CN108259374A - A kind of multi-core processor and message processing method - Google Patents
A kind of multi-core processor and message processing method Download PDFInfo
- Publication number
- CN108259374A CN108259374A CN201710734490.1A CN201710734490A CN108259374A CN 108259374 A CN108259374 A CN 108259374A CN 201710734490 A CN201710734490 A CN 201710734490A CN 108259374 A CN108259374 A CN 108259374A
- Authority
- CN
- China
- Prior art keywords
- core
- message
- work
- shunting
- group
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/12—Avoiding congestion; Recovering from congestion
- H04L47/125—Avoiding congestion; Recovering from congestion by balancing the load, e.g. traffic engineering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F15/00—Digital computers in general; Data processing equipment in general
- G06F15/16—Combinations of two or more digital computers each having at least an arithmetic unit, a program unit and a register, e.g. for a simultaneous processing of several programs
- G06F15/163—Interprocessor communication
- G06F15/173—Interprocessor communication using an interconnection network, e.g. matrix, shuffle, pyramid, star, snowflake
- G06F15/17306—Intercommunication techniques
- G06F15/17318—Parallel communications techniques, e.g. gather, scatter, reduce, roadcast, multicast, all to all
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/24—Traffic characterised by specific attributes, e.g. priority or QoS
- H04L47/2441—Traffic characterised by specific attributes, e.g. priority or QoS relying on flow classification, e.g. using integrated services [IntServ]
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multi Processors (AREA)
Abstract
The present invention provides a kind of multi-core processor and message processing method, the multi-core processor include：PKI modules for identifying Group corresponding with shunting Core for message distribution shunting, Group are identified and shunting mark is filled up to message and corresponds in Work；Shunt Core, for obtaining the Work for including shunting Core and corresponding to Group marks, quantity based on the Work shunting marks included and data Core distributes the first data Core from data Core, and the Group marks that Work includes are replaced with the corresponding Group of the first data Core identifies；First data Core, for obtaining the Work for including the corresponding Group marks of first data Core, reading Work corresponds to message and is handled from memory.The present invention can mitigate the processing pressure of individual data Core.
Description
Technical field
The present invention relates to network communication technology field more particularly to a kind of multi-core processors and message processing method.
Background technology
Multi-core processor refers to be integrated with the processing of multiple Core (being known as process cores or kernel) on a hardware chip
Device.Since multiple process cores in multi-core processor are relatively independent, can parallel processing message, therefore, can be promoted based at multinuclear
Manage the message processing capability of the network equipment of device.
Multi-core processor distributes Group (group) for message first, is then based on Group and Core after message is received
Correspondence, message is handled by the Group distributed corresponding Core.Due to the corresponding Group's of each Core
Quantity is different, causes the processing pressure for corresponding to the Core of more Group can be larger.
Invention content
The purpose of the present invention is to provide a kind of multi-core processor and message processing method, to mitigate the processing pressure of Core
Power.
For achieving the above object, the present invention provides following technical solutions：
The present invention provides a kind of multi-core processor, and the multi-core processor includes：PKI (Packet Input Unit, report
Text receive) module, SSO (Schedule/Synchronize/Order Unit, dispatching message) module, multiple Core；Wherein,
The multiple Core includes：Shunt Core and data Core；
The PKI modules, will for distributing the shunting mark of the shunting corresponding Group of Core and the message for message
The Group marks of the Group of distribution and the shunting are identified and are filled up in the corresponding Work (work item) of the message, and by institute
Message deposit memory is stated, the Work is sent to the SSO modules, wherein, belong to the shunting mark of the message of same data flow
It is sensible same；
The shunting Core, for including the Group of the corresponding Group of the shunting Core from SSO modules acquisition
The shunting mark and the quantity of data Core that includes of the multiple Core that the Work of mark, the Work based on acquisition include from
The first data Core is distributed in the data Core, the Group marks that the Work of acquisition includes are replaced with into first data
The Group marks of the corresponding Group of Core, the SSO modules are sent to by replaced Work；
The first data Core, for including the corresponding Group of the first data Core from SSO modules acquisition
Group marks Work, and message is read from the memory based on the Work of acquisition and is handled.
