CN103441937A - Sending method and receiving method of multicast data - Google Patents
Sending method and receiving method of multicast data Download PDFInfo
- Publication number
- CN103441937A CN103441937A CN2013103669357A CN201310366935A CN103441937A CN 103441937 A CN103441937 A CN 103441937A CN 2013103669357 A CN2013103669357 A CN 2013103669357A CN 201310366935 A CN201310366935 A CN 201310366935A CN 103441937 A CN103441937 A CN 103441937A
- Authority
- CN
- China
- Prior art keywords
- multicast
- multicast packets
- packets
- data
- packet
- 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.)
- Pending
Links
Images
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a sending method and a receiving method of multicast data. The sending method includes the steps that the multicast data needing sending are separated into multiple multicast packets; parameter information is added to a packet header of each multicast packet, and the parameter information is used for showing the position of the current multicast packet in the multicast data; multicast sending is performed on the multicast packets after the parameter information is added to the multicast packets. According to the sending method and the receiving method of the multicast data, by separating the multicast data, sending the separated multicast packets to provide an efficient multicast communication mechanism and shortening data time delay and by adding the packet headers which are used for determining the position of each multicast packet in the multicast data on the separated multicast packets, integrity and orderliness in sending the multicast packets are facilitated, and the communication efficiency is further improved.
Description
Technical field
The present invention relates to computer realm, and especially, relate to a kind of sending method and method of reseptance of multicast packet.
Background technology
The InfiniBand framework is a kind of how concurrent link " cable conversion " technology of supporting, is mainly used in the I/O communication of bottom, realizes the effect of two-forty, low delay on serial link.The InfiniBand framework is as a kind of unified interconnect architecture, not only can be for storing I/O and network I/O, and can be for interprocess communication.The main feature of InfiniBand framework is high bandwidth, the SDR(Single Date Rate of single link, single data rate) bandwidth is 2.5Gbps, DDR(Double Date Rate, double data rate (DDR)) bandwidth is 5Gbps, QDR(Quad Date Rate, four figures is according to rate) bandwidth is 10Gbps, if need larger bandwidth, only needing to increase number of links gets final product, such as the QDR bandwidth of 4 times can reach 40Gbps), low time delay (switch time delay 140ns, application program time delay 3 μ s, new network interface card technology will make the application program time delay be reduced to 1 μ s level), system extension good (can easily realize the complete InfiniBand network without congested tens thousand of end equipment).In addition, the InfiniBand standard is supported RDMA(Remote Direct Memory Access, direct distance input and output), make when using InfiniBand to construct server, storage network than ten thousand mbit ethernets and FC(Fibre Channel, netted passage) there is higher performance, efficiency and flexibility.
The realization of the multi-client based on InfiniBand,, after server end receives the connection request of a plurality of clients, connect with a plurality of clients, and then server end sends to a plurality of clients by same piece of data successively.As shown in Figure 1, the multicast of InfiniBand can be realized based on hardware, after server node (transmit leg) connects with a plurality of clients, hand over client the necessary information of protecting, the data that receive are added to QP(Queue paar, it is queue team, do not illustrate in the drawings, comprise transmit queue and receiving queue) receiving queue in, the multicast packet that is arranged in transmit queue can be sent to a plurality of recipients (Fig. 1 just lists as three clients as example) simultaneously effectively through switch, then, server node adds CQ(Completion queue by the data that have been sent completely, complete queue).After switch receives the multicast packets of server end transmission, can copy this multicast packets, and send this multicast packets to a plurality of recipients that add this multicast group, the recipient adds the data that receive in the receiving queue of QP, then the data in receiving queue are submitted to the user through transmit queue, the data of having submitted to add CQ.
Although InfiniBand has the characteristic of two-forty, low delay, by analyze and actual test known, in the design to the multi-client support, data delay during transmission can be basically along with the growth of client terminal quantity linear growth.
The problem of the data delay caused for multicast data transmission in correlation technique to multi-client, not yet propose effective solution at present.
Summary of the invention
The problem of the data delay caused for multicast data transmission in correlation technique to multi-client, the present invention proposes a kind of sending method and method of reseptance and the management devices of server data transmission and management devices of receiving terminal transfer of data of multicast packet, can reduce the data delay while propagating multicast packet, improve communication efficiency.
Technical scheme of the present invention is achieved in that
A kind of sending method of multicast packet is provided according to an aspect of the present invention.
