CN103905300B - A kind of data message sending method, equipment and system - Google Patents
A kind of data message sending method, equipment and system Download PDFInfo
- Publication number
- CN103905300B CN103905300B CN201210572445.8A CN201210572445A CN103905300B CN 103905300 B CN103905300 B CN 103905300B CN 201210572445 A CN201210572445 A CN 201210572445A CN 103905300 B CN103905300 B CN 103905300B
- Authority
- CN
- China
- Prior art keywords
- message
- data message
- subdata
- node
- downstream node
- 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.)
- Expired - Fee Related
Links
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The embodiment of the present invention provides a kind of data message sending method, equipment and system.This method includes：When present node has data message to be sent, judge whether the current cache remaining space of downstream node is more than or equal to the length of the data message；If so, the present node sends the data message to the downstream node；If it is not, the data message is split as at least two subdata messages by the present node, and sends the subdata message one by one to the downstream node, wherein, the length of the subdata message is less than or equal to the current cache remaining space of the downstream node.A kind of data message sending method, equipment and system are provided according to the present invention, effectively reduces the link idle time, it is abundant to utilize link circuit resource.
Description
Technical field
The present invention relates to flow control technique, more particularly to a kind of data message sending method, equipment and system, belong to logical
Believe technical field.
Background technology
One complicated network system is by various kinds of equipment（Such as interchanger, router, core net）, and between these equipment
Connecting path composition.When source node sends a data message, the data message is usually required by some on network
Node can just eventually arrive at destination.Contain buffering area in each node, for keeping in the data not being disposed in the node
Message.When the out of buffers of node is to cache received data message, data message can be caused to lose, by transmitting terminal
Again corresponding data message is sent to the node, so as to cause the waste of network bandwidth.
In view of the above-mentioned problems, generally use in the following manner in current network system, to improve the utilization rate of network bandwidth：
Whether when present node needs forwarding data packets to downstream node, this can be accommodated by detecting the remaining cache space of downstream node
Data message, if can accommodate, the data message is sent to downstream node；If can not accommodate, do not save downstream temporarily
Point sends datagram, and waits the congestion of downstream node to release.
When each node in network system, which uses, carries out data message transmission with upper type, although avoiding due to number
The problem of causing data message to be lost with data message retransmission according to the caching deficiency of message receiving node, carry to a certain extent
The high utilization rate of network bandwidth, but because the premise of data-message transmission is：Downstream node must possess the complete number of reception
According to the ability of message, the transmission of data message can be just carried out, the free time of link is caused and wastes.
The content of the invention
For in the prior art the defects of, the present invention provides a kind of data message sending method, equipment and system, to have
Effect reduces the link idle time, makes full use of link circuit resource.
First aspect, there is provided a kind of data message sending method, including：
When present node has data message to be sent, judge whether the current cache remaining space of downstream node is more than
Equal to the length of the data message；
If so, the present node sends the data message to the downstream node；If it is not, the present node will
The data message is split as at least two subdata messages, and sends the subdata message one by one to the downstream node,
Wherein, the length of the subdata message is less than or equal to the current cache remaining space of the downstream node.
In the first possible implementation of first aspect, the data message to be sent works as prosthomere including described
The whole subdata messages for the composition data message that point receives from upstream node；
Correspondingly, the present node sends the data message to the downstream node, including：
The present node will form whole subdata messages of the data message, merge into the data message, and
Send to the downstream node.
In second of possible implementation of first aspect, the present node by the data message be split as to
Few two sub- data messages, including：
The present node judges, if according to default fractionation subdata message amount threshold value, the data message is entered
Row is split, and whether the length for splitting the subdata message of acquisition is less than or equal to the current cache remaining space of the downstream node；
If so, the data message then is split as into quantity is less than or equal to the son for splitting subdata message amount threshold value
Data message；
If it is not, then wait the downstream node to update the current cache remaining space, and according to current slow after renewal
Remaining space is deposited, re-executes whether the length for judging to split the subdata message obtained is less than or equal to the downstream node
Current cache remaining space operation.
With reference to first aspect or the first of first aspect or second of possible implementation, in the third of first aspect
In possible implementation, after the data message is split as at least two subdata messages by the present node, and to
Before the downstream node sends the subdata message one by one, in addition to：
The present node sets unique sub-mark in each subdata message, and is provided for indicating the son
Identify effective enable bit；
Correspondingly, it is described send the subdata message one by one to the downstream node after, in addition to：
The present node receives confirmation message that the downstream node returns, carrying the sub-mark or negative is true
Recognize message, and the sub-mark according to entrained by the confirmation message or NACK message, carried out to broadcasting caching again
Clear up and/or retransmitted.
Second aspect, there is provided a kind of data message sends equipment, including：
Judge module, for when data message to be sent be present, judging the current cache remaining space of downstream node
Whether the length of the data message is more than or equal to；
Processing module is sent, if being judged as YES for the judge module, the data message is sent to the downstream
Node；If the judge module is judged as NO, the data message is split as at least two subdata messages, and under described
Trip node sends the subdata message one by one, wherein, the length of the subdata message is less than or equal to the downstream node
Current cache remaining space.
In the first possible implementation of second aspect, the data message to be sent is sent out including data message
The whole subdata messages for forming the data message for sending equipment to be received from upstream node；Correspondingly, it is described to send processing mould
Block is additionally operable to：
Whole subdata messages of the data message will be formed, merge into the data message, and sent under described
Swim node.