The present invention also provides a kind of message processing method, the shunting in the multiple Core included applied to multi-core processor
Core, the multi-core processor further include：PKI modules, SSO modules, the multiple Core further include data Core, the method
Including：
It is obtained from the SSO modules and includes the Work that the Group of the corresponding Group of the shunting Core is identified, described point
The Work that stream Core is obtained is the shunting mark that the PKI modules are the message distribution shunting corresponding Group of Core and the message
After knowledge, the Group of the Group of distribution mark and the shunting are identified and are filled up in the corresponding Work of the message, and will be filled out
It is in the block that Work after writing is sent to the SSO moulds, wherein, the shunting mark for belonging to the message of same data flow is identical；
The quantity of data Core that the shunting mark and the multiple Core that Work based on the acquisition includes include
The first data Core is distributed from the data Core；
The Group marks that the Work of the acquisition includes are replaced with the corresponding Group's of the first data Core
Group is identified；
Replaced Work is sent to the SSO modules, so that the first data Core is obtained from the SSO modules
The Work of Group marks including the corresponding Group of the first data Core, and read from memory based on the Work of acquisition
Message is handled, and the message that the first data Core is read from memory is the PKI modules to fill in message corresponding
After Work, message is stored in the memory.
It can be seen from above description the present invention first by PKI modules for the message distribution corresponding Group of shunting Core and
The shunting mark of message, and the Group marks for shunting the corresponding Group of Core is corresponding with the shunting mark write-in message of message
In Work；When shunting Core gets the Work including the corresponding Group of shunting Core, based on the shunting mark in Work
The first data Core, and the Group that Work is included are distributed from data Core with the quantity of data Core in multi-core processor
Mark replaces with the Group marks of the corresponding Group of the first data Core, so that the first data Core is obtained comprising first number
The Work of the Group marks of Group is corresponded to according to Core, and message is read from memory based on the Work of acquisition and is handled.Also
It is to say, the present invention has carried out message secondary distribution, and the first sub-distribution is that the correspondence based on Group and Core divides message
Dispensing shunts Core, and the second sub-distribution is to be distributed to message by shunting quantity of the Core based on data Core in multi-core processor
Data Core (shunting Core is not handled message, only carries out second of batch operation), is based on just because of shunting Core
Second sub-distribution of the quantity of data Core so that perform load balancing based on all data Core and be possibly realized, so as to mitigate
The processing pressure of individual data Core.
Description of the drawings
Fig. 1 is a kind of structure diagram of multi-core processor shown in the embodiment of the present invention；
Fig. 2 is a kind of flow chart of message processing method shown in the embodiment of the present invention.
Specific embodiment
Here exemplary embodiment will be illustrated in detail, example is illustrated in the accompanying drawings.Following description is related to
During attached drawing, unless otherwise indicated, the same numbers in different attached drawings represent the same or similar element.Following exemplary embodiment
Described in embodiment do not represent and the consistent all embodiments of the present invention.On the contrary, they be only with it is such as appended
The example of the consistent device and method of some aspects being described in detail in claims, of the invention.
It is only merely for the purpose of description specific embodiment in terminology used in the present invention, is not intended to limit the invention.
It is also intended in the present invention and " one kind " of singulative used in the attached claims, " described " and "the" including majority
Form, unless context clearly shows that other meanings.It is also understood that term "and/or" used herein refers to and wraps
Containing one or more associated list items purposes, any or all may be combined.
It will be appreciated that though various information, but this may be described using term first, second, third, etc. in the present invention
A little information should not necessarily be limited by these terms.These terms are only used for same type of information being distinguished from each other out.For example, not departing from
In the case of the scope of the invention, the first information can also be referred to as the second information, and similarly, the second information can also be referred to as
One information.Depending on linguistic context, word as used in this " if " can be construed to " ... when " or " when ...
When " or " in response to determining ".