This transfer approach comprises:
The multicast packet that needs are sent is split as a plurality of multicast packets;
Parameter information is added in packet header in each multicast packets, and parameter information is for meaning the position of current multicast packets at multicast packet;
Multicast packets after the interpolation parameter information is carried out to the multicast transmission.
Wherein, the parameter information of each multicast packets comprises:
The first data head and the second data head, wherein, the first data head is for meaning to be split by multicast packet the quantity of the multicast packets obtained, the second data head is for meaning the sequence of current multicast packets in a plurality of multicast packets.
Preferably, the length of this multicast packets is less than 2K.
Further, the multicast packets after the interpolation parameter information being carried out to the multicast transmission comprises:
Send the multicast packets after encapsulating based on the direct distance input and output rdma protocol.
And, the multicast packets after the interpolation parameter information is carried out to the multicast transmission and comprises:
Multicast packets is sent to switch, by switch, multicast packets is copied and the multicast packets copied is sent to receiving terminal.
A kind of method of reseptance of multicast packet is provided according to an aspect of the present invention.
This method of reseptance comprises:
Receive in advance and split by multicast packet a plurality of multicast packets that obtain;
Determine the position of current multicast packets in multicast packet according to the parameter information in packet header of each multicast packets;
According to the position of each multicast packets, this multicast packets is added to receiving queue.
Further, the parameter information of each multicast packets comprises:
The first data head and the second data head, wherein, the first data head is for meaning to be split by multicast packet the quantity of the multicast packets obtained, the second data head is for meaning the sequence of current multicast packets in a plurality of multicast packets.
A kind of management devices of server data transmission is provided according to an aspect of the present invention.
The management devices of this server data transmission comprises:
Main control module, for server is being carried out in initialized situation, module, sending module, inquiry transmission block of state are monitored in configuration;
Monitor module for setting up and maintaining and the communicating by letter of recipient;
Sending module will be for needing the multicast packet sent to be split as a plurality of multicast packets, and add parameter information in the packet header of each multicast packets, parameter information is used for meaning the position of current multicast packets at multicast packet, and the multicast packets after the interpolation parameter information is carried out to the multicast transmission;
Inquiry sends block of state for inquiring about the transmission state of each multicast packets, and the control sending module sends one by one to a plurality of multicast packets.
Wherein, above-mentioned monitoring module comprises for setting up and maintaining with communicating by letter of recipient:
Set up and maintain and the communicating by letter of recipient based on the direct distance input and output rdma protocol.
Further, the parameter information of each multicast packets comprises:
The first data head and the second data head, wherein, the first data head is for meaning to be split by multicast packet the quantity of the multicast packets obtained, the second data head is for meaning the sequence of current multicast packets in a plurality of multicast packets.
In addition, sending module is for carrying out multicast and send and comprise adding multicast packets after parameter information:
Multicast packets is sent to switch, by switch, multicast packets is copied and the multicast packets copied is sent to receiving terminal.
A kind of management devices of receiving terminal transfer of data is provided according to an aspect of the present invention.
The management devices of this receiving terminal transfer of data comprises:
Main control module, for receiving terminal is being carried out in initialized situation, configuration receiver module, enquiry module;
Receiver module splits by multicast packet a plurality of multicast packets that obtain for receiving in advance, determine the position of current multicast packets in multicast packet according to the parameter information in packet header of each multicast packets, and according to the position of each multicast packets, this multicast packets is added to receiving queue;
Enquiry module, in the situation that inquire receiving queue, having data, is submitted the data in receiving queue to.
Wherein, the parameter information of each multicast packets comprises:
The first data head and the second data head, wherein, the first data head is for meaning to be split by multicast packet the quantity of the multicast packets obtained, the second data head is for meaning the sequence of current multicast packets in a plurality of multicast packets.
The present invention is by splitting multicast packet, then the multicast packets sent after splitting can provide efficient cast communication mechanism, and minimizing data delay, in addition, add the packet header with definite each multicast packets position in multicast packet by giving the multicast packets after splitting, be conducive to integrality and order that multicast packets sends, further improve communication efficiency.