In second of possible implementation of second aspect, the transmission processing module is used for：
If judging to split the data message according to default fractionation subdata message amount threshold value, split and obtain
Subdata message length whether be less than or equal to the downstream node current cache remaining space；
If so, the data message then is split as into quantity is less than or equal to the son for splitting subdata message amount threshold value
Data message；
If it is not, then wait the downstream node to update the current cache remaining space, and according to current slow after renewal
Remaining space is deposited, re-executes whether the length for judging to split the subdata message obtained is less than or equal to the downstream node
Current cache remaining space operation.
With reference to second aspect or the first of second aspect or second of possible implementation, in the third of second aspect
In possible implementation, the transmission processing module is additionally operable to：
Sent one by one after the data message is split as at least two subdata messages, and to the downstream node
Before the subdata message, unique sub-mark is set in each subdata message, and be provided for indicating the son
Identify effective enable bit；
Correspondingly, it is described send the subdata message one by one to the downstream node after, receive the downstream joint
Confirmation message or NACK message that point returns, carrying the sub-mark, and it is true according to the confirmation message or negative
Recognize the sub-mark entrained by message, cleared up and/or retransmitted to broadcasting caching again.
The third aspect, there is provided a kind of data message sends system, including data message provided by the invention sends equipment, with
And the downstream node being connected with data message transmission equipment.
Data message sending method, equipment and system according to embodiments of the present invention, because present node is sending data
During message, whether fractionation transmission is carried out to data message according to the cache resources automatic decision on link, i.e. if downstream node
The spatial cache of offer can not accommodate complete data message to be sent, as long as meeting the existing spatial cache of downstream node
The partial data of data message to be sent can be accommodated, it is possible to big message is split as small messages, anticipatory staging is completely reported
A part for text when downstream node can continue to receive, retransmits remaining TLP messages to downstream node.Therefore,
The link idle time is effectively reduced, avoids the waste of link circuit resource, and reduces the overall delay of transmitting message.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
There is the required accompanying drawing used in technology description to be briefly described, it should be apparent that, drawings in the following description are only this
Some embodiments of invention, for those of ordinary skill in the art, without having to pay creative labor, may be used also
To obtain other accompanying drawings according to these accompanying drawings.
Fig. 1 is the typical case schematic diagram of a scenario of the data message sending method of the embodiment of the present invention.
Fig. 2 is the schematic flow sheet of the data message sending method of the embodiment of the present invention.
Fig. 3 be the embodiment of the present invention data message sending method in message retransmission an example.
Fig. 4 is that the data message of the embodiment of the present invention sends the structural representation of equipment.
Fig. 5 is that another data message of the embodiment of the present invention sends the structural representation of equipment.
Fig. 6 is that the data message of the embodiment of the present invention sends the system architecture diagram of system.
Embodiment
Fig. 1 is the typical case schematic diagram of a scenario of the data message sending method of the embodiment of the present invention.As shown in figure 1, should
Improved PCI bus（Peripheral Component InterconnectExpress, PCIE）The net of system
Exemplified by network, according to the transmitting procedure of data message in network, it is assumed that three class nodes in network be present, be upstream respectively
（Upstream）It is node 11, current（Current）Node 12 and downstream（Downstream）Node 13.Wherein, upstream node
Data message is sent to by present node 12 by physical link, and data message is sent to downstream by present node by physical link
Node 13.Also, such as application of each node is based on credit（Credit-Based）Mechanism, i.e., using credit（Credit）Come
Spatial cache is represented, present node 12 need to consume by the Credit for providing downstream node 13 with data message to be sent
Credit is compared, to judge that can downstream node 13 accommodate the data message.
Wherein, upstream node 11, present node 12 and downstream node 13, which are only used for distinguishing, is transmitting a data message
When, the precedence of each node-node transmission data message, and it is not intended to limit the position of each node in a network.On moreover,
It can be the arbitrary network equipment in network, such as interchanger, router to swim node 11, present node 12 and downstream node 13
Or equipment of the core network etc..Hereinafter, the application scenarios with reference to shown in Fig. 1, to the data message sending method of the embodiment of the present invention
It is described in detail.
Fig. 2 is the schematic flow sheet of the data message sending method of the embodiment of the present invention.As shown in Fig. 2 the data message
Sending method includes：
Step 201, when present node has data message to be sent, the current cache remaining space of downstream node is judged
Whether the length of the data message is more than or equal to；
Specifically, by application Credit-Based mechanism PCIE systems network exemplified by, when Current nodes prepare to
Downstream node sends a transaction layer bag（Transaction Lay Packet, TLP）When message, Current nodes are first
The Credit first to return according to Downstream node feeding backs, it is empty to judge whether Downstream nodes can provide enough cachings
Between receiving the TLP messages.Wherein, Current nodes obtain the Credit of Downstream node feeding backs mode, both may be used
To be by Current node active inquiries, can also be limited herein by Downstream node timing feedbacks.
Such as the memory space made an appointment in network corresponding to 1 Credit is a, wherein a is natural number, is, for example,
2DW.If the Credit that downstream node provides to Current nodes is x, show the data reception capabilities of current downstream node（I.e.
Current cache remaining space）X*a DW, for example, the total capacity of the data buffer storage of downstream node is 8 × 2DW=16DW, then when
When the caching of downstream node is in idle, the Credit that it is provided to present node is 8.If present node is sent to downstream node
The length of TLP messages is y DW, and wherein y is natural number, then the Credit for needing consumption is y/a DW, for example, y/2 DW.