The embodiment of the present invention proposes a kind of multi-core processor, which includes PKI modules, SSO modules and more
A Core, wherein, multiple Core include：Shunt Core and data Core.It is one kind shown in the embodiment of the present invention referring to Fig. 1
The structure diagram of multi-core processor, wherein, Core1 and Core2 are shunting Core, and Core3~Core9 is data Core, when
So, the present invention is not particularly limited the quantity of Core, and the quantity of Core shown in Fig. 1 is merely illustrative.
The modules that multi-core processor includes will be described below：
PKI modules：For distributing the shunting mark of the shunting corresponding Group of Core and the message for message, will distribute
Group Group marks and the shunting mark be filled up in the corresponding Work of the message, and will be in the message is stored in
It deposits, the Work is sent to the SSO modules.
Specifically, the range of the corresponding Group of shunting Core, the corresponding Group of shunting Core are configured in PKI modules
Range and shunting Core configured in SSO modules it is consistent with the correspondence of Group.Referring to table 1, be in SSO modules
The correspondence of the Core and Group of configuration, wherein, each Group uniquely corresponds to a Core.
Table 1
As previously mentioned, Core1 and Core2 is shunting Core, Core3~Core9 is data Core, therefore, is matched in PKI
Put ranging from Group0~Group3 of the corresponding Group of shunting Core.
After PKI receives message, Group is distributed for message in the range of the configured corresponding Group of shunting Core,
For example, can the purpose IP address based on message be the message distribution shunting corresponding Group of Core, as with identical destination IP
The message of location distributes identical Group, and certainly, the present invention is not limited to the side of the corresponding Group of message distribution shunting Core
Formula；In addition, PKI modules are also message distribution shunting mark, wherein, the shunting mark for belonging to the message of same data flow is identical.
The Group of the Group distributed for message mark and shunting mark are filled up in the corresponding Work of message by PKI, and
Work is sent to SSO modules, is saved in by SSO modules in the hardware array of itself, meanwhile, message deposit is located at more by PKI
Memory except core processor is medium pending.
Shunt Core：For including the Group marks of the corresponding Group of the shunting Core from SSO modules acquisition
Work, the shunting mark and the quantity of data Core that includes of the multiple Core that the Work based on acquisition includes are from described
The first data Core is distributed in data Core, the Group marks that the Work of acquisition includes are replaced with the first data Core corresponds to
Group Group mark, replaced Work is sent to the SSO modules.
Specifically, based on the Core being configured in aforementioned SSO modules and the correspondence of Group, shunting Core can be from SSO
The Work for the Group marks for including the corresponding Group of shunting Core is obtained in module, that is to say, that PKI is corresponding by message
In Work (Work includes the Group marks of the corresponding Group of shunting Core) deposit SSO modules, Core is again from SSO moulds for shunting
The Work identified including the Group of the corresponding Group of shunting Core is taken out in block, i.e., it is all into multi-core processor
The corresponding Work of message will pass through shunting Core processing, and processing procedure is as follows：
Core is shunted based on the data Core from the shunting mark and multi-core processor that the Work that SSO modules obtain includes
Quantity, determine to handle the data Core of the corresponding messages of the Work from data Core, be denoted as the first data Core.Specifically
For, in one embodiment, the number that the shunting mark and multi-core processor that shunting Works of the Core based on acquisition includes include
Hash operation is carried out according to the quantity of Core, is determined as the first data Core of message distribution based on operation result, if for example, message
Shunting be identified as 10, the quantity of data Core is 7, it is assumed that utilizes quantity (7) progress of the shunting mark (10) to data Core
The result of Hash operation is 3 (Hash result 0~6 corresponds to Core3~Core9 respectively), then should be current message distribution data
Core6 handles current message by Core6.
Group that the Work obtained from SSO modules includes is identified that replace with the first data Core corresponding by shunting Core
The Group marks of Group, referring to table 1, the corresponding Group of Core6 are Group113~Group117, then may be selected
Group113 replaces the Group marks that Work includes, and the present invention from the corresponding multiple Group of data Core to selecting to replace
The method of the Group marks of the Group changed is not construed as limiting.Replaced Work is sent to SSO modules by shunting Core, is saved in
In the hardware array of SSO modules.