The accompanying drawing explanation
Fig. 1 is the schematic diagram of multicast data transmission in prior art;
Fig. 2 is the flow chart according to the sending method of the multicast packet of the embodiment of the present invention;
Fig. 3 is the schematic diagram of transmission multicast packet according to an embodiment of the invention;
Fig. 4 is the block diagram according to the management devices of the server data transmission of the embodiment of the present invention;
Fig. 5 is the schematic diagram of server end operation according to another embodiment of the invention;
Fig. 6 is the flow chart according to the method for reseptance of the multicast packet of the embodiment of the present invention;
Fig. 7 is the flow chart of the method for reseptance of multicast packet according to still another embodiment of the invention;
Fig. 8 is the block diagram according to the management devices of the receiving terminal transfer of data of the embodiment of the present invention;
Fig. 9 is the flow chart for server end receiving step according to still a further embodiment;
Figure 10 is the schematic diagram according to the CS model of the embodiment of the present invention.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is clearly and completely described, obviously, described embodiment is only the present invention's part embodiment, rather than whole embodiment.Embodiment based in the present invention, the every other embodiment that those of ordinary skills obtain, belong to the scope of protection of the invention.
According to embodiments of the invention, provide a kind of sending method of multicast packet.
As shown in Figure 2, according to the sending method of the embodiment of the present invention, can comprise:
Step S201, the multicast packet that needs are sent is split as a plurality of multicast packets, and each multicast packets need to meet length requirement, and preferably, the length of this multicast packets is less than 2K;
Step S203, add parameter information in the packet header of each multicast packets, and parameter information is for meaning the position of current multicast packets at multicast packet.
Step S205, multicast packets after the interpolation parameter information is carried out to the multicast transmission, and preferably, send the multicast packets after encapsulating based on the direct distance input and output rdma protocol, multicast packets is sent to switch (can be stored in the buffer memory of switch), multicast packets is copied and the multicast packets copied is sent to a plurality of receiving terminals by switch, and, the quantity of the client based on needing the receiving group bag is determined the umber copied, in addition, a plurality of receiving terminals of receiving multicast data that need can be added to a multicast group, by arranging, multicast group can be avoided retransmitting or leaking discovery resembles.
Wherein, the parameter information of the multicast packets described in literary composition all can comprise: the first data head and the second data head, wherein, the first data head is for meaning to be split by multicast packet the quantity of the multicast packets obtained, the second data head is for meaning the sequence of current multicast packets in a plurality of multicast packets, and the recipient can judge whether a plurality of multicast packets finish receiving according to the first data head, further, the recipient can correctly sort a plurality of data according to the second data head.
As shown in Figure 3, in actual applications, according to the sending method of the multicast packet of the embodiment of the present invention, can be applied as following steps:
Main thread is carried out the initialization setting, carries out the buffer resource distribution, becomes to give birth to send-thread, inquiry thread, waits for thread and deletes thread, after success, data are hung over to queue, cancel other thread when needs finish, if do not finish, continue data are hung over to queue;
Wait for that thread monitors the connection that client is sent at any time, then receive, successfully with the client exchange data after, client is added to overall client's chained list, if the swap data failure, main thread is carried out and is cancelled other thread;
Whether send-thread judges in queue data, if nothing, dormancy, if having, obtain the multicast buffer memory, gets queuing data and copy in the multicast buffer memory, then the multicast buffer memory is held in to the QP transmit queue;
The inquiry thread judge whether data send successfully, if success continue to inquire about, if do not send successfully dormancy;
Delete thread and judge whether client disconnects, the client if it is this disconnected is deleted from overall client chained list, if not, deletes the thread dormancy.
According to embodiments of the invention, provide a kind of management devices of server data transmission.
As shown in Figure 4, according to the management devices of the server data of embodiment of the present invention transmission, can comprise:
Sending module 43, be split as a plurality of multicast packets and add parameter information in the packet header of each multicast packets for the multicast packet that needs are sent, parameter information is used for meaning the position of current multicast packets at multicast packet, and the multicast packets after the interpolation parameter information is carried out to the multicast transmission;
Inquiry sends block of state 44, and for inquiring about the transmission state of each multicast packets, and the control sending module sends one by one to a plurality of multicast packets.
Wherein, monitoring module 42 can comprise for setting up and maintaining with communicating by letter of recipient: based on the direct distance input and output rdma protocol, set up and maintain and the communicating by letter of recipient.
Further, the parameter information of each multicast packets can comprise: the first data head and the second data head, wherein, the first data head is for meaning to be split by multicast packet the quantity of the multicast packets obtained, and the second data head is for meaning the sequence of current multicast packets in a plurality of multicast packets.
In addition, when sending module 43 carries out the multicast transmission for the multicast packets to after adding parameter information, multicast packets can be sent to switch, by switch, multicast packets be copied and the multicast packets copied is sent to receiving terminal.
