CN103986585B - Message preprocess method and its device - Google Patents
Message preprocess method and its device Download PDFInfo
- Publication number
- CN103986585B CN103986585B CN201410200835.1A CN201410200835A CN103986585B CN 103986585 B CN103986585 B CN 103986585B CN 201410200835 A CN201410200835 A CN 201410200835A CN 103986585 B CN103986585 B CN 103986585B
- Authority
- CN
- China
- Prior art keywords
- message
- dma
- network
- accelerator engine
- network interface
- 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.)
- Active
Links
Abstract
The present invention provides a kind of message preprocess method and device, and applied to the business board for being configured with network accelerator engine, this method includes：The first DMA channel corresponding to PCI E ports is established, the message of the forwarding chip from the business board is write by internal memory by first DMA channel；By the DMA driving generation notification messages corresponding to first DMA channel, and the notification message is sent to the network accelerator engine, wherein including the attribute information of the message in the notification message；The network accelerator engine pre-processes according to the notification message to the message.In technical scheme, the message from forwarding chip is pre-processed by network accelerator engine, so as to avoid taking cpu resource, helps to lift the overall performance of the network equipment.
Description
Technical field
The present invention relates to technical field of data processing, more particularly to a kind of message processing method and a kind of Message processing dress
Put.
Background technology
For the existing network equipment, lead to as shown in figure 1, common hardware design is PCI-E ports as local cpu
Road.However, because PCIE ports are more single by function during DMA packet receivings, it is merely able to realize the function by queue packet receiving, nothing
Method realizes load balancing by pretreatments such as the parsing to message, classification, distribution.
In the related art, it is proposed that message is pre-processed by software mode, but can be caused to the big of cpu resource
Amount takes, and causes that the overall performance of the network equipment is extremely low, and availability is poor.
Therefore, how to realize that the message being passed to PCI-E ports is effectively pre-processed, and avoid largely taking CPU
Resource, turn into technical problem urgently to be resolved hurrily at present.
The content of the invention
In view of this, the present invention provides a kind of message processing method and a kind of message process device, by PIC-E ports
The bridge joint of corresponding DMA channel and network accelerator engine, the message that PIC-E ports are passed to is carried out in advance by network accelerator
Reason, avoids the occupancy to cpu resource, contributes to the network equipment to maintain higher overall performance.
To achieve the above object, it is as follows to provide technical scheme by the present invention：
A kind of message preprocess method, applied to business board, the business board is configured with network accelerator engine, this method
Including：
The first DMA channel corresponding to PCI-E ports is established, by first DMA channel by from the business board
The message write-in internal memory of forwarding chip；
By the DMA driving generation notification messages corresponding to first DMA channel, and the notification message is sent to institute
Network accelerator engine is stated, wherein including the attribute information of the message in the notification message；
The network accelerator engine pre-processes according to the notification message to the message.
Correspondingly, the invention also provides a kind of message pretreatment unit, applied to business board, the business board is configured with
Network accelerator engine, the device include：
Unit is established, for establishing the first DMA channel corresponding to PCI-E ports；
Writing unit, for by first DMA channel by the write-in of the message of the forwarding chip from the business board
Deposit；
Generation unit, for generating notification message by the DMA drivings corresponding to first DMA channel, wherein described logical
Know the attribute information that the message is included in message；
Transmitting element, for the notification message to be sent to the network accelerator engine；
Pretreatment unit, for, according to the notification message, being carried out by the network accelerator engine to the message pre-
Processing.
From above technical scheme, the present invention is by DMA channel corresponding to PIC-E ports and network accelerator engine
Bridge joint, by network accelerator to PIC-E ports be passed to message pre-process, avoid the occupancy to cpu resource, help
Higher overall performance is maintained in the network equipment.
Brief description of the drawings
Fig. 1 is the schematic diagram that the message being passed in correlation technique to PCI-E ports is handled；
Fig. 2 shows the schematic flow diagram of message preprocess method according to an embodiment of the invention；
Fig. 3 is the schematic diagram that the message being passed in correlation technique to network interface is handled；
Fig. 4 show it is according to an embodiment of the invention by network accelerator engine to PCI-E ports be passed to
The schematic diagram that message is pre-processed；
Fig. 5 A-5C show the schematic diagram that the message according to an embodiment of the invention in internal memory is retransmitted；
Fig. 6 shows being passed to by network accelerator engine to PCI-E ports according to another embodiment of the invention
The schematic diagram that is pre-processed of message；
Fig. 7 shows the schematic block diagram of message pretreatment unit according to an embodiment of the invention.