First data Core：For including the corresponding Group's of first data Core from SSO modules acquisition
The Work of Group marks, and message is read from memory based on the Work of acquisition and is handled.
Specifically, equally based on the Core being configured in aforementioned SSO modules and the correspondence of Group, the first data Core
The Work for the Group marks for including the corresponding Group of first data Core can be obtained from SSO modules, for example, Core6
The Work for including the corresponding Group113 of Core6 can be obtained from SSO modules.Message is further included in the Work usually obtained
Storage address, the first data Core read message based on the Work addresses included from memory and are handled.
It can be seen from the above description that the present invention is by shunting shunting mark and the quantity of data Core of the Core based on message
Carry out secondary distribution so that the messages with different shunting marks can be in equilibrium assignment to each data Core, so as to reduce
The processing pressure of individual data Core.
Further, the multi-core processor in the present invention further includes interrupt control unit (not shown in figure 1), interruption control
Device is used to determine that target Core, target Core correspond to the Group that Work present in SSO modules includes from multiple Core
Identify corresponding Group, Work existing for triggering target Core from SSO modules acquisition.That is, the shunting in the present invention
Core and data Core be based on down trigger from SSO modules acquisition include corresponding Group Group identify
Work.The mode of this down trigger avoids Core and periodically inquires in SSO modules when whether there is Work to Core resources
It occupies.
In addition, the shunting Core in the present invention, is additionally operable to the quantity that statistics distributes to the message of the first data Core processing.
First data Core is additionally operable to count the quantity of the message read from memory；And it obtains shunting Core and distributes to the first data
The quantity of the message of Core processing；When the first data Core reads message based on the Work obtained from SSO modules from memory
Afterwards, judge that shunting Core distributes to the quantity of message of oneself processing and the quantity of message read in the slave memory of itself statistics
Difference whether be more than preset amount threshold, if distributing to quantity and the first data of the message of the first data Core processing
The difference of the quantity for the message that Core is read is more than preset amount threshold, illustrates that shunt Core distributes report to the first data Core
The speed of text alreadys exceed the speed of the first data Core processing messages, causes to wait for what first data Core was handled in memory
The quantity of message is increased sharply, and excessively ties up memory source, and therefore, the first data Core will currently be read from memory in the present invention
Message directly abandons, to accelerate releasing memory.
The embodiment of the present invention also proposes a kind of message processing method, is at the message shown in the embodiment of the present invention referring to Fig. 2
The flow chart of reason method, the embodiment are described Message processing process from shunting Core sides.
Step 201, the Group marks of the corresponding Group of the shunting Core are included from SSO modules acquisition
Work, the Work that the shunting Core is obtained are that the PKI modules are the message distribution shunting corresponding Group of Core and the report
After the shunting mark of text, the Group of the Group of distribution mark and the shunting are identified and are filled up to the corresponding Work of the message
In, and to be sent to the SSO moulds in the block by the Work after will fill in, wherein, belong to the shunting mark of the message of same data flow
It is identical.
Step 202, the data Core's that the shunting mark and the multiple Core that the Work based on acquisition includes include
Quantity distributes the first data Core from the data Core.
Step 203, the Group marks that the Work of acquisition includes are replaced with the corresponding Group's of the first data Core
Group is identified.
Step 204, replaced Work is sent to the SSO modules, so that the first data Core is from the SSO
Module acquisition includes the Work that the Group of the corresponding Group of the first data Core is identified, and based on the Work of acquisition from interior
It deposits middle reading message to be handled, the message that the first data Core is read from memory is filling in report for the PKI modules
After the corresponding Work of text, message is stored in the memory.
Further, the multi-core processor further includes interrupt control unit, described to be obtained from the SSO modules including described
The Work of the Group marks of the corresponding Group of Core is shunted, including：
When the trigger signal for receiving the interrupt control unit, include the shunting Core from SSO modules acquisition
The Work of the Group marks of corresponding Group.