As shown in Figure 5, the schematic diagram moved for server end according to an embodiment of the invention.In order to support a plurality of clients of server (, described receiving terminal herein), server end need to design by the mode of multithreading/multi-process, wherein, multithreading/multi-process mainly comprises: main thread, watcher thread, send-thread, inquiry send the state thread, delete the context thread.Below main role's purposes of describing each thread.
1, main thread.The user sets some configuration parameters, establishment watcher thread, send-thread, the inquiry thread of RDMA Lib and deletes the context thread by main thread, allocation buffer, and do the necessary cleaning work of exiting.Wherein, configuration parameter mainly comprises port, the IB(InfiniBand of watcher thread) implementor name and port numbers, the length of every frame, the number of frame buffer zone, the IP address of multicast group, IP address of local IB card etc.In addition, after main thread generates other thread, each thread operates independently of one another.
2, watcher thread.Connect thread also referred to as waiting for.Set up and initially be connected for the client with initiating at any time connection request, both sides exchange the necessary information of RDMA communication.Due to before the RDMA intercommunication, must exchange identity and authentication information each other by another communication port, otherwise RDMA connects and can't set up.Wherein, another communication port is often referred to the socket channel, and this paper also adopts the socket channel.
3, send-thread.For every frame data are sent to client, and be written to storage accelerator card (SysCache) equipment.The main task of this thread is that user's multicast packet (can be 10MB) is split into to parcel (preferably, the length of single parcel can be 2KB), and carries out the wrapper head, then sends to switch/router.Copied by switch/router, and be forwarded to each client in multicast group.
Wherein, between watcher thread and send-thread, comprise the asynchronous transmission data step, for multicast packet is joined the team, then data go out except after row, entering send-thread.
4, inquiry transmission state thread.After each frame data is sent to switch, whether inquiry transmission state thread sends successfully for data query, if send successfully, can carry out the transmission of next frame data, otherwise, need to wait for that this Frame sends successfully.Adopt two parallel thread execution, that is, by separating of inquiry transmission state thread and send-thread, can improve the speed that data send.
5, delete the context thread.At server end, the necessary information (also can be called contextual information) of coming maintain customer end and server to carry out RDMA to communicate by letter by chained list, send heartbeat for RDMA Write operation timing.After client disconnects, the heartbeat that server sends to client can not get responding, and server can assert that this client has disconnected communication connection, and therefore, server can be deleted these information from chained list, and discharges local resource.Owing to before the operation chained list, needing to obtain lock, and it is more consuming time to discharge local resource, therefore, in practical operation, server end has also safeguarded that removes a chained list, delete from activating chained list for linking the contextual information that disconnects client (also can be described as the client of dying), put into and remove chained list.Deleting the context thread is mainly to be responsible for timing shift detection, except in chained list, whether element is arranged, if having, deletes and removes the element in chained list, discharge and disconnect local IB resource corresponding to client, otherwise, after sleep, continue to detect.
According to embodiments of the invention, provide a kind of method of reseptance of multicast packet.
As shown in Figure 6, according to the method for reseptance of the embodiment of the present invention, can comprise:
Step S601, receive in advance and split by multicast packet a plurality of multicast packets that obtain;
Step S603, determine the position of current multicast packets in multicast packet according to the parameter information in packet header of each multicast packets, further, the parameter information of each multicast packets can comprise: the first data head and the second data head, wherein, the first data head is for meaning to be split by multicast packet the quantity of the multicast packets obtained, and the second data head is for meaning the sequence of current multicast packets in a plurality of multicast packets;
Step S605, according to the position of each multicast packets, this multicast packets is added to receiving queue, and, after the first data head and the second data head need to being peeled off out from each multicast packets, a plurality of multicast packets after the position encapsulation according to multicast packets in a plurality of multicast packets is peeled off.
In actual applications, can adopt CS(Client-Server according to the method for sending and receiving of the multicast packet of the embodiment of the present invention) model support Single-Server and multi-client communicate by letter realize, and, can adopt different development platforms the server and client side, for example, server end is used (SuSE) Linux OS, and client can adopt the Windows system, also can use more known other operating system of computer realm.
As shown in Figure 7, in actual applications, according to the method for reseptance of the multicast packet of the embodiment of the present invention, can be applied as following steps:
Thereby main thread and server end carry out socket communicates by letter and carries out necessary information interaction, then carry out that initialization setting, buffer resource distribute, the generated query thread and, obtain data from queue, whether inquiry is obtained data and is finished, if finish close other thread, if do not finish continue to obtain data from queue.