If the Credit that Downstream nodes provide is more than or equal to the Credit sent needed for the TLP messages,
Then judge that Downstream nodes can receive the TLP messages, i.e. the current cache remaining space of Downstream nodes is more than
Equal to the length of the TLP messages；Sent, whereas if the Credit that Downstream nodes provide is less than needed for the TLP messages
Credit if, then judge Downstream nodes can receive the TLP messages, the i.e. current cache of Downstream nodes
Remaining space is less than the length of the TLP messages.
Step 202, if so, the present node sends the data message to the downstream node；If it is not, described work as
The data message is split as at least two subdata messages by front nodal point, and sends the subnumber one by one to the downstream node
According to message, wherein, the length of the subdata message is less than or equal to the current cache remaining space of the downstream node.
Specifically, if judging to know that the current cache remaining space of Downstream nodes is more than or equal to by step 201
The length of the TLP messages, then directly send the TLP messages to Downstream nodes.If judge to know by step 201
The current cache remaining space of Downstream nodes is less than the length of the TLP messages, then the TLP messages is split as into more height
Message（Hereinafter referred to, small TLP messages）, so that the length of each small TLP messages is less than or equal to working as Downstream nodes
Preceding cache residual space.
For example, the Credit required for being currently needed for the TLP messages of transmission is y, if the TLP messages are split as z small
TLP messages, then the Credit needed for the small TLP messages after each fractionation is exactly y/z.Wherein, for z selection, y/z should be met
Less than or equal to x, wherein x is the Credit that downstream node provides.Because the length of small TLP messages saves less than or equal to Downstream
The current cache remaining space of point, therefore present node first can send one small TLP messages to Downstream nodes, if
Downstream nodes receive the cache residual space after the small TLP messages still greater than the length equal to small TLP messages, then currently
Node sends one small TLP messages to Downstream nodes again；If Downstream nodes receive slow after the small TLP messages
The length that remaining space is less than small TLP messages is deposited, then present node waits the current cache of the renewal of Downstream nodes to remain
Complementary space, and when renewal current cache remaining space be more than or equal to small TLP messages length when, present node again to
Downstream nodes send next small TLP messages.
According to the data message sending method of above-described embodiment, when being sent datagram due to present node, according to link
On cache resources automatic decision whether fractionation transmission is carried out to data message, i.e. if downstream node provide spatial cache
Complete data message to be sent can not be accommodated, as long as it is to be sent to meet that the existing spatial cache of downstream node can accommodate
Data message partial data, it is possible to big message is split as small messages, a part for anticipatory staging complete message arrives down
Node is swum, when downstream node can continue to receive, retransmits remaining TLP messages.Therefore, link is effectively reduced
Free time, the waste of link circuit resource is avoided, and reduce the overall delay of transmitting message.
Further, in the data message sending method of above-described embodiment, the present node is by the data message
At least two subdata messages are split as, including：
The present node judges, if according to default fractionation subdata message amount threshold value, the data message is entered
Row is split, and whether the length for splitting the subdata message of acquisition is less than or equal to the current cache remaining space of the downstream node；
If so, the data message then is split as into quantity is less than or equal to the son for splitting subdata message amount threshold value
Data message；
If it is not, then wait the downstream node to update the current cache remaining space, and according to current slow after renewal
Remaining space is deposited, re-executes whether the length for judging to split the subdata message obtained is less than or equal to the downstream node
Current cache remaining space operation.
Specifically, for example it is previously provided with present node and splits subdata message amount threshold value, it is detachable for limiting
Subdata message maximum quantity.For example, it is 4 to split subdata message amount threshold value, with present node to sent
Before TLP messages are split, first judge whether to meet y/4≤x, wherein：X is the Credit that downstream node provides, and y is current
Need the Credit required for the TLP messages that transmit.Only when meeting y/4≤x, present node is split to TLP messages, such as
4 small TLP messages are split as, if in addition, also meeting y/2≤x, TLP messages can also be split as 2 small TLP by present node
Message.
According to the data message sending method of above-described embodiment, subdata message amount threshold is split due to preset in node
Value, the hard-wired complexity of node can be reduced.
Further, in the data message sending method of above-described embodiment, the data message to be sent includes institute
State the whole subdata messages for the composition data message that present node receives from upstream node；
Correspondingly, the present node sends the data message to the downstream node, including：
The present node will form whole subdata messages of the data message, merge into the data message, and
Send to the downstream node.
Specifically, if Current nodes have received the subdata message after the fractionation that Upstream nodes send over,
Such as totally 4, and the data volume of each subdata message is 4 DW, then the data total amount of this 4 sub- data messages is 16
DW.When Current nodes forward again to Downstream nodes, if finding the Credit that Downstream nodes provide
More than 8, i.e. 16 DW, then present node this 4 sub- data messages are synthesized into original big TLP messages and are transmitted.More have
Body, in the specific implementation, each TLP messages have unique corresponding sequence identifier（Sequence ID）, in both link ends
Realize the order-preserving mechanism of TLP message transmissions.TLP messages are split after obtaining small TLP messages, each small TLP messages carry this
Sequence ID, i.e., the small TLP messages obtained are split by same TLP messages and carry identical Sequence ID.Therefore,
Current nodes can be with according to whether continuously receive multiple TLP messages for carrying identical Sequence ID, to judge
Whether the multiple TLP messages received are to split the small TLP messages obtained by same TLP messages.