Further, the data that the shunting mark and the multiple Core that the Work based on acquisition includes include
The quantity of Core distributes the first data Core from the data Core, including：
The shunting mark and the quantity of data Core that includes of the multiple Core that Work based on the acquisition includes into
Row Hash operation distributes the first data Core based on operation result from the data Core.
Further, the method further includes：
Statistics distributes to the quantity of the message of the first data Core processing so that the first data Core be based on from
After the Work that the SSO is obtained reads message from the memory, obtain the shunting Core and distribute to first data
The quantity of the message of Core processing, if distributing to the quantity of the message of the first data Core processing and first data
When the difference of the quantity of the message read from the memory of Core statistics is more than preset amount threshold, abandon from described interior
Deposit the message of middle reading.
The realization process of above-mentioned message processing method specifically refers to the above-mentioned description to module each in multi-core processor, herein
It repeats no more.
Now still by taking multi-core processor shown in Fig. 1 as an example, the multi-core processor and Message processing process are introduced.
First, as shown in table 1, in SSO modules configured Core and Group correspondence, wherein, shunting Core correspond to
Group ranging from Group0~Group3；The range Group0 of the corresponding Group of configuration shunting Core in PKI modules
~Group3；The amount threshold for distributing to the message that individual data Core is handled stored in default memory is 30.
After PKI modules receive message (being denoted as Packet1), from the range of the corresponding Group of shunting Core of configuration
Group1 is distributed for Packet1 in Group0~Group3, meanwhile, the five-tuple information based on Packet1 is distributed for Packet1
Shunting mark (being denoted as Hash-tag1) is 10, in addition, the storage address (notes of Packet1 in memory are also predefined in PKI modules
For Address1), by the corresponding Work (being denoted as Work1) of Group1, Hash-tag1 and Address1 write-in Packet1
In, and Work1 is sent to SSO modules, Packet1 is stored in memory (not shown in figure 1).
Since Group1 corresponds to Core1, interrupt control unit (not shown in figure 1) by down trigger Core1 from
The Work1 for including the corresponding Group1 of Core1 is obtained in SSO modules, Core1 obtains the Hash- of Packet1 from Work1
The value (10) of tag1, the value (10) based on the Hash-tag1 to the quantity of data Core in multi-core processor (Core3~
Core9, totally 7 data Core) carry out Hash, it is assumed that and Hash result 3 distributes data according to the Hash result for Packet1
Core6 distributes Core6 processing Packet1.Core6 corresponds to Group113~Group117 as known from Table 1, from this
A Group, such as Group113 are selected in the range of Group113~Group117, is replaced in Work1 using the Group113
Group1, replaced Work (being denoted as Work2) is sent to SSO modules, meanwhile, statistics distribute to Core6 processing report
The quantity (being denoted as N1) of text.
Similarly, since Group113 corresponds to Core6, interrupt control unit is by down trigger Core6 from SSO modules
Middle to obtain the Work2 for including the corresponding Group113 of Core6, Core6 obtains the storages of Packet1 in memory from Work2
Address Address1 reads Packet1 based on Address1 from memory, meanwhile, count the message that itself has currently read
Quantity (is denoted as M1), and Core6 obtains the quantity for distributing to the message of itself that all shunting Core (Core1 and Core2) count,
It is assumed that Core2 distributes to the quantity of the message of Core6, it is denoted as N2, it is assumed here that, N1=50, N2=60, M1=90, then
Core6 calculates (N1+N2)-M1=20<Preset amount threshold (30) illustrates that is stored in memory waits for the report of Core6 processing
The quantity of text has not exceeded the processing capacity of Core6, and therefore, Core6 can normally be handled the Packet1 currently read.
It is assumed that PKI modules receive new message (being denoted as Packet2) again, from the corresponding Group of shunting Core of configuration
Range Group0~Group3 in for Packet2 distribute Group0, meanwhile, the five-tuple information based on Packet2 is
Packet2 distribution shunting marks (being denoted as Hash-tag2) are 7, PKI modules by Group0, Hash-tag2 and Packet2
In the corresponding Work (being denoted as Work3) of storage address (being denoted as Address2) write-in Packet2, and Work3 is sent to SSO moulds
Packet2 is stored in memory by block.