Whether receiving thread inquiry this locality is free buffer memory, if otherwise buffer memory is suspended to the QP receiving queue, whether if having dormancy 1ms, then continuing inquiry this locality is free buffer memory.
The generated query thread obtains user cache, judging whether to receive data, if the data of not receiving, dormancy 1ms, then continue to have judged whether a frame data restructuring, if the data of receiving, copy user cache, judged whether a frame data restructuring, if do not complete a frame data restructuring, by continuing to receive data, if complete a frame data restructuring, data are hung into queue.
According to embodiments of the invention, provide a kind of management devices of receiving terminal transfer of data.
As shown in Figure 8, according to the management devices of the embodiment of the present invention, can comprise:
Main control module 81, for receiving terminal is being carried out in initialized situation, configuration receiver module, enquiry module;
Receiver module 82, split by multicast packet a plurality of multicast packets that obtain for receiving in advance, determine the position of current multicast packets in multicast packet according to the parameter information in packet header of each multicast packets, and according to the position of each multicast packets, this multicast packets is added to receiving queue;
Enquiry module 83, in the situation that inquire receiving queue, having data, submit the data in receiving queue to.
Wherein, the parameter information of each multicast packets can comprise:
The first data head and the second data head, wherein, the first data head is for meaning to be split by multicast packet the quantity of the multicast packets obtained, the second data head is for meaning the sequence of current multicast packets in a plurality of multicast packets.
The client-side program overall framework as shown in Figure 5.In order to reach the performance requirement of low delay, client has also adopted the design of multithreading.
As shown in Figure 9, be the schematic diagram of server end receiving step according to an embodiment of the invention.In order to reach the performance requirement of low delay, client has also adopted the design of multithreading.The Role delineation of thread is as follows:
1, main thread.Set up TCP(Transmission Control Protocol transmission control protocol by socket with server end) be connected and carry out necessary data communication with server, obtain control information and the multicast group information of the multicast packet of server transmission, then distribute local IB resource and buffering area.Finally create receiving thread and inquiry thread, and do the necessary cleaning work of exiting.
2, receiving thread.Reception is from the multicast packet of server, obtains in packet header about after Data Position information, after packet header is weeded out, copies in the data receiver buffer memory of allocating in advance.If this buffer memory is filled full, add in receiving queue, wait for next step operation.
Wherein, after connecting, after depositing receiving thread, have the asynchronous data step of obtaining, receiving thread except after team, for multicast packet is carried out to asynchronous reception, thereby takes out queue by data, carries out timeout treatment simultaneously.
3, inquiry thread.This thread can periodically be waken up, checks on receiving queue and whether data is arranged, if data are arranged, data are submitted to the user; Otherwise continuation can enter sleep state, wait for and being waken up next time.
As shown in figure 10, be the schematic diagram of the CS model according to the embodiment of the present invention, wherein, the CS model is divided into server end, switch end and customer side.
The operation of server end comprises: will wait data to be sent (being said multicast packet in literary composition) to carry out the fractionation of packet, can multicast packets be divided into and wrap for No. 0 to No. N bag according to the size requirements of multicast packets, parcel after fractionation (is said multicast packets in literary composition,) can meet the size requirements of multicast packets, for example, in one embodiment, the length of multicast packets can be less than 2K; In addition, and calculate the total value of a multicast packets quantity according to the sum of multicast packets, and using this total value serializing as the first data head, the recipient can judge whether the multicast packets that a multicast packet comprises finishes receiving fully according to total value, utilize total value to carry out packet loss detection, can prevent that leak-stopping sends out multicast packets; Then a plurality of multicast packets after splitting are calculated to also serializing again and draw the second data head, this second data head can mean the position of each multicast packets in a plurality of multicast packets, and the main purpose that arranges of the second data head is in order to facilitate client can carry out according to this sequence number the restructuring of data after receiving multicast packets; Finally by two data heads with by rdma protocol, write in the buffer memory of switch after the corresponding multicast packets split is merged.The data head obtained by the total value serializing in Fig. 4 is the first data head, and the data head obtained by the sequence number serializing is the second data head.
The operation of switch comprises: the packet of receiving is copied, then be sent in the buffer memory of each client of multicast group by InfiniBand, set up in advance multicast group, a plurality of clients that need the same multicast data are added to a multicast group.