Embodiment
The present invention is by the bridge joint to DMA channel corresponding to PIC-E ports and network accelerator engine, by network accelerator
The message being passed to PIC-E ports pre-processes, and avoids the occupancy to cpu resource, contributes to the network equipment to remain higher
Overall performance.
For the present invention is further described, there is provided the following example：
Fig. 2 shows the schematic flow diagram of message preprocess method according to an embodiment of the invention.
As shown in Fig. 2 message preprocess method according to an embodiment of the invention, can apply to be configured with network
The business board of accelerator engine, this method include：
Step 202, the first DMA channel corresponding to PCI-E ports is established, will be come from by first DMA channel described
The message write-in internal memory of the forwarding chip of business board；
Step 204, by the DMA driving generation notification messages corresponding to first DMA channel, and by the notification message
Send to the network accelerator engine, wherein including the attribute information of the message in the notification message；
Step 206, the network accelerator engine pre-processes according to the notification message to the message.
In the above-described embodiments, by the utilization to network accelerator engine, the message from forwarding chip can be entered
Row pretreatment, so as to avoid taking cpu resource.In fact, NA (Network are often configured with the network equipment
Accelerator, network accelerator) engine, compared to the message pretreatment realized using software, NA engines need not take CPU
Resource, and with abundant Message processing function.Therefore, in technical scheme, actually need to utilize NA engines
Message processing function, realize and the message from PCI-E ports is pre-processed.
For example Fig. 3 shows the process that NA engines are normally handled message.As shown in Figure 3, it is assumed that NA engines are corresponding
Network interface be XGMII (10Gigabit Medium Independent Interface, 10 Gigabit Media independent interfaces).Report
The message transmissions queue that text is entered in network interface, i.e. GMAC FIFO (gigabit (Gigabit) media access control (Media
Access Control) FIFO (First Input First Output)) queue, the message in the queue drawn by NA
The network interface DMA channel for holding up foundation is passed in internal memory.
Meanwhile NA engines also include packet parsing engine, for being parsed (parse), classification to message
(classify) pretreatment such as (distribution), is distributed.Based on the pretreatment of packet parsing engine, corresponding message is obtained
Description information, NA engines can enter according to the message description information to its respective thread (Thread) in corresponding core (Core)
Row calls, then the thread reads corresponding message and handled from internal memory.Therefore, based on pretreatment behaviour of the NA engines to message
Make, be easy to implement load balancing.
It can be seen that in order to realize that NA engines pre-process the message from forwarding chip, actually need to realize PCI-E
Associating between port and NA engines, the then specific embodiment for proposing two kinds of technical schemes of the present invention, below in conjunction with accompanying drawing
It is described in detail.
Embodiment one
, can be by net corresponding to PCI-E DMA channels corresponding to PCI-E ports and NA engines as an exemplary embodiment
Mouth DMA channel is bridged, so as to which the message from forwarding chip is passed into NA engines, and by NA engine implementations to the pre- of message
Processing.
Specifically, Fig. 4 show it is according to an embodiment of the invention by network accelerator engine to PCI-E ports
The schematic diagram that incoming message is pre-processed.
As shown in figure 4, it is according to an embodiment of the invention by network accelerator engine to PCI-E ports be passed to
The process that message is pre-processed includes：
(1) chip configuration when initializing.Specifically, including：DMA is driven to corresponding to PCI-E DMA configurations in physics
Address is deposited, and network port driving is physical memory addresses corresponding to network interface DMA configurations, is led to be respectively used to storage from corresponding DMA
The message in road.
(2) forwarding chip is passed to message by PCI-E ports.
(3) the operation A shown in Fig. 4 is corresponded to, message is passed to internal memory by PCI-E DMA channels.Wherein, PCI-E DMA lead to
Road includes multiple message queues, with corresponding to different business.
Therefore, it is determined that needing to perform the message queue of pretreatment, and be able to will be come from according to the actual demand of current application
The message of the message queue performs following step (4) to the handling process of (7), is otherwise handled according to prior art.
(4) assume that outstanding message is message X, then DMA drives recorded message X attribute information.