Since Group0 corresponds to Core1, interrupt control unit obtains packet by down trigger Core1 from SSO modules
The Work3 of the corresponding Group0 of Core1 is included, Core1 obtains the value (7) of the Hash-tag2 of Packet2 from Work3, is based on
The value (7) of Hash-tag2 carries out Hash to the quantity (7) of data Core in multi-core processor, it is assumed that Hash result 0, according to
The Hash result distributes data Core3 for Packet2, that is, distributes Core3 processing Packet2.Core3 is corresponded to as known from Table 1
Group101~Group104 selects Group102 in the range of the Group101~Group104, is replaced using the Group102
The Group0 in Work3 is changed, replaced Work (being denoted as Work4) is sent to SSO modules, meanwhile, statistics distributes to Core3
The quantity (being denoted as N3) of the message of processing.
Similarly, since Group102 corresponds to Core3, interrupt control unit is by down trigger Core3 from SSO modules
Middle to obtain the Work4 for including the corresponding Group102 of Core3, Core3 obtains the storages of Packet2 in memory from Work4
Address Address2 reads Packet2 based on Address2 from memory, meanwhile, count the message that itself has currently read
Quantity (is denoted as M2), and Core3 obtains the quantity for distributing to the message of itself of Core1 and Core2 statistics, it is assumed that Core2 is distributed
To the quantity of the message of Core3, be denoted as N4, it is assumed here that, N3=70, N4=80, M2=100, then Core3 calculate (N3+N4)-
M2=50>Preset amount threshold (30), the message for waiting for Core3 processing for illustrating to store in memory is excessive, more than Core3
Processing capacity, and tie up the memory space of other data Core messages to be processed, therefore, Core3 is to currently reading
Packet2 is directly abandoned, with quick release memory headroom.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention
God and any modification, equivalent substitution, improvement and etc. within principle, done, should be included within the scope of protection of the invention.
Claims (8)
1. a kind of multi-core processor, which is characterized in that the multi-core processor includes：Message receives PKI modules, dispatching message
SSO modules, multiple core Core；Wherein, the multiple Core includes：Shunt Core and data Core；
The PKI modules for distributing the shunting mark of shunting corresponding group of Group of Core and the message for message, will divide
The Group marks of the Group matched and the shunting are identified and are filled up in the corresponding work item Work of the message, and by the report
Text deposit memory, the SSO modules are sent to by the Work, wherein, the shunting for belonging to the message of same data flow identifies phase
Together；
The shunting Core, for including the Group marks of the corresponding Group of the shunting Core from SSO modules acquisition
Work, the shunting mark and the quantity of data Core that includes of the multiple Core that the Work based on acquisition includes are from described
The first data Core is distributed in data Core, the Group marks that the Work of acquisition includes are replaced with into the first data Core
The Group marks of corresponding Group, the SSO modules are sent to by replaced Work；
The first data Core, for including the corresponding Group's of the first data Core from SSO modules acquisition
The Work of Group marks, and message is read from the memory based on the Work of acquisition and is handled.
2. processor as described in claim 1, which is characterized in that the processor further includes：
Interrupt control unit, for determining that target Core, the target Core correspond to the SSO modules from the multiple Core
Present in the Group that include of Work identify corresponding Group；It triggers described in the target Core from SSO modules acquisition
Existing Work.
3. processor as described in claim 1, which is characterized in that the shunting that the shunting Works of the Core based on acquisition includes
The quantity of data Core that mark and the multiple Core include distributes the first data Core from the data Core, wraps
It includes：
The number of data Core that the shunting mark and the multiple Core that the shunting Works of the Core based on acquisition includes include
Amount carries out Hash operation, and the first data Core is distributed from the data Core based on operation result.