The operation of client comprises: after receiving multicast packets, by peeling off data packet head (i.e. two data heads), after determining the position of multicast packets in receiving a plurality of multicast packets by two data heads, will split after a plurality of multicast packets that obtain combine and return to client by a multicast packet.
Further, can transfer of data be divided into to server and client side's two parts according to the embodiment of the present invention, server is responsible for data are mail to registered and add all clients of multicast group; Client is responsible for receiving these data.
In sum, by means of technique scheme of the present invention, the present invention is by splitting multicast packet, then the multicast packets sent after splitting can provide efficient cast communication mechanism, and the minimizing data delay, in addition, add the packet header with definite each multicast packets position in multicast packet by giving the multicast packets after splitting, be conducive to integrality and order that multicast packets sends, further improve communication efficiency.In addition, in the present invention, adopt the multicasting technology of InfiniBand that the communication mechanism of a kind of efficient Single-Server, multi-client is provided, almost can be by data simultaneously at a high speed send to multi-client, and time delay can be with the growth of client linear increment.
The foregoing is only preferred embodiment of the present invention, in order to limit the present invention, within the spirit and principles in the present invention not all, any modification of doing, be equal to replacement, improvement etc., within all should being included in protection scope of the present invention.
Claims (13)
1. the sending method of a multicast packet, is characterized in that, comprising:
The described multicast packet that needs are sent is split as a plurality of multicast packets;
Parameter information is added in packet header in each multicast packets, and described parameter information is for meaning the position of current multicast packets at described multicast packet;
Multicast packets after the interpolation parameter information is carried out to the multicast transmission.
2. sending method according to claim 1, is characterized in that, the parameter information of each multicast packets comprises:
The first data head and the second data head, wherein, described the first data head is for meaning to be split by described multicast packet the quantity of the multicast packets obtained, described the second data head is for meaning the sequence of current multicast packets in described a plurality of multicast packets.
3. sending method according to claim 1, is characterized in that, the length of described multicast packets is less than 2K.
4. sending method according to claim 1, is characterized in that, the multicast packets after the interpolation parameter information carried out to the multicast transmission and comprise:
Send the multicast packets after encapsulating based on the direct distance input and output rdma protocol.
5. sending method according to claim 1, is characterized in that, the multicast packets after the interpolation parameter information carried out to the multicast transmission and comprise:
Multicast packets is sent to switch, by described switch, multicast packets is copied and the multicast packets copied is sent to receiving terminal.
6. the method for reseptance of a multicast packet, is characterized in that, comprising:
Receive in advance and split by described multicast packet a plurality of multicast packets that obtain;
Determine the position of current multicast packets in described multicast packet according to the parameter information in packet header of each multicast packets;
According to the position of each multicast packets, this multicast packets is added to receiving queue.
7. sending method according to claim 6, is characterized in that, the parameter information of each multicast packets comprises:
The first data head and the second data head, wherein, described the first data head is for meaning to be split by described multicast packet the quantity of the multicast packets obtained, described the second data head is for meaning the sequence of current multicast packets in described a plurality of multicast packets.
8. the management devices of a server data transmission, is characterized in that, comprising:
Main control module, for server is being carried out in initialized situation, module, sending module, inquiry transmission block of state are monitored in configuration;
Described monitoring module is for setting up and maintaining and the communicating by letter of recipient;
Described sending module will be for needing the described multicast packet sent to be split as a plurality of multicast packets, and add parameter information in the packet header of each multicast packets, described parameter information is used for meaning the position of current multicast packets at described multicast packet, and the multicast packets after the interpolation parameter information is carried out to the multicast transmission;
Described inquiry sends block of state for inquiring about the transmission state of each multicast packets, and controls described sending module described a plurality of multicast packets are sent one by one.
9. management devices according to claim 8, is characterized in that, described monitoring module comprises for setting up and maintaining with communicating by letter of recipient:
Set up and maintain and the communicating by letter of recipient based on the direct distance input and output rdma protocol.
10. management devices according to claim 8, is characterized in that, the parameter information of each multicast packets comprises:
The first data head and the second data head, wherein, described the first data head is for meaning to be split by described multicast packet the quantity of the multicast packets obtained, described the second data head is for meaning the sequence of current multicast packets in described a plurality of multicast packets.
11. management devices according to claim 8, is characterized in that, described sending module carries out the multicast transmission for the multicast packets to after the interpolation parameter information and comprises:
Multicast packets is sent to switch, by described switch, multicast packets is copied and the multicast packets copied is sent to receiving terminal.
12. the management devices of a receiving terminal transfer of data, is characterized in that, comprising:
Main control module, for receiving terminal is being carried out in initialized situation, configuration receiver module, enquiry module;
Described receiver module splits by described multicast packet a plurality of multicast packets that obtain for receiving in advance, determine the position of current multicast packets in described multicast packet according to the parameter information in packet header of each multicast packets, and according to the position of each multicast packets, this multicast packets is added to receiving queue;
Described enquiry module, in the situation that inquire described receiving queue, having data, is submitted the data in described receiving queue to.
13. management devices according to claim 12, is characterized in that, the parameter information of each multicast packets comprises:
The first data head and the second data head, wherein, described the first data head is for meaning to be split by described multicast packet the quantity of the multicast packets obtained, described the second data head is for meaning the sequence of current multicast packets in described a plurality of multicast packets.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013103669357A CN103441937A (en) | 2013-08-21 | 2013-08-21 | Sending method and receiving method of multicast data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013103669357A CN103441937A (en) | 2013-08-21 | 2013-08-21 | Sending method and receiving method of multicast data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103441937A true CN103441937A (en) | 2013-12-11 |
Family
ID=49695603
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2013103669357A Pending CN103441937A (en) | 2013-08-21 | 2013-08-21 | Sending method and receiving method of multicast data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103441937A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105337888A (en) * | 2015-11-18 | 2016-02-17 | 华为技术有限公司 | Multinuclear forwarding-based load balancing method and device, and virtual switch |
| CN106021399A (en) * | 2016-05-12 | 2016-10-12 | 网易（杭州）网络有限公司 | Query request message processing method and apparatus |
| CN107508828A (en) * | 2017-09-18 | 2017-12-22 | 南京斯坦德云科技股份有限公司 | A kind of very-long-range data interaction system and method |
| CN108234327A (en) * | 2016-12-13 | 2018-06-29 | 英业达科技有限公司 | Data transmission system, data receiver method and data transmission method |
| WO2019033891A1 (en) * | 2017-08-14 | 2019-02-21 | 华为技术有限公司 | Rdma-based network traffic determination method and device |
| WO2019140556A1 (en) * | 2018-01-16 | 2019-07-25 | 华为技术有限公司 | Message transmission method and apparatus |
| CN110132222A (en) * | 2019-04-23 | 2019-08-16 | 广州南方卫星导航仪器有限公司 | The AM automatic monitoring of base station operating status and alarm method, electronic equipment, medium |
| CN114944966A (en) * | 2022-04-20 | 2022-08-26 | 南方电网深圳数字电网研究院有限公司 | RDMA multicast-based data transmission method and system |
| WO2023135674A1 (en) * | 2022-01-12 | 2023-07-20 | 日本電信電話株式会社 | Processing system, processing device, processing method, and program |
| CN113645662B (en) * | 2021-08-17 | 2023-08-18 | 中国联合网络通信集团有限公司 | Service processing method and device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050120140A1 (en) * | 2002-03-14 | 2005-06-02 | Bodlaender Maarten P. | Method of and system for multi-patch communication |
| CN101512526A (en) * | 2006-09-06 | 2009-08-19 | 微软公司 | Dynamic fragment mapping |
| CN101695202A (en) * | 2009-10-22 | 2010-04-14 | 中兴通讯股份有限公司 | Method and system for transmitting big accessory |
| CN102983884A (en) * | 2011-09-05 | 2013-03-20 | 国民技术股份有限公司 | A method for transmitting data via a magnetic channel |
-
2013
- 2013-08-21 CN CN2013103669357A patent/CN103441937A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050120140A1 (en) * | 2002-03-14 | 2005-06-02 | Bodlaender Maarten P. | Method of and system for multi-patch communication |
| CN101512526A (en) * | 2006-09-06 | 2009-08-19 | 微软公司 | Dynamic fragment mapping |
| CN101695202A (en) * | 2009-10-22 | 2010-04-14 | 中兴通讯股份有限公司 | Method and system for transmitting big accessory |
| CN102983884A (en) * | 2011-09-05 | 2013-03-20 | 国民技术股份有限公司 | A method for transmitting data via a magnetic channel |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105337888B (en) * | 2015-11-18 | 2018-12-07 | 华为技术有限公司 | Load-balancing method, device and virtual switch based on multicore forwarding |
| CN105337888A (en) * | 2015-11-18 | 2016-02-17 | 华为技术有限公司 | Multinuclear forwarding-based load balancing method and device, and virtual switch |
| CN106021399A (en) * | 2016-05-12 | 2016-10-12 | 网易（杭州）网络有限公司 | Query request message processing method and apparatus |
| CN106021399B (en) * | 2016-05-12 | 2019-12-06 | 网易（杭州）网络有限公司 | method and device for processing query request message |
| CN108234327A (en) * | 2016-12-13 | 2018-06-29 | 英业达科技有限公司 | Data transmission system, data receiver method and data transmission method |
| WO2019033891A1 (en) * | 2017-08-14 | 2019-02-21 | 华为技术有限公司 | Rdma-based network traffic determination method and device |
| CN107508828B (en) * | 2017-09-18 | 2019-10-18 | 南京斯坦德云科技股份有限公司 | A kind of very-long-range data interaction system and method |
| CN107508828A (en) * | 2017-09-18 | 2017-12-22 | 南京斯坦德云科技股份有限公司 | A kind of very-long-range data interaction system and method |
| WO2019140556A1 (en) * | 2018-01-16 | 2019-07-25 | 华为技术有限公司 | Message transmission method and apparatus |
| US11716409B2 (en) | 2018-01-16 | 2023-08-01 | Huawei Technologies Co., Ltd. | Packet transmission method and apparatus |
| CN110132222A (en) * | 2019-04-23 | 2019-08-16 | 广州南方卫星导航仪器有限公司 | The AM automatic monitoring of base station operating status and alarm method, electronic equipment, medium |
| CN113645662B (en) * | 2021-08-17 | 2023-08-18 | 中国联合网络通信集团有限公司 | Service processing method and device |
| WO2023135674A1 (en) * | 2022-01-12 | 2023-07-20 | 日本電信電話株式会社 | Processing system, processing device, processing method, and program |
| CN114944966A (en) * | 2022-04-20 | 2022-08-26 | 南方电网深圳数字电网研究院有限公司 | RDMA multicast-based data transmission method and system |
| CN114944966B (en) * | 2022-04-20 | 2024-04-19 | 南方电网数字平台科技(广东)有限公司 | RDMA multicast-based data transmission method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103441937A (en) | Sending method and receiving method of multicast data | |
| JP6697556B2 (en) | Multipath transfer design | |
| CN105516191B (en) | System based on the FPGA 10,000,000,000 net Transmission Control Protocol unloading engine TOE realized | |
| US9503394B2 (en) | Clustered dispersion of resource use in shared computing environments | |
| CN103370911B (en) | Streaming communication system | |
| US9319347B1 (en) | Deadlock-resistant fabric tree replication in a network device | |
| EP2356775B1 (en) | Central controller for coordinating multicast message transmissions in distributed virtual network switch environment | |
| US10148744B2 (en) | Random next iteration for data update management | |
| US8392575B1 (en) | Clustered device dispersion in a multi-tenant environment | |
| CN103236949B (en) | Monitoring method, device and the system of a kind of server cluster | |
| US8370496B1 (en) | Reducing average link bandwidth in an oversubscribed environment | |
| US8949472B2 (en) | Data affinity based scheme for mapping connections to CPUs in I/O adapter | |
| EP2003823B1 (en) | Autonegotiation over an interface for which no autonegotiation standard exists | |
| US20110238839A1 (en) | Network intrusion detection apparatus | |
| US8539094B1 (en) | Ordered iteration for data update management | |
| CN110177118A (en) | A kind of RPC communication method based on RDMA | |
| CN104821887A (en) | Device and Method for Packet Processing with Memories Having Different Latencies | |
| Gibson et al. | Aquila: A unified, low-latency fabric for datacenter networks | |
| RU2011140978A (en) | METHOD AND DEVICE OF I / O ROUTING AND MAP | |
| EP2355404A1 (en) | Method, device and system for forwarding multicast packets | |
| EP2680536A1 (en) | Methods and apparatus for providing services in a distributed switch | |
| EP1570369A2 (en) | An apparatus and method for receive transport protocol termination | |
| CN105162623B (en) | A kind of cloud processing method and Cloud Server | |
| US20050169309A1 (en) | System and method for vertical perimeter protection | |
| US9203895B1 (en) | System and method for lateral connection between interface devices with bypass of external network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20131211 |