Further, the unique sub- mark set by fractionation node can also be carried by splitting in each small TLP messages of acquisition
Know（SubID）, and by fractionation node set be used for indicate the effective enable bit of sub-mark.With upstream node to TLP messages
Split, i.e., exemplified by present node receives the small TLP messages after splitting, carried out as described below.
For example, original TLP messages are split as 4 small TLP messages by upstream node, and each small TLP messages are entered according to table 1
Row is set.Table 1 is SubID fields and sub-mark enable bit（SubID Enabled）Setting example, in this example, such as
Using PCIE agreements, prompt message is handled to the TLP among TLP headings（TLP Processing Hint, TPH）In field
Reservation（Reserve）Position carries out self-defined extension, and the TH positions of TLP headings are arranged into 1, to indicate to contain in current TLP
There is prompt message（Hint）.More specifically, the length of Reserve positions is 2 bytes, several bits therein are arranged to
SubID fields, it can be determined according to default fractionation subdata message amount threshold value, for example,.Split subdata message amount threshold
It is worth for 4, then using 2 bits as SubID fields, if it is 8 to split subdata message amount threshold value, needs to use 3 bits
As SubID fields.In addition, 1 bit is also arranged to SubID Enabled fields.In table 1, to use Bit 0 and Bit
1 is used as SubID fields, and using Bit 2 as exemplified by SubID Enabled fields.
Table 1
As shown in table 1, upstream node is respectively 00,01,10 and 11 for the SubID of 4 small messages distribution after fractionation, and
And the SubID Enabled fields in each small messages are set to indicate that sub-mark is effective, such as SubID Enabled take 1
Bit, when SubID Enabled are 0, instruction sub-mark is invalid, i.e., the message is the complete TLP messages not split, when
When SubID Enabled are 1, instruction sub-mark is effective, i.e., the message is the small TLP messages after fractionation.
Assuming that the Sequence ID of the TLP messages before splitting are ' b0001, in the 4 small TLP messages after fractionation, to need
To add SubID fields on the basis of Sequence ID, reform into：' b0001_00, ' b0001_01, ' b0001_10 and
‘b0001_11。
After present node receives this 4 small TLP messages, if SubID Enabled are 1, small TLP messages are parsed
Sequence id fields and SubID fields.Because the Sequence id fields of this 4 small TLP messages are identical, but SubID words
Duan Butong, then this 4 small TLP messages are can determine that to form the sub- message of same TLP messages, and take 2 according to SubID fields
Bit, present node can determine that this 4 small TLP messages are to form Sequence ID as ' all son reports of b0001 TLP messages
Text.Therefore, present node by the data summation of this 4 small TLP messages compared with the Credit that downstream node provides,
If transmit all small TLP messages belonged under a Sequence ID group（I.e. above-mentioned 4 small TLP messages）Data institute
Need Credit numerical value, the Credit that can be provided less than or equal to downstream node, then present node TLP this 4 small is reported
Text is combined into original big TLP message formats and is transmitted, and by the SubID Enabled positions 0 of the TLP messages after merging,
So that downstream node knows that the TLP messages are the complete TLP messages without fractionation.
With the main distinction of TLP messages after merging it is length before merging（Length）Field and check field.After merging,
The Length fields of TLP messages are exactly that the Length of four small TLP messages is directly added, because the TLP before splitting
The length of message must meet PCIE protocol specifications, so the length of the TLP messages after merging is also to meet specification.
In PCIE agreements, other fields of TLP messages, such as type（Type）, order（Order）Relation etc. remains unchanged, will not
Because big TLP messages are split as into small TLP messages, and small TLP messages are merged into big TLP messages and changed.
According to the data message sending method of above-described embodiment, because node can not only be in the cache resources on link not
Fractionation transmission is carried out to data message when sufficient, to reduce the link idle time, avoids the waste of link circuit resource, additionally it is possible to
In the cache resources abundance on link, the data message split is merged and sent, so as to reduce transmission delay, and improves band
It is wide.It is thereby achieved that according to the different situations of link circuit resource, message transmissions are more neatly carried out.
Further, in the data message sending method of above-described embodiment, the mark of the subdata message of acquisition is split,
That is Sequence ID and SubID, it can be also used for realizing the order-preserving mechanism of small TLP message transmissions at the both ends of link.Specifically
Ground, TLP messages are split with present node, the small TLP messages after fractionation transmit between present node and downstream node
Exemplified by this process, carry out as described below.
Downstream node, i.e., the receiving terminal of small TLP messages, it is cached with the relevant information that every pen data message transmits.Table 2 is to connect
The list of the data cached message transmission relevant information of receiving end.
Table 2
| Sequence ID | SubID | SubID Enabled | Data payload (Data Payload) |
| ‘b0001 | ‘b00 | 1 | ...... |
| ‘b0001 | ‘b01 | 1 | ...... |
| ‘b0001 | ‘b10 | 1 | ...... |
| ‘b0001 | ‘b11 | 1 | ...... |
| ‘b0010 | ‘b00 | 0 | ...... |
As shown in table 2, receiving terminal records original Sequence ID, is split as the SubID added after small TLP messages, with
And the whether effective enable bits of SubID.
Following ack/nack（ACK/NAK）When agreement carries out data transmission, if receiving terminal is reported according to small TLP
Check information in text（Table 2 is not shown）, verify and know that Sequence ID+SubID are that ' b0001_10 small TLP messages occur
Mistake is, it is necessary to by present node（The transmitting terminal of i.e. small TLP messages）During re-transmission, then NAK messages are returned to transmitting terminal, and in NAK
The Sequence ID+SubID, i.e. ' b0001_01 of last small TLP message being properly received are carried in message.Transmitting terminal
After receiving the NAK messages, by ' small TLP messages corresponding to b0001_01 and the TLP messages that transmit before it from broadcasting again
Removed in caching, and to ' small TLP messages corresponding to b0001_10 and the TLP messages transmitted after which retransmit.
More specifically, Fig. 3 be the embodiment of the present invention data message sending method in message retransmission an example.It is logical
The example shown in Fig. 3 is crossed, the situation that NAK messages are received to transmitting terminal is described in more detail.
As shown in figure 3, transmitting terminal sends TLP1 ~ TLP5 to receiving terminal, wherein TLP1 is first message, and TLP5 is most
The latter message.Receiving terminal sequentially receives TLP1 ~ 3, and TLP4 and TLP5 are still in transmit process.In receiving terminal, TLP3 does not have
By completeness check, now receiving terminal sends NAK messages to transmitting terminal, and include No. Sequence of the NAK messages is
' b0001_01, i.e. last small TLP message being properly received Sequence ID+SubID.Transmitting terminal, which receives, to be carried
' after b0001_01 NAK messages, learn that TLP1 and TLP2 have been successfully received.Now transmitting terminal stop from upstream node after
New TLP is received in continued access, TLP1 and TLP2 is removed among the caching of broadcasting again of itself, and it is slow to be sent in broadcasting again again
Message in depositing since TLP3, in this example, transmitting terminal will resend TLP3 ~ TLP5.
According to the data message sending method of above-described embodiment, pass through SequenceID, SubID and SubID Enable
Position has been able to the integrity control mechanism after one big data message of guarantee is splitted into subdata message.Due to each process
Subdata message after fractionation includes all complete information of a data message, so by each son after fractionation
Data message regards an independent TLP message as, only difference is that it has the possibility being merged again, can but this
Its any behavior in a link can be had no effect on.Therefore group data message makes a mistake, it is necessary to when retransmitting, only need to be corresponding
Subdata start of heading part subdata message retransmission, without the result after whole split is retransmitted.
Fig. 4 is that the data message of the embodiment of the present invention sends the structural representation of equipment.As shown in figure 4, the data message
Sending equipment 40 includes：
Judge module 41, the current cache for when data message to be sent be present, judging downstream node are remaining empty
Between whether be more than or equal to the data message length；
Processing module 42 is sent, if being judged as YES for the judge module 41, the data message is sent to described
Downstream node；If the judge module 41 is judged as NO, the data message is split as at least two subdata messages, and to
The downstream node sends the subdata message one by one, wherein, the length of the subdata message is less than or equal to the downstream
The current cache remaining space of node.
It is, for example, in the data message sending method of above-described embodiment that the data message of the embodiment of the present invention, which sends equipment,
Present node, its corresponding hardware device can be the arbitrary network equipment in network, such as interchanger, router or core net
Equipment etc..In hardware realization, above judge module and transmission processing module can be CPU, or single
Piece machine, and transmission processing module can also be integrated with transmitter, to realize that the message to downstream node is sent.
The idiographic flow and the number of above-described embodiment that the data message transmission equipment of the embodiment of the present invention sends datagram
It is identical according to file transmitting method, therefore here is omitted.
Data message according to embodiments of the present invention sends equipment, due to when sending datagram, according on link
Whether cache resources automatic decision carries out fractionation transmission to data message, i.e. if the spatial cache that downstream node provides can not
Complete data message to be sent is enough accommodated, as long as meeting that the existing spatial cache of downstream node can accommodate number to be sent
According to the partial data of message, it is possible to big message is split as into small messages, the part of anticipatory staging complete message to downstream joint
Point, when downstream node can continue to receive, retransmit remaining TLP messages.Therefore, link idle is effectively reduced
Time, the waste of link circuit resource is avoided, and reduce the overall delay of transmitting message.
Further, in the data message of above-described embodiment sends equipment, the data message to be sent includes number
The whole subdata messages for the composition data message that equipment receives from upstream node are sent according to message；Correspondingly, the hair
Processing module is sent to be additionally operable to：
Whole subdata messages of the data message will be formed, merge into the data message, and sent under described
Swim node.
Further, in the data message of above-described embodiment sends equipment, the transmission processing module is used for：
If judging to split the data message according to default fractionation subdata message amount threshold value, split and obtain
Subdata message length whether be less than or equal to the downstream node current cache remaining space；
If so, the data message then is split as into quantity is less than or equal to the son for splitting subdata message amount threshold value
Data message；
If it is not, then wait the downstream node to update the current cache remaining space, and according to current slow after renewal
Remaining space is deposited, re-executes whether the length for judging to split the subdata message obtained is less than or equal to working as the downstream node
The operation of preceding cache residual space.
Further, in the data message of above-described embodiment sends equipment, the transmission processing module is additionally operable to：
Sent one by one after the data message is split as at least two subdata messages, and to the downstream node
Before the subdata message, unique sub-mark is set in each subdata message, and be provided for indicating the son
Identify effective enable bit；
Correspondingly, it is described send the subdata message one by one to the downstream node after, receive the downstream joint
Confirmation message or NACK message that point returns, carrying the sub-mark, and it is true according to the confirmation message or negative
Recognize the sub-mark entrained by message, cleared up and/or retransmitted to broadcasting caching again.
Equipment is sent according to the data message of above-described embodiment, passes through SequenceID, SubID and SubID Enable
Position has been able to the integrity control mechanism after one big data message of guarantee is splitted into subdata message.Due to each process
Subdata message after fractionation includes all complete information of a data message, so by each son after fractionation
Data message regards an independent TLP message as, only difference is that it has the possibility being merged again, can but this
Its any behavior in a link can be had no effect on.Therefore group data message makes a mistake, it is necessary to when retransmitting, only need to be corresponding
Subdata start of heading part subdata message retransmission, without the result after whole split is retransmitted.
Fig. 5 is that another data message of the embodiment of the present invention sends the structural representation of equipment.As shown in figure 5, the data
Message, which sends equipment 50, includes memory 51 and the processor 52 being connected with memory, wherein：One group of journey is stored in memory 51
Sequence code, and processor 52 is used to call the program code stored in memory 51, for performing following operation：
When data message to be sent be present, judge the current cache remaining space of downstream node whether more than or equal to described
The length of data message；
If so, the data message is sent to the downstream node；If it is not, the data message is split as at least two
Individual sub- data message, and the subdata message is sent one by one to the downstream node, wherein, the length of the subdata message
Less than or equal to the current cache remaining space of the downstream node.
It is, for example, in the data message sending method of above-described embodiment that the data message of the embodiment of the present invention, which sends equipment,
Present node, its corresponding hardware device can be the arbitrary network equipment in network, such as interchanger, router or core net
Equipment etc..
The idiographic flow and the number of above-described embodiment that the data message transmission equipment of the embodiment of the present invention sends datagram
It is identical according to file transmitting method, therefore here is omitted.
Data message according to embodiments of the present invention sends equipment, due to when sending datagram, according on link
Whether cache resources automatic decision carries out fractionation transmission to data message, i.e. if the spatial cache that downstream node provides can not
Complete data message to be sent is enough accommodated, as long as meeting that the existing spatial cache of downstream node can accommodate number to be sent
According to the partial data of message, it is possible to big message is split as into small messages, the part of anticipatory staging complete message to downstream joint
Point, when downstream node can continue to receive, retransmit remaining TLP messages.Therefore, link idle is effectively reduced
Time, the waste of link circuit resource is avoided, and reduce the overall delay of transmitting message.
Further, in the data message of above-described embodiment sends equipment, the data message to be sent includes number
The whole subdata messages for the composition data message that equipment receives from upstream node are sent according to message；Correspondingly, processor
It is additionally operable to call the program code stored in memory, whole subdata messages of the data message will be formed, merge into institute
Data message is stated, and is sent to the downstream node.
Further, in the data message of above-described embodiment sends equipment, processor is additionally operable to call and deposited in memory
The program code of storage, if judging to split the data message according to default fractionation subdata message amount threshold value, tear open
Whether the length of the subdata message separately won is less than or equal to the current cache remaining space of the downstream node；
If so, the data message then is split as into quantity is less than or equal to the son for splitting subdata message amount threshold value
Data message；
If it is not, then wait the downstream node to update the current cache remaining space, and according to current slow after renewal
Remaining space is deposited, re-executes whether the length for judging to split the subdata message obtained is less than or equal to the downstream node
Current cache remaining space operation.
Further, in the data message of above-described embodiment sends equipment, processor is additionally operable to call and deposited in memory
The program code of storage, after the data message is split as at least two subdata messages, and to the downstream node by
Before the individual transmission subdata message, following operate is performed：
Unique sub-mark is set in each subdata message, and is provided for indicating that the sub-mark effectively makes
Can position；
Correspondingly, after the subdata message is sent one by one to the downstream node, following operate also is performed：
Receive confirmation message or NACK message that the downstream node returns, carrying the sub-mark, and root
According to the sub-mark entrained by the confirmation message or NACK message, cleared up and/or entered to broadcasting caching again
Row retransmits.
Fig. 6 is that the data message of the embodiment of the present invention sends the system architecture diagram of system.As shown in fig. 6, the data message
The data message that transmission system includes any of the above-described embodiment sends equipment 61, and sends equipment 61 with the data message and connect
The downstream node 62 connect.
In the data message transmission system of the embodiment of the present invention, the idiographic flow that sends datagram and above-described embodiment
Data message sending method is identical, therefore here is omitted.
Data message according to embodiments of the present invention sends system, is being sent datagram because data message sends equipment
When, whether fractionation transmission is carried out to data message according to the cache resources automatic decision on link, i.e. if downstream node provides
Spatial cache can not accommodate complete data message to be sent, as long as meeting that the existing spatial cache of downstream node can
Accommodate the partial data of data message to be sent, it is possible to big message is split as small messages, anticipatory staging complete message
A part arrives downstream node, when downstream node can continue to receive, retransmits remaining TLP messages.Therefore, effectively
Reduce the link idle time, avoid the waste of link circuit resource, and reduce the overall delay of transmitting message.
One of ordinary skill in the art will appreciate that：Realizing all or part of step of above-mentioned each method embodiment can lead to
The related hardware of programmed instruction is crossed to complete.Foregoing program can be stored in a computer read/write memory medium.The journey
Sequence upon execution, execution the step of including above-mentioned each method embodiment；And foregoing storage medium includes：ROM, RAM, magnetic disc or
Person's CD etc. is various can be with the medium of store program codes.
Finally it should be noted that：The above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although
The present invention is described in detail with reference to the foregoing embodiments, it will be understood by those within the art that：It still may be used
To be modified to the technical scheme described in foregoing embodiments, or equivalent substitution is carried out to which part technical characteristic；
And these modification or replace, do not make appropriate technical solution essence depart from various embodiments of the present invention technical scheme spirit and
Scope.
Claims (3)
- A kind of 1. data message sending method, it is characterised in that including：When present node has data message to be sent, judge whether the current cache remaining space of downstream node is more than or equal to The length of the data message；If so, the present node sends the data message to the downstream node；If it is not, the present node is by described in Data message is split as at least two subdata messages, and sends the subdata message one by one to the downstream node, wherein, The length of the subdata message is less than or equal to the current cache remaining space of the downstream node, and each subdata message is taken Sequence identifier with the data message；After the data message is split as at least two subdata messages by the present node, and to the downstream node by Before the individual transmission subdata message, in addition to：The present node sets unique sub-mark in each subdata message, and is provided for indicating the sub-mark Effective enable bit；Correspondingly, it is described send the subdata message one by one to the downstream node after, in addition to：The present node receives confirmation message or NACK report that the downstream node returns, carrying the sub-mark Text, and the sub-mark according to entrained by the confirmation message or NACK message, are cleared up broadcasting caching again And/or retransmitted；The data message to be sent includes the complete of the composition data message that the present node receives from upstream node Portion's subdata message；Correspondingly, the present node sends the data message to the downstream node, including：The present node will form whole subdata messages of the data message, merge into the data message, and send To the downstream node；The data message is split as at least two subdata messages by the present node, including：The present node judges, if according to default fractionation subdata message amount threshold value, the data message is torn open Point, whether the length for splitting the subdata message of acquisition is less than or equal to the current cache remaining space of the downstream node；If so, the data message then is split as into quantity is less than or equal to the subdata for splitting subdata message amount threshold value Message；If it is not, then waiting the downstream node to update the current cache remaining space, and remained according to the current cache after renewal Whether complementary space, the length for re-executing the subdata message for judging fractionation and obtaining are less than or equal to working as the downstream node The operation of preceding cache residual space.
- 2. a kind of data message sends equipment, it is characterised in that including：Judge module, for when data message to be sent be present, judge downstream node current cache remaining space whether More than or equal to the length of the data message；Processing module is sent, if being judged as YES for the judge module, the data message is sent to the downstream node； If the judge module is judged as NO, the data message is split as at least two subdata messages, and to the downstream joint Point sends the subdata message one by one, wherein, the length of the subdata message is less than or equal to the current of the downstream node Cache residual space, each subdata message carry the sequence identifier of the data message；The transmission processing module is additionally operable to：After the data message is split as at least two subdata messages, and send one by one to the downstream node described Before subdata message, unique sub-mark is set in each subdata message, and be provided for indicating the sub-mark Effective enable bit；Correspondingly, it is described send the subdata message one by one to the downstream node after, receive the downstream node and return Confirmation message or NACK message return, carrying the sub-mark, and according to the confirmation message or NACK report The sub-mark entrained by text, cleared up and/or retransmitted to broadcasting caching again；The data message to be sent includes data message and sends the composition data message that equipment receives from upstream node Whole subdata messages；Correspondingly, the transmission processing module is additionally operable to：Whole subdata messages of the data message will be formed, merge into the data message, and send to the downstream joint Point；The transmission processing module is used for：If judging the data message to be split according to default fractionation subdata message amount threshold value, split the son of acquisition Whether the length of data message is less than or equal to the current cache remaining space of the downstream node；If so, the data message then is split as into quantity is less than or equal to the subdata for splitting subdata message amount threshold value Message；If it is not, then waiting the downstream node to update the current cache remaining space, and remained according to the current cache after renewal Whether complementary space, the length for re-executing the subdata message for judging fractionation and obtaining are less than or equal to working as the downstream node The operation of preceding cache residual space.
- 3. a kind of data message sends system, it is characterised in that equipment is sent including the data message described in claim 2, with And the downstream node being connected with data message transmission equipment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210572445.8A CN103905300B (en) | 2012-12-25 | 2012-12-25 | A kind of data message sending method, equipment and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210572445.8A CN103905300B (en) | 2012-12-25 | 2012-12-25 | A kind of data message sending method, equipment and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103905300A CN103905300A (en) | 2014-07-02 |
| CN103905300B true CN103905300B (en) | 2017-12-15 |
Family
ID=50996457
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210572445.8A Expired - Fee Related CN103905300B (en) | 2012-12-25 | 2012-12-25 | A kind of data message sending method, equipment and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103905300B (en) |
Families Citing this family (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105933230A (en) * | 2016-04-01 | 2016-09-07 | 浪潮电子信息产业股份有限公司 | Multi-node data transmission method, node and computer system |
| CN109565699B (en) * | 2016-08-19 | 2020-12-08 | 华为技术有限公司 | Network node and method thereof |
| CN106534161B (en) * | 2016-12-02 | 2019-08-02 | 浪潮金融信息技术有限公司 | A kind of transaction message processing method, device and system |
| CN109309563B (en) * | 2017-07-27 | 2021-12-24 | 科大国盾量子技术股份有限公司 | Information error correction processing method and system |
| CN107404340A (en) * | 2017-08-29 | 2017-11-28 | 广东欧珀移动通信有限公司 | Transmission method, device, terminal and the storage medium of Bluetooth audio frequency data |
| CN109039900B (en) * | 2018-08-10 | 2020-06-12 | 北京邮电大学 | Method and device for sending stop signal of credit packet |
| CN111416776A (en) * | 2019-01-07 | 2020-07-14 | 华为技术有限公司 | Method and network device for transmitting data |
| CN112463739A (en) * | 2019-09-09 | 2021-03-09 | 山东省计算中心（国家超级计算济南中心） | Data processing method and system based on ocean mode ROMS |
| CN110708682B (en) * | 2019-09-12 | 2023-04-28 | 中国神华能源股份有限公司神朔铁路分公司 | Vehicle-ground data communication method and device, computer equipment and storage medium |
| CN111162830B (en) * | 2019-12-12 | 2022-03-04 | 航天东方红卫星有限公司 | Satellite-ground data transmission routing method based on track forecast |
| CN113645678B (en) * | 2020-04-27 | 2023-01-03 | 阿里巴巴集团控股有限公司 | Data message sending and receiving methods, network distribution method, related device and system |
| CN113300983B (en) * | 2020-05-29 | 2022-05-31 | 阿里巴巴集团控股有限公司 | Flow control method, device, system and storage medium |
| CN112769710A (en) * | 2021-01-21 | 2021-05-07 | 中国电子科技集团公司电子科学研究院 | Router and operation method thereof |
| CN113992317A (en) * | 2021-09-28 | 2022-01-28 | 山东云海国创云计算装备产业创新中心有限公司 | SPI communication method, device and storage medium |
| CN116743324B (en) * | 2023-08-16 | 2023-11-03 | 苏州浪潮智能科技有限公司 | Data transmission method, device, receiving end, storage medium, system and server |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1492644A (en) * | 2002-10-22 | 2004-04-28 | 华为技术有限公司 | Processing method for super long message in two layered virtual special network |
| CN101924848A (en) * | 2009-06-11 | 2010-12-22 | 富士施乐株式会社 | Electronic device, and image forming device and system |
| CN102185868A (en) * | 2011-05-20 | 2011-09-14 | 杭州华三通信技术有限公司 | Authentication method, system and equipment based on extensible authentication protocol (EAP) |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4066613B2 (en) * | 2000-04-05 | 2008-03-26 | 富士ゼロックス株式会社 | Display control apparatus, image forming apparatus, and display control program |
| US20090103124A1 (en) * | 2005-08-31 | 2009-04-23 | Canon Kabushiki Kaisha | Image forming apparatus, mobile device, and control method therefor |
| CN101699914A (en) * | 2009-11-11 | 2010-04-28 | 青岛海信移动通信技术股份有限公司 | Up-link and down-link service data processing method and relevant device in mobile terminal |
| CN101883041B (en) * | 2010-06-29 | 2015-07-22 | 中兴通讯股份有限公司 | Storing and forwarding system and message storing method thereof |
| CN101951339B (en) * | 2010-08-24 | 2012-08-15 | 北京星网锐捷网络技术有限公司 | Message speed limitation method, device and network equipment |
-
2012
- 2012-12-25 CN CN201210572445.8A patent/CN103905300B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1492644A (en) * | 2002-10-22 | 2004-04-28 | 华为技术有限公司 | Processing method for super long message in two layered virtual special network |
| CN101924848A (en) * | 2009-06-11 | 2010-12-22 | 富士施乐株式会社 | Electronic device, and image forming device and system |
| CN102185868A (en) * | 2011-05-20 | 2011-09-14 | 杭州华三通信技术有限公司 | Authentication method, system and equipment based on extensible authentication protocol (EAP) |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103905300A (en) | 2014-07-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103905300B (en) | A kind of data message sending method, equipment and system | |
| US6338091B1 (en) | System for optimistic transmission flow control including receiver data discards upon inadequate buffering condition | |
| CN104484295B (en) | Receiver sliding window-based data transmission method in parallel computer system | |
| US11381514B2 (en) | Methods and apparatus for early delivery of data link layer packets | |
| US6393023B1 (en) | System and method for acknowledging receipt of messages within a packet based communication network | |
| US8166227B2 (en) | Apparatus for processing peripheral component interconnect express protocol | |
| US8310934B2 (en) | Method and device for controlling information channel flow | |
| US8774203B2 (en) | One-way message notification with out-of-order packet delivery | |
| US6542513B1 (en) | Optimistic, eager rendezvous transmission mode and combined rendezvous modes for message processing systems | |
| CN111030747B (en) | FPGA-based SpaceFibre node IP core | |
| EP2061211A1 (en) | Packet recovery method, communication system, information processing device, and program | |
| CN105897385A (en) | Method for acknowledging transmission of RLC (Radio Link Control) data packet and RLC AM (Acknowledge Mode) entity sender | |
| US8792512B2 (en) | Reliable message transport network | |
| WO2010022629A1 (en) | Method, equipment and system for segment data recombining on multilink protocol | |
| CN103944880B (en) | A kind of method of ZigBee data transfers | |
| CN104065465B (en) | A kind of method of message retransmission, request end, responder and system | |
| CN100589477C (en) | Method for realizing data sending based on binding link | |
| CN104980257A (en) | Internet of things communication method and device | |
| CN103368703B (en) | Data package retransmission method, data packet receiving method and device | |
| CN106304188B (en) | A kind of AM RLC is for retransmitting the receiving handling method of the optimization of data | |
| CN102201901A (en) | Data retransmission method and device | |
| CN107222379A (en) | A kind of method and apparatus of serial communication | |
| CN101742451A (en) | Method, system and device for transmitting charging information | |
| CN113973091A (en) | Message processing method, network equipment and related equipment | |
| CN103391257A (en) | Message storing and forwarding method, device and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20171215 Termination date: 20201225 |