4. processor as described in claim 1, it is characterised in that：
The shunting Core is additionally operable to the quantity that statistics distributes to the message of the first data Core processing；
The first data Core is additionally operable to, and counts the quantity of the message read from the memory；Obtain the shunting Core
Distribute to the quantity of the message of the first data Core processing；
The first data Core reads message from the memory based on the Work of acquisition and is handled, including：
The first data Core reads message based on the Work of acquisition from the memory, if being currently assigned to first number
Difference according to the quantity and the quantity of the first data Core messages read of the Core messages handled is more than preset quantity
Threshold value then abandons the message read from the memory.
5. a kind of message processing method, the shunting Core in the multiple core Core included applied to multi-core processor, feature exists
In the multi-core processor further includes：Message receives PKI modules, dispatching message SSO modules, and the multiple Core further includes number
According to Core, the method includes：
Include the work item Work of the Group marks of corresponding group of Group of the shunting Core, institute from SSO modules acquisition
It is point that the PKI modules are the message distribution shunting corresponding Group of Core and the message to state the Work that shunting Core is obtained
After traffic identifier, the Group of the Group of distribution mark and the shunting are identified and are filled up in the corresponding Work of the message, and
It is in the block that Work after will fill in is sent to the SSO moulds, wherein, the shunting mark for belonging to the message of same data flow is identical；
The shunting mark and the quantity of data Core that includes of the multiple Core that Work based on the acquisition includes are from institute
It states and the first data Core is distributed in data Core；
The Group that the Group marks that the Work of the acquisition includes are replaced with to the corresponding Group of the first data Core is marked
Know；
Replaced Work is sent to the SSO modules, so that the first data Core includes from SSO modules acquisition
The Work of the Group marks of the corresponding Group of the first data Core, and the Work based on acquisition reads message from memory
It is handled, the message that the first data Core is read from memory is filling in the corresponding Work of message for the PKI modules
Afterwards, message is stored in the memory.
6. method as claimed in claim 5, which is characterized in that the multi-core processor further includes interrupt control unit, it is described from
The SSO modules acquisition includes the Work of the Group marks of the corresponding Group of the shunting Core, including：
When the trigger signal for receiving the interrupt control unit, include the shunting Core from SSO modules acquisition and correspond to
Group Group mark Work.
7. method as claimed in claim 5, which is characterized in that shunting that the Work based on the acquisition includes identify with
And the quantity of data Core that the multiple Core includes distributes the first data Core from the data Core, including：
The quantity of data Core that the shunting mark and the multiple Core that Work based on the acquisition includes include is breathed out
Uncommon operation, the first data Core is distributed based on operation result from the data Core.
8. method as claimed in claim 5, which is characterized in that the method further includes：
Statistics distributes to the quantity of the message of the first data Core processing, so that the first data Core is based on from described
After the Work that SSO is obtained reads message from the memory, obtain the shunting Core and distribute at the first data Core
The quantity of the message of reason, if the quantity for distributing to the message of the first data Core processing is counted with the first data Core
The message read from the memory quantity difference be more than preset amount threshold when, abandon and read from the memory
Message.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710734490.1A CN108259374B (en) | 2017-08-24 | 2017-08-24 | A kind of multi-core processor and message processing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710734490.1A CN108259374B (en) | 2017-08-24 | 2017-08-24 | A kind of multi-core processor and message processing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108259374A true CN108259374A (en) | 2018-07-06 |
| CN108259374B CN108259374B (en) | 2019-07-09 |
Family
ID=62721256
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710734490.1A Active CN108259374B (en) | 2017-08-24 | 2017-08-24 | A kind of multi-core processor and message processing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108259374B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109067667A (en) * | 2018-07-13 | 2018-12-21 | 哈尔滨工程大学 | A kind of dynamic load balancing method based on message classification |
| WO2021073473A1 (en) * | 2019-10-16 | 2021-04-22 | 中兴通讯股份有限公司 | Data packet processing method and apparatus, communication device, and storage medium |
| CN113176940A (en) * | 2021-03-29 | 2021-07-27 | 新华三信息安全技术有限公司 | Data flow splitting method and device and network equipment |
| CN115022423A (en) * | 2022-06-22 | 2022-09-06 | 上海弘积信息科技有限公司 | IPv4 identifier distribution method in load balancing equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110145461A1 (en) * | 2008-08-19 | 2011-06-16 | Zte Corporation | Method and device for balancing interrupt load of multicore processor |
| CN102521047A (en) * | 2011-11-15 | 2012-06-27 | 重庆邮电大学 | Method for realizing interrupted load balance among multi-core processors |
| CN103514043A (en) * | 2012-06-29 | 2014-01-15 | 华为技术有限公司 | Multi-processor system and data processing method thereof |
| CN103744735A (en) * | 2014-01-09 | 2014-04-23 | 福建星网锐捷网络有限公司 | Method and device for scheduling multi-core resource |
-
2017
- 2017-08-24 CN CN201710734490.1A patent/CN108259374B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110145461A1 (en) * | 2008-08-19 | 2011-06-16 | Zte Corporation | Method and device for balancing interrupt load of multicore processor |
| CN102521047A (en) * | 2011-11-15 | 2012-06-27 | 重庆邮电大学 | Method for realizing interrupted load balance among multi-core processors |
| CN103514043A (en) * | 2012-06-29 | 2014-01-15 | 华为技术有限公司 | Multi-processor system and data processing method thereof |
| CN103744735A (en) * | 2014-01-09 | 2014-04-23 | 福建星网锐捷网络有限公司 | Method and device for scheduling multi-core resource |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109067667A (en) * | 2018-07-13 | 2018-12-21 | 哈尔滨工程大学 | A kind of dynamic load balancing method based on message classification |
| CN109067667B (en) * | 2018-07-13 | 2022-03-18 | 哈尔滨工程大学 | Dynamic load balancing method based on message classification |
| WO2021073473A1 (en) * | 2019-10-16 | 2021-04-22 | 中兴通讯股份有限公司 | Data packet processing method and apparatus, communication device, and storage medium |
| CN113176940A (en) * | 2021-03-29 | 2021-07-27 | 新华三信息安全技术有限公司 | Data flow splitting method and device and network equipment |
| CN115022423A (en) * | 2022-06-22 | 2022-09-06 | 上海弘积信息科技有限公司 | IPv4 identifier distribution method in load balancing equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108259374B (en) | 2019-07-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108259374B (en) | A kind of multi-core processor and message processing method | |
| US10313362B2 (en) | Systems and methods for real-time configurable load determination | |
| US20190273656A1 (en) | Consistent hashing for network traffic dispatching | |
| CN104283806B (en) | Business chain processing method and equipment | |
| US9225668B2 (en) | Priority driven channel allocation for packet transferring | |
| CN102457571B (en) | Method for uniformly distributing data in cloud storage | |
| CN104067578B (en) | Dynamic load leveling without packet reordering | |
| CN107547391B (en) | Message transmission method and device | |
| CN108259346B (en) | Equivalent routing table item establishing method and device | |
| US9525727B2 (en) | Efficient and scalable pull-based load distribution | |
| CN102047619B (en) | Methods, systems, and computer readable media for dynamically rate limiting slowpath processing of exception packets | |
| CN108055203A (en) | A kind of equivalent route load sharing method and device | |
| EP2730067A1 (en) | Method and apparatus for load balancing | |
| CN104683255A (en) | Load sharing balance method and device for physical ports, and link aggregation system | |
| US8305889B2 (en) | Method for allocating a resource among consumers in proportion to configurable weights | |
| CN112948313B (en) | Session data management method and device | |
| CN104461727A (en) | Memory module access method and device | |
| US9063841B1 (en) | External memory management in a network device | |
| CN108170533A (en) | The processing method and processing device of message, computer readable storage medium | |
| CN106941457B (en) | The method and device of IP packet frament reassembling based on multi-core processor | |
| CN104580008A (en) | Method and device for improving multi-queue random message discarding accuracy based on hardware | |
| CN101848158B (en) | Load balance method and device for data channels and network exchange equipment | |
| US7533109B2 (en) | Item queue management | |
| CN108595370A (en) | A kind of multi-core processor and message processing method | |
| WO2003103234A1 (en) | Packet transfer circuit and packet transfer method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |