. Whether the required critical resource has been occupied by another thread, and the sending module 420 is configured to send the data received by the first thread to the processing waiting queue of the other thread when the detection result of the detecting module 410 is positive.

The device 400 may further include an occupation module 430 and an execution module 440, wherein the occupation module 430 is configured to occupy the critical resource for the first thread when the detection result of the detection module 410 is negative, and the execution module 440 is configured to utilize the The occupied critical resource for the first thread performs a specified process on the data received by the first thread.

The sending module 420 may further include an identifying module and a transmitting module, wherein the identifying module is configured to identify a priority of the data received by the first thread when the detection result of the detecting module 410 is positive, and the transmitting module is configured to The data received by one thread is sent to the processing wait queue of the other thread corresponding to the priority of the data received by the first thread. The device 400 may further include an checking module 450 and a releasing module 460, wherein the checking module 450 is configured to check whether the data sent by the other thread is saved in the processing waiting queue of the first thread after performing the specified processing on the data received by the first thread. The release module 460 is configured to release the occupied critical resource for the first thread when the check result of the check module 450 is negative, or when the execution module 440 processes the data held by the processing queue of the first thread. The execution module 440 is further configured to: when the check result of the check module 450 is affirmative, use the occupied critical resource pair for the first thread according to the priority of the processing queue of the first thread from the highest to the lowest priority. The processing of the first thread waits for the data held by the queue to perform the specified processing.

The detecting module 410 may further include an applying module and a determining module, wherein the applying module is configured to apply a mutual exclusion lock of the first thread to a critical resource required to process the data received by the first thread, and the determining module is configured to apply When the mutex fails, it is determined that the critical resource is already occupied by another thread.

For example, the data received by the first thread may be an SSSAR message that needs to be split into ATM cells, and the specified processing may be AAL2 fragmentation processing, and the critical resources required to process the data received by the first thread may be It is a critical resource related to the VC connection of ATM cells that send SSSAR messages. Alternatively, the data received by the first thread may be an ATM cell that needs to be reassembled into an SSSAR message, the specified process may be AAL2 reassembly processing, and the critical resource required to process the data received by the first thread may be the received SSSAR. The critical resource of the VC connection of the ATM cell of the message.

Those skilled in the art should understand that each module included in the device 400 can be implemented by using software, hardware or a combination of software and hardware.

Embodiments of the present invention also provide a machine readable storage medium storing machine executable instructions, wherein when the machine executable instructions are executed, causing a machine to implement a method for processing data in a first thread, the method comprising the Step: When the first thread receives the data that needs to perform the specified processing, it detects whether the critical resource required to perform the specified processing on the received data has been occupied by another thread; when the detection result is positive, the received The data is sent to the processing queue of the other thread.

The method may further include the following steps: when the detection result is negative, occupying the required critical resource for the first thread; performing the designation on the received data by using the occupied critical resource for the first thread deal with. The method may further include the following steps: after performing the specified processing on the received data, checking whether the data sent by the other thread is saved in the processing waiting queue of the first thread; when the check result is positive, according to the first The processing of the thread waits for the priority of the queue from high to low, and performs the specified processing on the data held by the processing queue of the first thread by using the critical resource occupied by the first thread; when the check result is negative, Release the critical resources used for the first thread.

The transmitting step described above may further include the following steps: when the detection result is positive, identifying the priority of the received data; transmitting the received data to the other thread and the received The priority of the data corresponds to the processing wait queue.

The detecting step described above may further include the steps of: applying a mutex of the first thread to the required critical resource; determining that the required critical resource has been used by the other when the mutex is failed to be applied A thread is occupied.

For example, the received data may be an SSSAR message that needs to be split into ATM cells, and the specified processing may be AAL2 fragmentation processing, and the required critical resource may be an ATM cell that sends an SSSAR message. The virtual path is connected to the related critical resource, or the received data may be an ATM cell that needs to be reassembled into an SSSAR message, and the specified process may be AAL2 reassembly processing, and the required critical resource may be the same as receiving the SSSAR message. The virtual path of the ATM cell is connected to the critical resource.

It should be understood by those skilled in the art that the present invention may be variously modified and changed without departing from the spirit of the invention. Within the scope. Therefore, the scope of the invention is defined by the appended claims.

Claims

Claim

1. A method for processing data in a first thread, comprising the steps of:

When the first thread receives the data that needs to perform the specified processing, it is detected whether the critical resource required to perform the specified processing on the data has been occupied by another thread;

When the result of the check is affirmative, the process of sending the data to the other thread waits for the queue.

2. The method according to claim 1, further comprising the steps of:

When the detection result is negative, occupying the required critical resource for the first thread;

The specified process is performed on the data using the occupied critical resource for the first thread.

3. The method of claim 2, further comprising the steps of:

After performing the specifying process on the data, checking whether the data sent by another thread is saved in the processing waiting queue of the first thread;

When the check result is affirmative, processing the first thread by using the occupied critical resource for the first thread according to a priority of processing priority queues of the first thread waiting for the first thread Waiting for the data saved by the queue to perform the specified processing;

When the check result is negative, the occupied critical resource for the first thread is released.

4. The method according to claim 1, wherein the transmitting step further comprises: identifying a priority of the data when the detection result is affirmative;

The data is sent to a processing wait queue of the other thread that corresponds to the priority of the data.

5. The method of claim 1, wherein the detecting step further comprises: applying a mutex lock of the first thread to the critical resource;

When the mutex is failed to be applied, it is determined that the critical resource has been occupied by another thread.

6. The method according to any one of claims 1 to 5, wherein

The data is an SSSAR packet that needs to be split into ATM cells, and the specified processing is AAL2 fragmentation processing, and the critical resource is related to a virtual path connection of the ATM cell that sends the SSSAR message. Critical resource,

Or the data is an ATM cell that needs to be reassembled into an SSSAR message, where the specified process is an AAL2 reassembly process, and the critical resource is related to a virtual path connection of the ATM cell that receives the SSSAR message. Critical resources.

7. An apparatus for processing data in a first thread, comprising:

a detecting module, configured to: when the first thread receives data that needs to perform specified processing, detect whether a critical resource required to perform the specified processing on the data has been occupied by another thread;

And a sending module, configured to send the data to a processing waiting queue of the another thread when the detection result is positive.

8. The apparatus according to claim 7, further comprising:

An occupation module, configured to occupy the critical resource for the first thread when the detection result is negative;

And an execution module, configured to perform the specified processing on the data by using the occupied critical resource for the first thread.

9. The apparatus according to claim 8, further comprising:

An checking module, configured to check whether data sent by another thread is saved in a processing waiting queue of the first thread after performing the specifying process on the data;

And a release module, configured to release the occupied critical resource for the first thread when the check result is negative.

The execution module is further configured to: when the check result is positive, use the occupied threshold for the first thread according to a priority of processing priority queues of the first thread The resource performs the specified processing on the data held by the processing wait queue of the first thread.

The device of claim 7, wherein the sending module further comprises: an identifying module, configured to: when the detection result is positive, identify a priority of the data; a transmitting module, configured to: Data is sent to the processing queue of the other thread corresponding to the priority of the data.

The device of claim 7, wherein the detecting module further comprises: an applying module, configured to apply a mutual exclusion lock of the first thread to the critical resource; and a determining module, configured to When the mutex fails, it is determined that the critical resource is already occupied by another thread.

12. The apparatus according to any one of claims 7-11, wherein

13. A machine readable storage medium storing machine executable instructions which, when executed, cause a machine to perform the steps of any one of claims 1-6.