(5) message X attribute information can be informed that NA draws based on the change driven to DMA, then DMA of the invention driving
Hold up, it is started network interface DMA transmission operation.Message is sent, include message X's in the message
Attribute information.
(6) correspond to the operation B shown in Fig. 4, the attribute information by network interface DMA according to message X, the report is read from internal memory
Literary X, and message X is retransmitted in internal memory, i.e. message X is transmitted to right by from the physical memory addresses corresponding to PCI-E DMA
Should be in network interface DMA physical memory addresses.
(7) because message copy is in physical memory addresses corresponding to network interface DMA, thus the message copy can directly lead to
Cross network interface DMA and enter NA engines, and and then enter packet parsing engine, to realize that message pre-processes.Specifically, for example pre-process
During " distribution " can be：Hash is carried out to message X by the packet parsing module in NA engines (actual is message copy)
(HASH) value calculates, so as to be distributed to corresponding VCPU (virtual processor), to realize load balancing.
In this embodiment, after message X is stored in internal memory, by being informed to the message of NA engines so that message X is (actual
On be corresponding message copy) be stored to physical memory addresses corresponding to network interface DMA, so as to enter NA engines carry out it is pre-
Processing, compared to software processing mode, above-described embodiment avoids a large amount of occupancy to cpu resource.
It should be noted that：, may be according to actual conditions not for the attribute information of the message X of record in step (4)
Change together, be described in detail with reference to Fig. 5 A-5C.
Scene one
As shown in Figure 5A, it is assumed that message X will not carry out any place when being transmitted via PCI-E DMA to message X
Reason, then attribute information is storage address PktAddresss of the message X in internal memory, and message X length length.
| Type | Source address | Destination address | Length |
| 1 | PktAddress | D-PktAddress | length |
Table 1
As an exemplary embodiment, it is assumed that the form of the message is as shown in table 1, wherein containing " type ", " source
Location ", " destination address ", " length " etc..For example when " type " is 1, show that current message is used for：Indicate recipient according to current
" source address ", " destination address ", " length " in message etc., read corresponding message X, and be retransmitted to destination address from internal memory
D-PktAddress；And " source address ", " destination address ", " length " etc., then it is used to indicate storage positions of the message X in internal memory
Put.
So, in step (6), storage address PktAddress can be jumped to, and read according to length length
Data, you can obtain message X.
Scene two
The message X sent on forwarding chip may be added a head, be then transmitted by PCI-E DMA.Therefore, belong to
Property length length of the information except including message X storage address PktAddress in internal memory and message X, also comprising addition
Head length offset.
As an exemplary embodiment, as shown in Figure 5 B, length is message X length, and with the head length of addition
Offset values are unrelated, and storage address PktAddress is then influenceed by offset.
| Type | Source address | Destination address | Length |
| 1 | PktAddress+offset | D-PktAddress | length |
Table 2
| Type | Source address | Offset | Destination address | Length |
| 1 | PktAddress | offset | D-PktAddress | length |
Table 3
Correspondingly, table 2 shows the message format as an exemplary embodiment, and " source address " therein is by table 1
" PktAddress " is changed to " PktAddress+offset ", then message X attribute information actually includes storage address
PktAddress+offset (offset offset is included in " storage address ") and message X length length.Or
As the message format of another exemplary embodiment, " offset " is used as single information in table 3, so as to message X attribute
Information actually includes storage address PktAddress (actually shoulding be PktAddress+offset), offset offset
With message X length length.
So, in step (6), storage address (PktAddress+offset) can be jumped to, and according to length
Length reads data, you can obtains message X.
Scene three
As another exemplary embodiment, as shown in Figure 5 C, length is that message X is added the total length behind head, i.e.,
Offset values are contained in length length.
| Type | Source address | Destination address | Length |
| 1 | PktAddress+offset | D-PktAddress | length-offset |
Table 4
Correspondingly, table 4 shows the message format as an exemplary embodiment, and " source address " therein is by table 1
" PktAddress " is changed to " PktAddress+offset ", and " length " is changed to " length- by " length " in table 1
Offset ", then message X attribute information actually include storage address PktAddress+offset and message X length
Offset offset (is included in " storage address " and " length ") by length-offset.It is or real as another exemplary
The message format of example is applied, it is same that " offset " (is shown in Table 3, but actual message X length is different from as single information
Scene two), actually (actually it is should be comprising storage address PktAddress so as to message X attribute information
PktAddress+offset), offset offset and message X length length (actually should be length+
offset)。
So, in step (6), storage address (PktAddress+offset) can be jumped to, and according to length
(length-offset) data are read, you can obtain message X.
Further, since in embodiment one, NA engine final process is actually message copy corresponding to message X, is made
It is in itself and not processed to obtain message X.Therefore, after message copy generates or is pre-processed into NA engines, will can report
Memory address corresponding to literary X is labeled as that can cover, to avoid wasting corresponding memory address.
Embodiment two
As another exemplary embodiment, after message is passed into internal memory by PCI-E DMA channels, can be added
To the queue of network interface corresponding to NA engines, than GMAC fifo queues as shown in Figure 3, so as to via the network interface by message weight
After new transmission once, by pretreatment of the NA engine implementations to message.
Specifically, Fig. 6 show according to another embodiment of the invention by network accelerator engine to PCI-E ends
The schematic diagram that the message that oral instructions enter is pre-processed.
As shown in fig. 6, being passed to by network accelerator engine to PCI-E ports according to another embodiment of the invention
The process that is pre-processed of message include：
(1) chip configuration when initializing.Specifically, including：1) DMA is driven to physics corresponding to PCI-E DMA configurations
Memory address, and network port driving is physical memory addresses corresponding to network interface DMA configurations, and corresponding DMA is come to be respectively used to storage
The message of passage；2) network interface (the XGMII ports i.e. shown in Fig. 6) is configured to loopback, can be data link layer loopback or
Physical layer loopback.Such as when configuring physical layer loopback, can be by configuring PHY (Physical Layer Device, physics
Layer device) register (such as control registers) of chip realizes.
(2) forwarding chip is passed to message by PCI-E ports.
(3) the operation A ' shown in Fig. 6 is corresponded to, message is passed to internal memory by PCI-E DMA channels.Wherein, PCI-E DMA lead to
Road includes multiple message queues, with corresponding to different business.
Therefore, it is determined that needing to perform the message queue of pretreatment, and be able to will be come from according to the actual demand of current application
The message of the message queue performs following step (4) to the handling process of (8), is otherwise handled according to prior art.
(4) assume that outstanding message is message X ', then DMA drives recorded message X ' attribute information.Wherein, message X '
Attribute information it is identical with the message X in embodiment one, can be with specific reference to the scene shown in Fig. 5 A-5C.
(5) message X ' attribute information can be informed by NA based on the change driven to DMA, then DMA of the invention driving
Engine, and call network port driving.
| Type | Source address | Length |
| 0 | PktAddress | length |
Table 5
Message X ' attribute information is similar to the attribute information of the message X in embodiment one, corresponding to shown in Fig. 5 A
Under scene, attribute information can include storage address PktAddresss and message X ' of the message X ' in internal memory length
Length, and NA engines can be informed by the message of form shown in table 5 and call network port driving.Wherein, " the class of the message
Type " can be " 0 ", then show that current message is used for：Recipient is indicated according to " source address " in current message, " length " etc.,
Corresponding message X ' is read from internal memory, and calls network port driving.
| Type | Source address | Length |
| 0 | PktAddress+offset | length |
Table 6
| Type | Source address | Offset | Length |
| 0 | PktAddress | offset | length |
Table 7
Similarly, it is necessary to consider offset offset to message X ' in internal memory under corresponding to the scene shown in Fig. 5 B
Storage address influence.Accordingly, as an exemplary embodiment, the message format shown in table 6 can be used so that " source
Address " is changed to " PktAddress+offset " by " PktAddress " in table 5, then message X ' attribute information is actually
Include storage address PktAddress+offset (offset offset is included in " storage address ") and message X ' length
length.Or as another exemplary embodiment, the message format shown in table 7 can be used, by " offset " as independent
Information, so as to which message X ' attribute information actually (is actually should be comprising storage address PktAddress
PktAddress+offset), offset offset and message X ' length length.
| Type | Source address | Length |
| 0 | PktAddress+offset | length-offset |
Table 8
Similarly, it is necessary to consider offset offset to message X ' in internal memory under corresponding to the scene shown in Fig. 5 C
Storage address and length influence.Accordingly, as an exemplary embodiment, the message format shown in table 8 can be used, is made
" source address " is changed to " PktAddress+offset " by " PktAddress " in table 5, and " length " by table 5
" length " is changed to " length-offset ", then message X ' attribute information actually includes storage address PktAddress+
Offset and message X ' length length-offset (offset offset is included in " storage address " and " length ").
Or the message format as another exemplary embodiment, it is same that " offset " (is shown in Table 7, but reality as single information
On message X ' length it is different), so as to message X ' attribute information actually comprising storage address PktAddress (actually
Should be PktAddress+offset), offset offset and message X ' length length (actually should be length+
offset)。
(6) correspond to the operation B ' shown in Fig. 6, the attribute information by network port driving according to message X ', read from internal memory
Message X ', and the message queue (GMAC fifo queues as shown in Figure 3) by message X ' added to network interface, so as to realize
Transmission to message X '.
(7) data link layers or physical layer loopback of the message X ' in network interface.
(8) the operation C ' shown in Fig. 6 is corresponded to, message X ' returns to network interface and enters NA engines, by the message in NA engines
Parsing module (as shown in Figure 3) is pre-processed.
Compared with the prior art shown in Fig. 3：Message X ' adds GMAC fifo queues after being taken out from internal memory,
So as to perform the processing step similar to prior art, only because the network interface in embodiment illustrated in fig. 6 is configured as loopback,
So that message X ' comes back to network interface, and NA engines can be entered.
And message X ' be stored in internal memory after, by the calling to network port driving and by network interface be arranged to data link layer or
Physical layer loopback so that message X ' can be pre-processed by network interface into NA engines, above-mentioned compared to software processing mode
Embodiment avoids a large amount of occupancy to cpu resource.
Fig. 7 shows the schematic block diagram of message pretreatment unit according to an embodiment of the invention.
As shown in fig. 7, message pretreatment unit according to an embodiment of the invention can be understood as on the network equipment
One logic device of operation.In hardware view, the network equipment includes processor, internal memory and nonvolatile memory, when
The hardware being so also possible that required for other business.The processor calculates corresponding to being read from nonvolatile memory
Machine program is into internal memory and then runs, and the user management device is formed on logic level, and the message pretreatment unit can be with
It is interpreted as the part that application system externally services, such as a part of network interaction system external service.Certainly, except
Outside software realization mode, the application is not precluded from other implementations, such as the mode of logical device or software and hardware combining
Etc., that is to say, that the executive agent of the handling process of following management method is not limited to each logic unit, management method
Executive agent can also be hardware or logical device.
In Software Implementation, the message pretreatment unit can include establish unit, writing unit, generation unit,
Transmitting element, pretreatment unit, judging unit, recording unit and indexing unit etc., correspond to above-described embodiment for realizing
Message preprocess method, here is omitted.
Wherein, as an exemplary embodiment, pretreatment unit may further include the first calling subelement, generation
Unit, storage subelement and the first transmission subelement；As another exemplary embodiment, pretreatment unit can also be wrapped further
Include the second calling subelement, addition subelement and the second transmission subelement.
Therefore, the present invention is by the bridge joint to DMA channel corresponding to PIC-E ports and network accelerator engine, by network
The message that accelerator is passed to PIC-E ports pre-processes, and avoids the occupancy to cpu resource, contributes to the network equipment to tie up
Hold higher overall performance.
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 any modification, equivalent substitution and improvements done etc., should be included within the scope of protection of the invention with principle.
Claims (12)
1. a kind of message preprocess method, applied to business board, it is characterised in that the business board is configured with network accelerator and drawn
Hold up, this method includes：
The first DMA channel corresponding to PCI-E ports is established, by first DMA channel by the forwarding from the business board
The message write-in internal memory of chip；
By the DMA driving generation notification messages corresponding to first DMA channel, and the notification message is sent to the net
Network accelerator engine, wherein including the attribute information of the message in the notification message；Wherein, the attribute information of the message
Including：The length of storage address of the message in internal memory and the message；
The network accelerator engine pre-processes according to the notification message to the message.
2. message preprocess method according to claim 1, it is characterised in that also include：
The DMA drivings judge the source of the message of write-in internal memory；
Wherein, if the message is from the message queue corresponding to first DMA channel, and the message queue is corresponding to default
Business, then the DMA drivings record the attribute information of the message, to generate the notification message.
3. message preprocess method according to claim 1, it is characterised in that if the message when writing internal memory, quilt
The head of preset length is added, then the attribute information of the message also includes：Corresponding to the address offset amount of the preset length.
4. message preprocess method according to any one of claim 1 to 3, it is characterised in that the network accelerator
Engine carries out pretreatment to the message and specifically included according to the notification message：
Call the network interface DMA corresponding to the network accelerator engine；
Attribute informations of the network interface DMA in the notification message, the message is obtained from internal memory and is generated corresponding
Message copy；
The network interface DMA deposits in the message copy corresponding network interface DMA physical memory addresses；And
The message copy is passed to the network accelerator engine by the network interface DMA, and by the network accelerator engine pair
The message copy is pre-processed.
5. message preprocess method according to claim 4, it is characterised in that also include：
After the message copy is generated, storage address corresponding to the message is labeled as upstate.
6. message preprocess method according to any one of claim 1 to 3, it is characterised in that the network accelerator
Engine carries out pretreatment to the message and specifically included according to the notification message：
In the case where network interface corresponding to the network accelerator engine is preconfigured as loopback, calling adds corresponding to the network
The network port driving of fast device engine；
Attribute information of the network port driving in the notification message, reads the message, and added from internal memory
Into transmission message queue corresponding to the network interface；And
The network interface receives the message of loopback, is passed to the network accelerator engine, and by the network accelerator
Engine pre-processes to the message.
7. a kind of message pretreatment unit, applied to business board, it is characterised in that the business board is configured with network accelerator and drawn
Hold up, the device includes：
Unit is established, for establishing the first DMA channel corresponding to PCI-E ports；
Writing unit, for the message of the forwarding chip from the business board to be write into internal memory by first DMA channel；
Generation unit, for generating notification message by the DMA drivings corresponding to first DMA channel, wherein the notice disappears
The attribute information of the message is included in breath；Wherein, the attribute information of the message includes：Storage of the message in internal memory
Address and the length of the message；
Transmitting element, for the notification message to be sent to the network accelerator engine；
Pretreatment unit, for, according to the notification message, being located in advance to the message by the network accelerator engine
Reason.
8. message pretreatment unit according to claim 7, it is characterised in that also include：
Judging unit, the source for the message for judging write-in internal memory is driven for the DMA；
Recording unit, in the message from the message queue corresponding to first DMA channel, and the message queue pair
In the case of pre-set business, the attribute information of the message should be recorded by DMA drivings, to generate the notification message.
9. message pretreatment unit according to claim 7, it is characterised in that if the message when writing internal memory, quilt
The head of preset length is added, then the attribute information of the message also includes：Corresponding to the address offset amount of the preset length.
10. the message pretreatment unit according to any one of claim 7 to 9, it is characterised in that the pretreatment unit
Specifically include：
First calls subelement, for calling the network interface DMA corresponding to the network accelerator engine；
Subelement is generated, for attribute informations of the network interface DMA in the notification message, the report is obtained from internal memory
Text simultaneously generates corresponding message copy；
Subelement is deposited, the message copy is deposited in corresponding network interface DMA physical memory for the network interface DMA
Location；And
First transmission subelement, the message copy is passed to the network accelerator engine for the network interface DMA, and by institute
Network accelerator engine is stated to pre-process the message copy.
11. message pretreatment unit according to claim 10, it is characterised in that also include：
Indexing unit, after being generated in the message copy, by storage address corresponding to the message labeled as available
State.
12. the message pretreatment unit according to any one of claim 7 to 9, it is characterised in that the pretreatment unit
Including：
Second calls subelement, in the case of being preconfigured in network interface corresponding to the network accelerator engine as loopback,
Call the network port driving corresponding to the network accelerator engine；
Add subelement, for attribute information of the network port driving in the notification message, read from internal memory described in
Message, and add it to and sent corresponding to the network interface in message queue；And
Second transmission subelement, the message of loopback is received for the network interface, is passed to the network accelerator engine,
And the message is pre-processed by the network accelerator engine.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410200835.1A CN103986585B (en) | 2014-05-13 | 2014-05-13 | Message preprocess method and its device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410200835.1A CN103986585B (en) | 2014-05-13 | 2014-05-13 | Message preprocess method and its device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103986585A CN103986585A (en) | 2014-08-13 |
| CN103986585B true CN103986585B (en) | 2018-03-16 |
Family
ID=51278412
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410200835.1A Active CN103986585B (en) | 2014-05-13 | 2014-05-13 | Message preprocess method and its device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103986585B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105991472A (en) * | 2015-03-03 | 2016-10-05 | 中兴通讯股份有限公司 | Message moving method and device |
| CN105786618B (en) * | 2016-02-24 | 2019-06-18 | 华为技术有限公司 | The method and apparatus of message are routed in accelerator network |
| CN105791021A (en) * | 2016-04-12 | 2016-07-20 | 上海斐讯数据通信技术有限公司 | Hardware acceleration device and method |
| CN108616414B (en) * | 2017-01-26 | 2021-12-03 | 华为技术有限公司 | Method and server for processing message |
| CN107124316B (en) * | 2017-05-23 | 2019-10-11 | 烽火通信科技股份有限公司 | Hardware based quick switching action implementation method in a kind of data communications equipment |
| CN111427814B (en) * | 2020-03-05 | 2021-12-07 | 深圳震有科技股份有限公司 | Inter-core communication method based on AMP system, terminal and storage medium |
| CN112929183B (en) * | 2021-01-26 | 2023-01-20 | 北京百度网讯科技有限公司 | Intelligent network card, message transmission method, device, equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101860485A (en) * | 2010-06-02 | 2010-10-13 | 上海融亿信息技术有限公司 | Network message filtering engine chip |
| WO2013055812A1 (en) * | 2011-10-14 | 2013-04-18 | Citrix Systems, Inc. | Systems and methods for dynamic adaptation of network accelerators |
| CN103200084A (en) * | 2013-03-18 | 2013-07-10 | 中兴通讯股份有限公司 | Message preprocessing method and device based on network processing unit and network processing unit |
-
2014
- 2014-05-13 CN CN201410200835.1A patent/CN103986585B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101860485A (en) * | 2010-06-02 | 2010-10-13 | 上海融亿信息技术有限公司 | Network message filtering engine chip |
| WO2013055812A1 (en) * | 2011-10-14 | 2013-04-18 | Citrix Systems, Inc. | Systems and methods for dynamic adaptation of network accelerators |
| CN103200084A (en) * | 2013-03-18 | 2013-07-10 | 中兴通讯股份有限公司 | Message preprocessing method and device based on network processing unit and network processing unit |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103986585A (en) | 2014-08-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103986585B (en) | Message preprocess method and its device | |
| US20220103452A1 (en) | Tracing logical network packets through physical network | |
| US10778557B2 (en) | Tracing network packets through logical and physical networks | |
| CN108055202B (en) | Message processing equipment and method | |
| WO2013136522A1 (en) | Computer system and method for communicating data between computers | |
| CN105656808A (en) | Message processing method and system thereof | |
| CN102752219B (en) | Method for implementing virtual device (VD) interconnection and switching equipment | |
| US9112708B1 (en) | Processing multicast packets in a network device | |
| CN104618432B (en) | A kind of processing method and processing system that event sends and receives | |
| CN105939324A (en) | Message forwarding method and device | |
| CN110380992A (en) | Message processing method, device and network flow acquire equipment | |
| CN107147543A (en) | A kind of socket communication means of server towards pc client | |
| CN106453156A (en) | Communication method and apparatus between virtual devices | |
| TW200814634A (en) | Systems and methods for packet forward control | |
| CN106850440B (en) | Router, routing method and chip for multi-address shared data routing packet | |
| CN107086960B (en) | Message transmission method and device | |
| CN103685062A (en) | Cache management method and device | |
| CN106502841B (en) | Data back up method and device | |
| CN106656656A (en) | Network device package capture method and device | |
| CN109714337A (en) | A kind of data encryption and transmission method and equipment | |
| CN109005122A (en) | File transmitting method, device and the network equipment | |
| CN107911317A (en) | A kind of method for dispatching message and device | |
| CN103368966A (en) | Server system and transmission method for data flow | |
| CN103491079B (en) | Message generating device, server and method | |
| CN108880892B (en) | Message processing method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 310052 Binjiang District Changhe Road, Zhejiang, China, No. 466, No. Applicant after: Xinhua three Technology Co., Ltd. Address before: 310052 Binjiang District Changhe Road, Zhejiang, China, No. 466, No. Applicant before: Huasan Communication Technology Co., Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |