WO2007083687A1 - 通信方法、通信システム、ノードおよびプログラム - Google Patents
通信方法、通信システム、ノードおよびプログラム Download PDFInfo
- Publication number
- WO2007083687A1 WO2007083687A1 PCT/JP2007/050660 JP2007050660W WO2007083687A1 WO 2007083687 A1 WO2007083687 A1 WO 2007083687A1 JP 2007050660 W JP2007050660 W JP 2007050660W WO 2007083687 A1 WO2007083687 A1 WO 2007083687A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- packet
- control information
- leader
- block
- data
- Prior art date
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L69/00—Network arrangements, protocols or services independent of the application payload and not provided for in the other groups of this subclass
- H04L69/14—Multichannel or multilink protocols
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W28/00—Network traffic management; Network resource management
- H04W28/02—Traffic management, e.g. flow control or congestion control
- H04W28/06—Optimizing the usage of the radio link, e.g. header compression, information sizing, discarding information
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L69/00—Network arrangements, protocols or services independent of the application payload and not provided for in the other groups of this subclass
- H04L69/28—Timers or timing mechanisms used in protocols
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W28/00—Network traffic management; Network resource management
- H04W28/02—Traffic management, e.g. flow control or congestion control
Definitions
- the present invention relates to a packet communication technique and a multiplexed communication technique.
- packet data communication digital data is stored in a packet called a packet and transferred over a network.
- the packet includes a data body to be transferred and a header containing information used for transfer control on the network.
- the header of the upper layer is included in the data body at the lower layer.
- Each node in the network that performs packet data communication is equipped with a transfer protocol function that decodes the header of each input packet, edits it as necessary, and sends the same data to the next node.
- the packet transfer protocol is an asynchronous protocol, and each node transfers different packets on the same network at an arbitrary time. Therefore, packet transfer can be performed at any rate as long as the performance of the sending node allows.
- TCP which is the fourth layer protocol in the OSI layer model
- Figure 16 shows an example in which TCP packets are transmitted to the receiving node as well as the transmitting node when IP (Internet 'protocol) is used in the third layer.
- TCP is a sliding window method to avoid packet delay and loss due to excessive load on the end-to-end connection provided by the third layer and below, and excessive bandwidth pressure on other sessions. (See Non-Patent Document 2 for details).
- the network is constant. By holding packets that are less than the amount (for example, the amount that can be transmitted before the arrival of an ACK corresponding to the transmitted packet), avoiding a transfer failure due to overload, the bandwidth provided by the lower layer connection Realize effective utilization.
- the upper limit of the amount of data held in the network is called a window size.
- a window size proportional to the product of the bandwidth and the round-trip delay in order to effectively use the bandwidth. For example, if the processing delay in the lower layer forwarding node can be ignored and the physical link force on the path can be ignored only in a short-distance wired line, the packet transmission delay is inversely proportional to the bandwidth.
- the window size can be the same regardless of the bandwidth.
- a bucket transfer session is set up between the two nodes for each path, and the load is distributed to each session and transferred in parallel to communicate between the two nodes.
- a demultiplexing technique for wideband transmission For example, a method has been proposed in which a TCP session is set up on each path and transferred in parallel, and the transfer between two nodes is wider than when using a single path (see Non-Patent Document 5 for details). .
- MIM monitors the speed and delay of each route, but the delay is large! /, And the route including the wireless link has a delay in the feedback of the monitoring result, so refer to the past transmission history that is held. Therefore, the delay when the current packet is transmitted is predicted from the history after the time when the feedback result becomes valid. By performing flow control based on the estimated delay value for each path, jitter is reduced when the paths are multiplexed while effectively using the bandwidth.
- a packet order control function is implemented so that the order of packets forwarded via different routes is not reversed during forwarding to downstream nodes.
- the sending node assigns a sequence number to each packet for each flow for which the order is to be stored, and the receiving node arranges the packet order correctly using the received sequence number and transfers it downstream. This is realized.
- Non-Patent Document 1 RFC793
- Non-Patent Document 2 Mastering TCP / IP, by Phillip Miller, Ohm Corp. Development Department (1998)
- Non-Patent Document 4 Kakuzawa et al., ⁇ Realization of arbitration between parallel TCP streams in long distance and high bandwidth communication '', SACSIS 2004.
- Non-Patent Document 5 Maki, Hasegawa, Murata, Murase, ⁇ Performance Analysis and Evaluation of TCP Overlay Network '', IEICE Technical Report IN04-96 (2004).
- Non-Patent Document 7 Dovrolis, Ramanathan, and Moore, “What Do PacketDispersion Tech niques Measuere ?, "IEEE INFOCOM 2001)
- Non-Patent Document 8 Ono et al., “Mobile Internet (3) —Retransmission Control Method 1”, 2004 IEICE General Conference, Paper B-5-165 (2004).
- Non-Patent Literature 9 L.S.Brakmo and L.L.Peterson, TCP Vegas: ⁇ to End ongestion Avoidance on a Global Internet, IEEE Journal of Selected Areas in Communications, Vol.13, No.8, pl465 (1995).
- Non-Patent Document 10 Okanoue et al., “Mobile Internet (1) Basic Concept and System Configuration”, 2004 IEICE General Conference, Paper B-5-163 (2004).
- Non-Patent Document 11 Nakata et al., “Mobile Internet (2) Flow Control Method 1”, 2004 IEICE General Conference, Paper B-5-164 (2004).
- TCP which has been widely used as a flow control method for conventional packet transfer sessions, uses multiple sessions in order to effectively use the bandwidth in a route having a delay / bandwidth product of a certain degree or more. It is necessary to provide a mechanism that keeps the delay 'bandwidth product that each session should keep by extending
- this requires complicated control such as load distribution among the sessions.
- the delay increases in inverse proportion to the bandwidth.
- the bandwidth of the route changes when the route bandwidth increases as the load on the route increases.
- Use efficiency can be kept high.
- the bandwidth of the path changes in a smaller direction, the higher the load on the path, the greater the transfer delay and the more time it takes to feedback, so the time to perform load distribution based on incorrect state recognition increases. Bigger. If an incorrect state recognition occurs in some of the multiple paths, the jitter between the multiple paths will increase, and the packet discard rate or retransmission rate will also increase in the overall communication that combines the multiple paths. Therefore, if the change in the bandwidth of the path is increasing, if the window size is increased and the packet transmission rate is increased, there is a trade-off problem that adversely affects the downward trend.
- these flow control protocols have a problem in that the bandwidth that can be used for data transfer is high because the occupation rate of the header is high with respect to the data size of the entire packet.
- the header compression method can be applied.
- it is necessary for the sending side and the receiving side to share the state variables necessary for compression and decompression.
- the protocol becomes complicated.
- the transmitting node determines control information to be included in the header for each transmission packet, and the receiving node performs header reading and analysis for each packet reception.
- the processing load of the node increases as the transmission / reception data rate increases, and the processing capability of the node determines the maximum transfer rate of the node in a specific protocol.
- the maximum transfer rate of the node is lower than the bandwidth of the route, and the flow control processing speed that is originally introduced for effective use of the bandwidth of the route is a bottle. As a result, there is a problem that the utilization efficiency of the bandwidth of the route is lowered.
- an object of the present invention is to provide a communication technique that reduces the load caused by transmission of control information related to a data packet and increases communication efficiency.
- an object of the present invention is to use a low-delay route or a highly reliable route for a leader packet storing control information when performing packet communication between nodes connected by a plurality of routes. It is to provide a technology that can suppress loss and delay of signaling information by transferring.
- the packet transmission is controlled based on the control information of the received leader packet.
- a second invention for solving the above-described problems is the above-mentioned first invention
- the control information is flow control information.
- a third invention for solving the above-described problem is the above-described first invention
- the control information is retransmission control information.
- a fourth invention for solving the above-described problem is the above first invention,
- the control information is routing information.
- a fifth invention for solving the above-mentioned problems is the above-mentioned first invention
- the control information is route information.
- a sixth invention for solving the above-mentioned problems is any one of the first to fifth inventions.
- the control information is control information common to the data packets.
- a seventh invention for solving the above-mentioned problems is any one of the first to sixth inventions described above.
- the data packet includes identification information for uniquely identifying a leader packet.
- An eighth invention for solving the above-described problems is any one of the first to seventh inventions.
- the reader bucket includes identification information for uniquely identifying a data packet in which control information is collected.
- a ninth invention for solving the above-mentioned problems is any one of the first to eighth inventions.
- the plurality of data packets or leader packets are transmitted / received via a plurality of lines.
- a tenth invention for solving the above-mentioned problem is the above-mentioned ninth invention.
- the leader packet is transmitted using the fastest line among the plurality of lines.
- a line on which the leader packet or the data packet is transmitted is selected based on the control information.
- a twelfth invention for solving the above-described problems is any one of the first to eleventh inventions,
- the amount of block data to be transmitted based on the control information contained in the received reader packet It is characterized by determining.
- a thirteenth invention for solving the above-mentioned problems is any one of the first to twelfth inventions,
- the transmission time of the block to be transmitted is determined based on the control information included in the received reader packet.
- a fourteenth invention for solving the above-described problems is any one of the first to thirteenth inventions,
- the speed of the communication path is estimated from the reception time of multiple data packets belonging to the same block.
- a packet block creating means for handling a plurality of data packets and a leader packet as one block, and collecting control information on the plurality of data packets in the leader packet;
- a sixteenth invention for solving the above-described problem is the fifteenth invention, in which
- the control information is flow control information.
- the control information is retransmission control information.
- the control information is routing information.
- a nineteenth aspect of the invention for solving the above-described problems is the fifteenth aspect of the invention.
- the control information is route information.
- a twentieth invention for solving the above-mentioned problems is any one of the fifteenth to nineteenth inventions,
- the packet block creating means collects control information common to the data packets in the leader packet.
- the data packet includes identification information for uniquely identifying a leader packet.
- a twenty-second invention for solving the above-described problems is any one of the fifteenth to twenty-first inventions,
- the reader bucket includes identification information for uniquely identifying a data packet in which control information is collected.
- a twenty-third invention for solving the above-described problems is any one of the fifteenth to twenty-second inventions.
- It has transmission means for transmitting the plurality of data packets or leader packets via a plurality of lines.
- a twenty-fourth aspect of the present invention for solving the above-described problem is the above twenty-third aspect of the invention.
- the scheduling means controls to transmit the leader packet using the fastest line among the plurality of lines.
- the scheduling means selects a line for transmitting the leader packet or the data packet based on control information.
- a twenty-sixth aspect of the present invention for solving the above-mentioned problems is any one of the fifteenth to twenty-fifth aspects of the invention.
- the scheduling means determines a data amount of a block to be transmitted based on control information included in the received leader packet.
- a twenty-seventh aspect of the present invention for solving the above-mentioned problems is any one of the fifteenth to twenty-sixth aspects of the invention.
- the scheduling means determines a transmission time of a block to be transmitted based on control information included in the received leader packet.
- a twenty-eighth aspect of the present invention for solving the above-described problems is any one of the fifteenth to twenty-seventh aspects of the present invention, It has a packet analysis unit that estimates the speed of a communication path from the reception time of a plurality of data packets belonging to the same block.
- a twenty-ninth invention for solving the above-mentioned problems is
- a packet block creating means for handling a plurality of data packets and a leader packet as one block, and collecting control information on the plurality of data packets in the leader packet;
- a thirtieth invention for solving the above-described problem is the above-mentioned twenty-ninth invention.
- the packet block creating means collects control information common to the data packets in the leader packet.
- a thirty-first invention for solving the above-mentioned problems is characterized in that, in the above-mentioned twenty-ninth or thirtieth invention, identification information for uniquely identifying a leader packet is included in the data packet.
- a thirty-second invention for solving the above-described problems is any one of the twenty-ninth to thirty-first inventions,
- the reader bucket includes identification information for uniquely identifying a data packet in which control information is collected.
- a thirty-third invention for solving the above-mentioned problems is as described in any one of the twenty-ninth to thirty-second inventions,
- It has transmission means for transmitting the plurality of data packets or leader packets via a plurality of lines.
- the thirty-fourth invention for solving the above-mentioned problems is the above-mentioned thirty-third invention.
- the scheduling means controls to transmit the leader packet using the fastest line among the plurality of lines.
- the scheduling means is a control grouped in a leader packet received from another node. Based on the control information, a line for transmitting the leader packet or the data packet is selected.
- a thirty-sixth aspect of the present invention for solving the above-described problems is the invention according to any one of the twenty-ninth to thirty-fifth aspects of the invention.
- the scheduling means determines a data amount of a block to be transmitted based on control information included in a leader packet received from another node.
- a thirty-seventh aspect of the present invention for solving the above-described problems is the invention according to any one of the twenty-ninth to thirty-sixth aspects of the invention.
- the scheduling means determines a transmission time of a block to be transmitted based on control information included in a leader packet received from another node.
- a thirty-eighth aspect of the present invention for solving the above-described problems is any one of the twenty-ninth to thirty-seventh aspects of the invention.
- It has a packet analysis unit that estimates the speed of a communication path from the reception time of a plurality of data packets belonging to the same block.
- a packet block creation process in which a plurality of data packets and a leader packet are treated as one block, and control information related to the plurality of data packets is collected in the leader packet;
- a plurality of data packets and a reader packet are made into one block, the control information of the plurality of data packets is collected into the reader packet, and based on the control information of the reader packet, Control the transmission of packets. For this reason, since the control information can be known at the node at an early stage, it becomes possible to perform signaling at high speed.
- the present invention provides a packet block between two nodes that can select one or more routes. By feeding back the results of route state estimation based on the received sequence, flow control and load distribution based on more accurate route state estimation than the conventional example can be realized.
- the processing load is reduced compared to the conventional example in which the flow control and the route selection process are performed for each transmission packet. .
- the present invention aggregates the flow control information, load distribution information, ARQ information, and order control information of each route into a leader packet, and transmits it as the head of the block from the lowest delay route or the highly reliable route. By doing so, the loss and delay of signaling information can be suppressed, and the trade-off between bandwidth utilization efficiency when the path bandwidth is increasing and control tracking when the path bandwidth is decreasing can be mitigated.
- the receiving node can predict the packet reception sequence in the block based on the information of the reader, if it is different from the prediction, it is determined that the communication is abnormal, thereby early detection of the communication abnormality. Is possible.
- the transfer of data in the sequence control buffer is not held for a longer time than necessary, and end-end jitter is improved.
- the present invention achieves header information reduction with a simpler implementation than the conventional header compression example that requires state information sharing between the transmission side and the reception side.
- FIG. 1 is a system configuration diagram showing a first embodiment of the present invention.
- FIG. 2 is a configuration diagram of a leader packet according to the present invention.
- FIG. 3 is a configuration diagram of a data packet according to the present invention.
- FIG. 4 is an example in which a flow control algorithm (PAC scheduler) is applied to the first embodiment of the present invention.
- PAC scheduler a flow control algorithm
- FIG. 5 is a flowchart followed by the scheduling unit.
- FIG. 6 is a flowchart followed by the packet creation unit.
- FIG. 7 is a system configuration diagram showing a second embodiment of the present invention.
- FIG. 8 is an example in which a flow control algorithm (PAC scheduler) is applied to the second embodiment of the present invention.
- FIG. 9 is an example in which the feedback period of the second embodiment of the present invention can be shortened.
- FIG. 10 is an example in which the speed estimation accuracy of the second embodiment of the present invention is kept constant.
- FIG. 11 is a diagram for explaining a third embodiment of the present invention.
- FIG. 12 is a diagram for explaining a fourth embodiment of the present invention.
- FIG. 13 is a diagram for explaining a fifth embodiment of the present invention.
- FIG. 14 is a diagram for explaining the outline of the present invention.
- FIG. 15 is a diagram for explaining the outline of the present invention (when there are three or more data packets).
- FIG. 16 is a diagram for explaining the outline of a conventional invention.
- FIG. 14 shows a state in which the packet transfer according to the present invention is performed from the transmission node 001 to the reception node 002.
- TCP which is the fourth layer protocol of the prior art, as shown in FIG.
- the packet is composed of the IP header, TCP header, and payload in that order from the top.
- TCP when multiple packets are transmitted simultaneously or within a short time that the control information does not change, the TCP header control information assigned to each packet is all equal.
- the control information when a plurality of packets are transmitted simultaneously or within a short time such that the control information does not change, the control information is stored only in the header for the first leader packet 003, and the present invention
- An IP header indicating the destination node is attached to the reader packet header as in the conventional case, and a reader packet comprising the IP header and the reader packet header of the present invention is first transmitted.
- leader packet 003 data packets 004 and 005 including the data packet header of the present invention having a smaller amount of information than the leader packet header are transmitted.
- the control information stored in the header of each packet is stored in the header of the leading reader packet 003, and the time required to receive the packet 003 is the same as that of the conventional packet.
- the time required for receiving 0040 is shorter. Therefore, the receiving node can know the control information (for example, ACK information) earlier than the conventional packet transfer method. As a result, signaling based on the control information can be performed at high speed.
- each data packet 004a, 005a, 006a is given a fourth layer header with the same data amount that is duplicated
- control information corresponding to the overlapping portion of each packet header in the conventional example is stored together in the header of the leading leader packet 003a, and each data packet is lighter than the TCP header.
- the header is assigned as the 4th layer header.
- the information that has been duplicated in a number of packets in the past is represented as a leader packet, so that the amount of signaling information by the header is kept equal to that of TCP.
- the total header size of all packets can be reduced compared to TCP. As a result, it is possible to improve the transmission efficiency of the entire packet transmission.
- one leader packet is created for many packets.
- a further effect of sending packets is that the accuracy of path speed measurement is increased.
- Speed is a measurable force with at least two packets. The more packets, the higher the scheme.
- Another reason is that, as the number of packets included in a packet block increases, the difference between the reception times of the beginning and end of a packet related to one measurement increases, which is required by the receiving node. This is because the time resolution may be large, and if the nodes have the same time resolution, the speed system is improved by using the packet block of the present invention.
- FIG. 1 shows a first embodiment of the present invention.
- packets are transferred between the forwarding node 101 and the forwarding node 102 using the packet forwarding method of the present invention. Since this embodiment assumes a bidirectionally symmetric protocol, the configuration of the forwarding node 101 and the forwarding node 102 is the same. In the following, the power to explain the packet transfer from the forwarding node 101 to the forwarding node 102 as an example.
- the transfer node 101 also receives data to be transferred to the transfer node 102 as well as data generation means such as a user application on another node or the same node, and stores it in the buffer unit 201.
- Scheduling section 202 controls transmission based on scheduling information stored in storage section 205. For example, when the time for packet transmission and the configuration of a data packet (hereinafter referred to as a packet group) to be transmitted following the leader packet are determined as described later, and the time for packet transmission is reached, the packet block The creation unit 203 is notified of the configuration of the leader packet and the packet group. The packet block creation unit 203 notified of the configuration of the leader packet and the packet group collects the control information of the data packet into the leader packet.
- a packet group data packet
- a number of data based on the notified configuration is extracted from the buffer unit 201, and predetermined header information is added to each of the extracted data to create a packet group.
- predetermined header information is added to each of the extracted data to create a packet group.
- a set of packets transmitted at this time is hereinafter referred to as a packet block.
- the packet transmission unit 204 sequentially transfers the packets passed from the packet block creation unit 203 to the packet reception unit 206 in the transfer node 102.
- 300-1 is a schematic diagram of a packet block on the communication path transmitted from the packet transmission unit 204, and shows that a plurality of data packets 301 follow the leader packet 302.
- the packet analysis unit 207 extracts predetermined information from the packet block 300-1 as described later, performs predetermined measurement on the received packet as described later, and stores the result in the storage unit 205 as flow control information. .
- the packet data that has been processed by the analysis unit 207 is sequentially transferred to the packet transfer unit 208, and the packet transfer unit 208 receives the data received from the packet analysis unit 207 as the next transfer node or a user application in the same node, etc. To the data receiving means.
- the scheduling unit 202 in the forwarding node 102 performs the same operation as the scheduling unit 202 in the forwarding node 101, based on the scheduling information stored in the storage unit 205, and the time and packet to be transmitted.
- the county configuration is determined as described later, and when the time for packet transmission is reached, the packet block creation unit 203 is notified of the configuration of the leader packet and the packet county.
- the packet block creation unit 203 in the forwarding node 102 creates a leader packet based on the notified configuration and the flow control information stored in the storage unit 205, and together with the data stored in the nota unit 201, the packet block 300- 2 is transmitted to the packet receiving unit 206 in the forwarding node 101 via the packet transmitting unit 204. After transmission, the transmission history is stored in the storage unit 205.
- the packet analysis unit 207 in the forwarding node 101 extracts predetermined information from the packet block 300-2 and performs predetermined measurement on the received packet, The result is stored as flow control information 2
- the data is stored in 05 and the scheduling unit 202 is notified that the flow control information has been updated.
- the scheduling unit 202 similarly updates the scheduling information stored in the storage unit 205 based on the updated flow control information. Subsequent packet transmission timing and packet block configuration are determined based on the updated scheduling information.
- FIG. 2 shows an example of the structure of the leader packet used in the present invention.
- the present invention can be implemented as a protocol on any communication layer.
- the protocol is assumed to be implemented as a fourth layer protocol, and an example of the structure when the lower layer transfer protocol is IP is shown.
- the leader packet in this embodiment is included in the header for block management information and flow control information S leader packet following the leading IP header.
- the structure of the data packet is shown in Fig. 3.
- the block management information is included in the data packet header and the data body is included in the payload.
- the block management information is attribute information generated by the transmitting side node, such as the number of packets in the packet block and the priority of the packet.
- the block management information is identification information for uniquely identifying a reader packet. That is, it is information for identifying the packet block to which each data packet belongs.
- the transmitting node assigns a unique sequence number to each packet block, which is used as block management information.
- Each packet in the block is sequentially transmitted starting from the leader packet (passed from the packet block creation unit 203 to the packet transmission unit 204).
- the time stamp at the time of transmission and the sequence of packets belonging to the packet block are assumed.
- the range of numbers is used as block management information.
- the receiving node measures the delay and describes the packet train method described in non-patent literature (total data amount Z based on the difference in reception time between the last packet and the first packet).
- the bandwidth of the communication path is estimated by the above method.
- the identification information for uniquely identifying the data packet in which the control information is collected may be included in the reader bucket to secure the relationship between the leader packet and the data packet group.
- the packet analysis unit 207 in the forwarding node 102 receives the reader packet 302 of the packet block 300-1 from the packet reception unit 206, the packet analysis unit 207 receives the packet from the block management information. Extracts the range of timestamps and sequence numbers for the first block 300—1. Further, the difference between the reception time and the extracted time stamp is written in the storage unit 205 as a path delay estimation result.
- the subsequent data packet 301 is received, if the sequence number is within the sequence number range of the packet block 300-1, the data size is stored.
- the total packet size of the received data packet in the packet block is calculated by the difference between the reception time of the packet and the reception time of the reader packet.
- the divided one is written in the storage unit 205 as the estimated bandwidth of the communication path.
- the flow control information (path delay estimation result and path bandwidth estimation result) recorded in the storage unit 205 by the packet analysis unit 207 in the forwarding node 102 by the above operation is also transmitted to the forwarding node 101.
- the packet is stored in the reader packet of the packet block 300-2 to be sent to the storage unit 205 via the packet analysis unit 207 in the forwarding node 101.
- the flow control information update is notified to the scheduling unit 202, and the scheduling unit 202 that receives the notification notifies the updated flow control information.
- the scheduling information is updated using the estimation result of the path delay and the estimation result of the path bandwidth, and thereafter, a packet block is configured based on the updated scheduling information.
- the reader packet 302 transmitted from the forwarding node 102 to the forwarding node 101 includes Ack information similar to that used in TCP and the reception window size (the amount of data that can be consulted determined on the receiving node side).
- the forwarding node 101 that has received this leader packet 302 performs transmission processing to the forwarding node 102 in the following procedure.
- the packet analysis unit 207 The Ack content included in the packet 302 is recorded in the storage unit 205, and the storage unit 205 notifies the scheduling unit 202 of a new Ack reception event.
- the scheduling unit 202 can transmit the difference between the reception window size extracted from the received leader packet and the data amount of the packet transmitted after the ACKed packet extracted from the transmission history recorded in the storage unit The data amount is notified to the packet block generation unit 203.
- the packet block generation unit 203 extracts the maximum amount of data from the buffer unit so that the packet block size including the leader packet 302 does not exceed the amount of data that can be transmitted, and adds block management information to each data. Create a data packet. Also, a leader packet including Ack information and block management information is created, and a packet block having both powers is passed to the packet transmission unit 204. If the transmittable data amount is less than the predetermined lower limit, packet block creation is suspended until the next notification of the transmittable data amount from the scheduling unit 202.
- the transmission history is recorded in the storage unit 205.
- the reception window size used by the forwarding node 101 is determined from the route speed estimated when the packet analysis unit 207 of the forwarding node 102 receives the packet block from the forwarding node 101.
- the value of the reception window size is given by, for example, a product of a predetermined maximum round-trip delay and an estimated speed.
- an upper limit based on the remaining amount of the reception buffer may be separately provided, and if it is less than this upper limit, a value based on the estimated speed may be used to determine the reception window size.
- the node processing load is reduced by having the information represented by a leader packet, which is conventionally held redundantly in a plurality of buckets.
- header size reduction effect occurs.
- the round-trip delay time is controlled to be less than or equal to the predetermined maximum round-trip delay time regardless of the path speed difference between the forward path and the return path.
- delay dispersion is reduced compared to conventional window control for paths with speed fluctuations.
- the window size is obtained directly using the speed estimation results, there is a feature that the bandwidth can be effectively used even on a route with large loss and delay variation. This is an advantage over Non-Patent Documents 1 and 9.
- Non-Patent Documents 1 and 9 the smaller of the congestion window size and the reception window size on the transmission side is adopted as the window size, so control is performed to reduce the window size when packet loss occurs. End up.
- the window size is determined using the speed estimation result, and therefore the window size itself is not affected! Can be used.
- the leader packet may include the speed estimation result instead of the reception window size, and the window size calculation may be performed by the packet analysis unit 207 of the transmission side node. Furthermore, it is also possible to divide the amount of transmittable data calculated at the time of receiving the reader packet into multiple packet blocks that do not send all at the same time, and send them over a transmission interval that also determines the estimated speed power! / ⁇ . However, the transmission interval at this time needs to be an interval at which the packet block transmitted earlier and the packet block transmitted later also contact on the route. In addition, by dividing and transmitting the bucket block in this way, the size of the burst sent to the path is reduced, so the effect of lowering the packet loss probability can be expected.
- Non-Patent Document 6 a method using the PAC scheduler method disclosed in Non-Patent Document 6 will be described.
- the basic operation of the PAC scheduler at the sending node is to predict the current route delay based on the delay and estimated speed fed back from the receiving node, and the transmission history after the packet used to measure the route state. Then, the next packet is transmitted at a time when the predicted delay is less than a predetermined threshold.
- An example of the flow control protocol operation when this is applied to the packet block transfer method is described with reference to FIG.
- the “101 time” number line represents the time on forwarding node 101
- the “102 time” number line represents the time on forwarding node 102.
- the forwarding node 101 transmits the packet block 300-a at time ts (a), and the head thereof reaches the forwarding node 102 at time tb (a).
- tb (a) means the time when the head of the packet block 300-a reaches the receiving node when there is no load.
- the head of packet block 300-a reaches node 102 only with a delay at no load.
- the estimated speed calculated from the reception time of the leader packet 302-a and the data packet 301-1 and the leader packet 302 -a The time stamp installed in a and the delay information estimated from the reception time of the forwarding node 102 are transmitted to the forwarding node 101 from the forwarding node 102 to the forwarding node 101 in ts (l).
- the node 101 is notified by accommodating the node 101.
- the node 101 After the packet block 300-a, the node 101 considers that the bandwidth of the path is constant and transmits the packet block 300-b. After transmitting packet block 300-b, the path information (estimated speed and delay) is updated by receiving 302-f. Since this is information before time T, node 101 still has the path bandwidth when 300-a is transmitted. Similarly, it is determined that there is no load, and packet block 300-c is transmitted at time ts (c). At this time, the node 101 determines time ts (c) by the following method. That is, the packet block 300 obtained from the route information (estimated speed and delay) included in the received packet 302-f and the transmission history after the packet block 300-a used for measurement of the route information. The time c) at which the head of -c is predicted to be received is determined as the same or earlier than the value obtained by adding the threshold th to the predicted arrival time tb (c) at no load.
- the predicted no-load arrival time tb (c) is the estimated time when the packet arrives at the receiving node 102 when the packet is transmitted with no load at ts (c). is there. Therefore, at the time of ts (c), the node 101 expects the beginning of the packet block 300-c to reach tb (c) + th.
- the size of the packet block 300-c is determined so that the predicted arrival time at the end is within tb (c) + to where tb (c) includes the maximum surplus delay to.
- the maximum surplus delay indicates a threshold value that prevents packet transmission from a route that is expected to have a longer delay.
- the path r of the packet P is used until the transmission time-the predicted arrival time when the transmitting node transmits the packet P from the path r reaches the transmission time after the transmission time-to. Do not send. Conversely, packets can be sent immediately from the route that is the predicted arrival time and the current time to.
- the propagation delay of packet block 300-b is larger than the prediction of node 101.
- the start is delayed until tf '(c), and the reception completion time of the last data packet 301-9 is delayed from the predicted tb (c) + to, and becomes (d).
- Node 102 transmits packet block 300-g to ts (g), and this leader packet 302-g receives packet block 300-b. More detected route information (that is, information in which the route bandwidth is reduced) is included.
- node 101 knows that the bandwidth of the route has been reduced by receiving packet 302-g, so it updates the route information accordingly, and the next packet block 300 that was scheduled to be transmitted to ts (d) before the update is updated. Change the transmission time of -d to ts, (d).
- the size of the packet block sent from node 101 to 102 is also 300-b,
- the above determination is based on the updated route information and transmission history, and the predicted arrival delay d) at the beginning of the block is equal to or less than tb (d) + th, and the predicted arrival time at the end is tb. (d) This is the result of calculating the transmission time and block size so that it is less than + to. Actually, the route state fluctuates between ts (g) and after (d), so in the figure, if the time when the head actually arrives, (d) is deviated from d).
- the PAC scheduler performs transmission timing control so as to correct the deviation of the predicted arrival time, and achieves both effective use of bandwidth and delay suppression.
- the window control described above provides control over the round trip delay
- the PAC scheduler further measures and hoods back the delay for each one-way path, so that each path has a different path status. There is an advantage that appropriate control can be performed. For example, when the return path band is large with respect to the current load and the return path band is small with respect to the current load, only the increase in the return path delay increases.
- the sending node on both the forward path and the return path reduces the load on the path, and as a result, the bandwidth of the forward path cannot be used effectively.
- the outbound and inbound delays are monitored separately at this time, it will be understood that it is necessary to reduce the load on the outbound path, so the outbound bandwidth can be used effectively.
- FIG. 5 shows a flow chart followed by the scheduling unit 202 to realize the operation of the flow control protocol
- FIG. 6 shows a flowchart followed by the packet block creation unit 203.
- the scheduling unit 202 determines the time when the packet block can be transmitted next and the block size each time the storage content of the storage unit 205 is updated, and notifies the packet block creation unit 203 of the time.
- FIG. 5 will be specifically described.
- the processing of S51 when the path state information or the packet transmission history information is updated from the storage unit 205, the notification is received.
- the predicted arrival time at the beginning of the next packet to be transmitted is the predicted arrival time at no load.
- the next packet transmission time (tl) is calculated so as to be equal to the time obtained by adding the threshold (th) to (tb) (S52).
- the amount of data (d) that is predicted to cause the propagation delay of the maximum surplus delay to is calculated from the latest transmission path state information (S53).
- the packet transmission time (tl) and the data amount (d) obtained as described above are notified to the packet block creating unit 203 (S54), and the process is terminated.
- the packet block creation unit 203 configures a packet block having a size not larger than the block size notified from the scheduling unit 202.
- the data is passed to the transmission unit 204.
- Fig. 6 there are two types of power starting from wait and idle. This is because the process of “transmitting a packet block at the time when transmission is possible” is realized with the aid of a timer. is there. While waiting for the timer to expire, the scheduling unit 202 remains in the wait state. On the other hand, the idle state does not wait for the timer and does not receive any data to be processed.
- FIG. 6 will be described below.
- the idle state will be described.
- a “new packet reception notification is received from the buffer unit 201” (S61) or “update notification of packet transmission time (tl) and data amount (d) is received from the scheduling unit 202” If so (S62), the process is started.
- the process it is first determined whether the packet transmission time (tl) ⁇ present (S63). If the result of this determination is No, the timer is started so as to expire at the time, the processing is terminated (S64), and a wait state is entered. If the determination result in S63 is Yes, a leader packet including the latest reception path state information and Ack information is created (S65).
- the buffer unit 201 also takes out the maximum amount of data for which the header power packet block size of the notafer is equal to or smaller than the data amount (d), and creates a data packet group (S66).
- the leader packet and the data packet group are transmitted to the packet transmission unit 204 (S67).
- the transmission record is stored in the storage unit 205. Write (S68) ends. Return to Udle state.
- the notification to the scheduling unit 202 when the storage contents of the storage unit 205 are updated is the power that the storage unit performs. This is indicated by the packet analysis unit 207 and the packet block creation unit 203. Alternatively, the same operation is realized even when the packet transmission unit 204 performs.
- the leader packet 302-b of the packet block 300-b is received at an interval from the end of reception of the immediately preceding packet block 300-a, so there is no load. Arrived in state. When such an interval is detected, forwarding node 102 forwards no-load delay detection by including information indicating that packet 302-b has arrived without load in leader packet 302-g addressed to the transmitting node. Tell node 101.
- the forwarding node 101 Upon receiving the leader packet 302-g, the forwarding node 101 updates the no-load path delay value to the delay of 302-b.
- a no-load delay value is used periodically. For this reason, if no load delay is detected for more than a certain time, the excess delay th, which is the block transmission threshold, is set to a negative value to forcibly induce no-load transfer. Thus, a load path delay value in a cycle within a certain time can be realized. For this reason, the path is completely loaded as th 0, creating a state in which data can be transmitted only in the state.
- each of the components in the forwarding nodes 101 and 102 feeds back the monitoring result of the delay and speed of the communication path and performs the packet flow rate control. Packet transmission considering (estimated speed and delay) is possible.
- the signaling information is aggregated into the leader packet and communicated in this way, both the extraction of the flow control information on the receiving side and the updating of the scheduling information on the transmitting side can be performed as in the conventional example. Since it is performed at the frequency of packet block transmission / reception that is not performed at the same frequency, the processing load on the forwarding node is reduced.
- the signaling information is omitted from the data packet and aggregated into the leader packet. There is an effect of keeping the total of
- high-accuracy route bandwidth estimation using the packet train method can be performed using a dedicated probe packet (dummy packet) that is used only for speed measurement without impairing bandwidth utilization efficiency.
- the estimated speed cannot be obtained if the bottleneck bandwidth of the path is equal to or higher than the transmission speed.
- an application power path that requires a specific speed can be transferred at a speed higher than that speed. It can be determined whether or not. The reason is that when the reception rate is lower than the transmission rate, the transfer rate is limited by the path speed, and the speed can be measured. Even if this is not the case, it is clear that the route speed is equal to or higher than the transmission rate (because there is no link with a lower rate than the transmission rate on the route). By keeping the transmission speed below a certain level, the load on the path can be reduced compared to the case of simultaneous transmission. Therefore, the possibility of packet loss due to excessive load can be reduced.
- flow delay and bandwidth monitoring information (route information) is included as flow control information, but reception confirmation information and the like may also be included. For example, if the reception confirmation of the leader packet and reception within the corresponding packet block are included, and the identification information (retransmission control information) of the data packet is included, the packet to be retransmitted by the transmission side can be accurately identified.
- leader packet When a leader packet is lost, for example, information indicating the loss of the leader packet and Ack information based on a sequence number similar to that used in TCP are used together to maintain transmission integrity. Be drunk. For example, if a sequence number is also assigned to the leader packet so that the receiving side returns Ack information, the sender side resends the bad packet so that both the leader packet and the data packet can be transferred without loss. Can be guaranteed.
- a single leader packet may be used. In that case, you may create a clone of the same packet, or you may combine multiple reader packets to cover the block control information for the entire block.
- the leader packet in the present invention may include a data body when the packet block is composed of a plurality of data packets, and the block management information is an existing identifier for identifying the packet in the block. For example, when the Identification field in the IP header is used, the block management information is not required for the data packet.
- the reader packet which is a feature of the present invention, includes the intra-block packet and the reader packet, except for the limitations that "the reader packet does not include user data" and "the data packet includes block cost management information". This is because it is possible to give information that can identify the flow control information generation method from them, and to generate flow control information according to the packet block to which the received packet belongs at the receiving node.
- the above control information may be routing information! /.
- a packet block is configured and transmitted only when periodic line monitoring is performed. Each data packet may be transmitted alone until the time.
- FIG. 7 shows the configuration of the transfer nodes 101 and 102 used in this embodiment.
- the configuration of the forwarding node 101 is the same as that of FIG. 1 used in the first embodiment.
- the forwarding node 102 has a plurality of packet receiving units 206 and a packet transmitting unit 204.
- the IP network 400 between the forwarding nodes 101 and 102 also provides a route to the packet receiving unit 206-1 of the forwarding node 102 and a route to the packet receiving unit 206-2.
- Each path generally includes physically separated links, and the bandwidth and delay vary independently of each other.
- the bandwidth and delay of the path from the packet transmitting unit 206-1 of the forwarding node 102 to the packet receiving unit 204 of the forwarding node 101 and the route from the packet receiving unit 204-2 also vary independently of each other. Let's say.
- This embodiment is an application example of the present invention to the case where there are a plurality of selectable paths between two nodes as described above.
- the packet structure and signaling mechanism for flow control in this embodiment are the same as in the first embodiment, but the contents of flow control information and scheduling information, and the configuration of packet blocks are different from those in the first embodiment.
- the packet block in this embodiment includes a data packet sent to each of a plurality of paths and one or more reader packets.
- the sequence number is given for each route
- the leader packet includes a time stamp as block management information and a sequence number range of packets in the packet block for each route.
- the delay and speed estimation by the packet analysis unit 207 is also performed for each path.
- the flow control information on the leader packet includes the delay and speed estimation results for each route.
- Packet blocks 300-1 and 300-2 shown in Fig. 7 are examples in which there are two selectable routes and one leader packet.
- the packet block creation unit 203 creates one packet block for a plurality of routes, and the packet transmission unit 204 sends out each packet in the block via the route designated by the packet block creation unit 203.
- Route bandwidth estimation by the packet train method is possible when two or more packets are transmitted simultaneously on the route or at a transmission rate that is equal to or higher than the bottleneck bandwidth of the route. To do this, it is necessary to send two or more packets from each route in one packet block. For this reason, the packet sent as a result of packet block configuration A dummy packet is transmitted along with the packet to be transmitted for the route that has become a si force. The contents of the dummy packet need only be a sequence number, for example.
- the flow control algorithm used in the present embodiment can be extended and applied.
- the window control method described as an example of the first flow control algorithm in the first embodiment is extended as follows in the second embodiment.
- the Ack information of all routes and the reception window size are included in the leader packet.
- the scheduling unit 202 of the transfer node that has received the leader packet calculates the transmittable data amount of each path as shown later and notifies the packet block generation unit 203 of it.
- the packet block generation unit 203 extracts the data in the buffer unit so that each path part of the packet block does not exceed the transmittable data amount! /, And determines the allocation to each path.
- the transmittable data amount does not satisfy the predetermined lower limit, data is not allocated to the route.
- the leader packet is assigned to the route with the largest amount of data that can be transmitted.
- the same leader packet may be assigned to other routes for redundancy.
- a dummy packet is also assigned to a route that cannot be assigned with one packet including the leader and data.
- the allocation route for each packet in the block is determined, all the packets in the block are passed to the packet transmission unit 204 together with the allocation information.
- the packet information passed to the packet transmission unit 204 is recorded in the storage unit as a transmission history.
- the reception window size in the receiving side node is determined in the same manner as in the first embodiment.
- the PAC scheduler method described as an example of the second flow control algorithm in the first embodiment is extended as follows in the second embodiment.
- the operation when two paths are included will be described with reference to FIG. In Fig. 8, there are two time lines corresponding to the two routes, and the arrival prediction sequences of packets passing through each route are shown.
- tb, b, th, and to are shown as tb (n), n), th (n), and to (n), respectively, and their meanings are shown in FIG.
- the transmitting node can transmit a packet block for any route if tb (n) + th (n) tn). In Fig. 8, tb (l) + th (l) (1), so transmission is possible immediately.
- Scheduling section 202 calculates the estimated arrival time of leader packet 302 for each route, and configures a packet block to transmit leader packet 302 to the route predicted to arrive earliest.
- route 2 is selected as the leader packet transmission route.
- the data size of the first packet 301-1 in the buffer unit 201 is acquired, and placed in the route predicted to arrive earliest when transmitted together with the reader packet.
- the first data packet 301-1 is routed to route 1.
- increase the number of packets in the block one by one and arrange them in the same way, and for each route, find the maximum number of transmitted packets so that the sum of the transmitted packet sizes is less than tb (n) + to (n).
- the packet block configuration is determined.
- a packet block 300 including five data packets 301-1 to 301-5 including path 1 and 2 is configured! /.
- the packet reception power flow that is not limited to the transmission delay of the flow control information. Includes waiting time before sending control information
- this waiting time is the maximum block transmission interval, the node processing load is reduced, the accuracy of route bandwidth estimation in the bucket train method, and the amount of data buffered on the route for effective use of bandwidth exceeds a certain level. Increasing the amount of allocation to each path per block for the purpose of maintaining, etc., increases the flow control information feedback delay.
- the “101 time” number line represents the time on the forwarding node 101
- the “102 time” number line represents the time on the forwarding node 102.
- route 1 and route 2 there are two routes, route 1 and route 2, between forwarding node 101 and forwarding node 102, and for the packet passing through each route, the receiving sequence on forwarding node 102 by forwarding node 101 is The predictions are shown on the “Route 1 Reception Prediction” and “Path 2 Reception Prediction” lines.
- Each packet block basically consists of only one route.
- the block size of the packet block of each route is determined so that the transmission delay of the packet in the block is constant, that is, the block of all routes (the sum of the packet sizes contained in the block / route speed) is the same. It is.
- each packet block is transmitted by shifting the block transmission period ti defined by the number of paths that can be used for the transmission delay Z signaling of the intra-packet packet, and this is the leader packet transmission period.
- the arrival of the packet block using route 1 and the packet block using route 2 are alternated.
- the transmission interval (feedback period) of leader packets can be shortened compared to the length of the packet train used for speed estimation.
- the arrival interval of the leader packet is almost equal to the length of the packet train. This is because when a packet block is transmitted using a plurality of routes as described in, the arrival of the leader packet can be received at intervals of ti shorter than the length of the packet train.
- the node 101 detects the speed change based on the information of the leader packet transmitted from the node 102 at ts (l) or ts (g), and the packet block 300-d transmitted at ts (d). Is adjusted so that the time required for reception from the beginning to the end (reception completion time) at node 102 is less than or equal to t ports.
- the packet block to be transmitted next to d is also transmitted from route 2 with a size that the reception completion time is t port or less. Thereafter, until route 1 is ready for transmission, that is, until tb + th becomes tf, a packet block having a size that the reception completion time is less than t port degrees is transmitted only from route 2. As described above, the transmission / reception cycle of the leader packet can be kept below about ti even when the delay of the route 1 is increased due to the change of the route state.
- the node 101 detects the speed change between ts (c) and ts (d), and the data that can be reached before the arrival time e) of the leader packet via the next route 1 Add as many packets over route 1 to packet block 300-d sent to ts (d).
- the packet block 300-d includes the packets of both paths 1 and 2.
- the block size is re-determined so that the transmission delay is within the number of available paths of ti for the new speed.
- the arrival interval of the leader bucket is kept within ti as before the speed change.
- the bandwidth utilization efficiency is improved.
- the present invention determines the transmission timing and the transmission path on the transmission side in units of packet blocks rather than being performed for each single packet or packet pair as in the conventional example, the data amount of the packet block Increasing the value reduces the load on the transmission processing of the forwarding node.
- the state information of the route with a large delay is transmitted using the route with the large delay itself. Compared to feedback, there is an effect that feedback can be made in a short time.
- the number of packets included in the packet block for each route If this is the case, the amount of signaling information is omitted from the data packet, and the total packet size can be kept lower than the conventional example for the same amount of data.
- the reception speed is estimated on the assumption that the data packets 301-2 and 301-3 included in the packet block 300-b are received immediately after 301-1. Similarly, it is assumed that the data packet 301-4 included in the packet block 300-c is received immediately after 301-3. Then, a virtual packet block such as the packet block 400 is configured. At this time, the time required to receive data packets 301-2 and 301-3 is the difference between the reception times of packet 302-b and packet 301-3 measured when packet block 300-b is received, and data packet 301-4. The time required to receive the packet is the difference between the reception time of the packet 302-c and the packet 301-4 measured when the packet block 300-c is received.
- the time required for receiving the data packet of the virtual packet block 400 is the sum of the time required for receiving the data packet measured at the time of receiving 300-a, 300-b, and 300-c.
- the receiving speed is estimated to be the time required to receive the data packet of the virtual packet block 400 divided by the total amount of data of four data packets from 301-1 to 301-4.
- the block management information 500 includes the speed of each link predicted by the transmitting side node and the size of the data in the block, thereby enabling early detection of communication abnormality.
- the outline of the operation to perform is shown.
- Fig. 11 shows block management information transmitted in the third embodiment when a packet block including packets with sequence numbers 1 to 6 for a certain path is estimated to have a link speed of 350 Kbps, and reception. The relationship of data packet reception timing at the side node is shown. The number line at the bottom of Fig. 11 shows the actual packet arrival sequence at the receiving node.
- the receiving node does not do anything because it reaches the packet 2 at the speed predicted by the transmitting node, but detects an increase in the arrival interval in packet 3, and the sending side's predicted speed and actual transfer speed. Recognize that a shift has occurred. When the deviation is determined to be greater than or equal to a certain level, an error report is created and sent back to minimize the time during which the sending node has misrecognized the link status and improve the followability of the flow control operation.
- FIG. 12 shows an example of block management information and a data packet reception sequence at the receiving side node in the fourth exemplary embodiment of the present invention.
- the block management information 500 includes user flow information for each packet in the block.
- the user flow refers to a packet transfer session between end hosts that is a target for storing the packet arrival order, and each user flow is assigned a unique ID.
- each packet in the user flow is assigned a sequence number that is unique in the user flow, separately from the sequence number for each route.
- the sending node contains two user flow forces 2049 and 2050 in a packet block containing packets with sequence numbers 4 to 11 per route.
- Figure 12 Number of times at the bottom The straight line shows the actual packet arrival sequence at the receiving node.
- the loss of packet 5 and packet 6 as a result of transmission. Loss is detected by reception of packet 7. Also, since the flow management information 500 shows that lost packets 5 and 6 are # 1 4 of flow 2049 and # 41 of flow 2050, respectively, packets belonging to these two flows that arrived so far It is transferred without waiting for arrival.
- Non-Patent Document 8 in the conventional example in which a separate sequence number is assigned to each route and each user flow.
- user flow information of a lost packet is not provided.
- the receiving node can know the sequence number of each lost user packet for each user flow, the packet loss can be reliably detected without being confused with the order inversion. Therefore, transfer hold can be omitted when packet loss is detected, and this has the effect of reducing jitter between end-host sessions.
- the data packet to be compressed includes a common part in the data area.
- the start sequence number and end sequence number indicate the range of packets subject to data compression.
- the start bit and the end bit indicate a common range in the data area of the compression target packet by the viewpoint and end point of the bit position.
- the data between the start bit and the end bit contains common data from the start bit to the end bit. While the above compressed information is included in the block management information, the common data specified there is deleted from the data field of the data packet to reduce the total amount of transmission packet size.
- Data compression as described above is particularly effective when the data itself includes a header of another protocol and the common part is large. Packet combinations and bits with different common parts If there are multiple items in the range, a plurality of pieces of compressed information as shown in FIG. 13 may be included in the block management information.
- the transmission data compression according to this embodiment is simpler in control than the conventional data compression method that requires signaling for sharing the compression state between the transmission and reception nodes.
- the above control information may be routing information! /.
- the nodes 101, 102, etc. of the present invention not only realize the operation as a node, but also realize the operation as a software by executing a program that executes the function of each part on a computer. You can also This program is held in a magnetic disk, a semiconductor storage device, or other recording medium, and the recording medium force is read into a computer, and the above-described functions can be realized by controlling its operation.
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer Security & Cryptography (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
- Communication Control (AREA)
Abstract
Description
Claims
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US12/161,922 US20090059958A1 (en) | 2006-01-23 | 2007-01-18 | Communication method, communication system, nodes and program |
JP2007554937A JP4780343B2 (ja) | 2006-01-23 | 2007-01-18 | 通信方法、通信システム、ノードおよびプログラム |
EP07706964A EP1981220A4 (en) | 2006-01-23 | 2007-01-18 | COMMUNICATION PROCEDURE, COMMUNICATION SYSTEM, NODES AND PROGRAM |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2006014458 | 2006-01-23 | ||
JP2006-014458 | 2006-06-05 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2007083687A1 true WO2007083687A1 (ja) | 2007-07-26 |
Family
ID=38287639
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2007/050660 WO2007083687A1 (ja) | 2006-01-23 | 2007-01-18 | 通信方法、通信システム、ノードおよびプログラム |
Country Status (6)
Country | Link |
---|---|
US (1) | US20090059958A1 (ja) |
EP (1) | EP1981220A4 (ja) |
JP (1) | JP4780343B2 (ja) |
KR (1) | KR20080079335A (ja) |
CN (1) | CN101379781A (ja) |
WO (1) | WO2007083687A1 (ja) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2009253663A (ja) * | 2008-04-07 | 2009-10-29 | Mitsubishi Electric Corp | モバイルルータアドホックネットワーク通信システム |
WO2011040006A1 (ja) * | 2009-09-30 | 2011-04-07 | パナソニック株式会社 | 送信装置、受信装置、送受信システム、及び送受信方法 |
JP2011160039A (ja) * | 2010-01-29 | 2011-08-18 | Brother Industries Ltd | パケット送信装置、パケット送信方法及びパケット送信プログラム |
Families Citing this family (44)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP5103900B2 (ja) * | 2006-12-28 | 2012-12-19 | 富士通株式会社 | パス状態監視方法及び装置 |
JP2008199605A (ja) * | 2007-02-12 | 2008-08-28 | Asustek Computer Inc | 無線通信システムにおいてリソース使用効率を向上させる方法及び装置 |
EP2003799A1 (en) * | 2007-06-12 | 2008-12-17 | Sony Deutschland Gmbh | Adaptive history aware beam steering |
JP4941753B2 (ja) * | 2007-08-31 | 2012-05-30 | 横河電機株式会社 | フィールド制御システム |
US8270404B2 (en) * | 2008-02-13 | 2012-09-18 | International Business Machines Corporation | System, method, and computer program product for improved distribution of data |
US9047421B2 (en) * | 2008-04-30 | 2015-06-02 | Alcatel Lucent | Serial link buffer fill-level compensation using multi-purpose start of protocol data unit timing characters |
US8374091B2 (en) * | 2009-03-26 | 2013-02-12 | Empire Technology Development Llc | TCP extension and variants for handling heterogeneous applications |
WO2011033568A1 (ja) * | 2009-09-17 | 2011-03-24 | 富士通株式会社 | 通信方法、通信システム、送信装置、受信装置 |
US8824285B1 (en) * | 2009-12-16 | 2014-09-02 | Dnutch Associates, Inc. | System and method for collision detection and avoidance during packet based communications |
US9515925B2 (en) | 2011-05-19 | 2016-12-06 | Qualcomm Incorporated | Apparatus and methods for media access control header compression |
US9125181B2 (en) * | 2011-08-23 | 2015-09-01 | Qualcomm Incorporated | Systems and methods for compressing headers |
EP2760182B1 (en) * | 2011-09-21 | 2017-03-01 | Fujitsu Limited | Data communication apparatus, data transmission method, and computer system |
US9479979B2 (en) * | 2012-06-29 | 2016-10-25 | Nokia Solutions And Networks Oy | Success rate improvements for ANR measurements while reducing data loss at a UE |
US9451469B2 (en) * | 2012-07-23 | 2016-09-20 | Intel Corporation | Apparatus and method for tunneled GPM |
EP2936739B1 (en) * | 2012-12-21 | 2017-12-13 | Telefonaktiebolaget LM Ericsson (publ) | Method and node arrangement for providing more accurate estimation of data path conditions |
KR102094718B1 (ko) * | 2013-09-26 | 2020-05-27 | 삼성전자주식회사 | 무선 네트워크에서 학습에 기반한 중계 노드 선택 방법 및 중계 장치 |
US9729439B2 (en) * | 2014-09-26 | 2017-08-08 | 128 Technology, Inc. | Network packet flow controller |
US10277506B2 (en) | 2014-12-08 | 2019-04-30 | 128 Technology, Inc. | Stateful load balancing in a stateless network |
US9736184B2 (en) | 2015-03-17 | 2017-08-15 | 128 Technology, Inc. | Apparatus and method for using certificate data to route data |
US9729682B2 (en) | 2015-05-18 | 2017-08-08 | 128 Technology, Inc. | Network device and method for processing a session using a packet signature |
US9762485B2 (en) | 2015-08-24 | 2017-09-12 | 128 Technology, Inc. | Network packet flow controller with extended session management |
US9871748B2 (en) | 2015-12-09 | 2018-01-16 | 128 Technology, Inc. | Router with optimized statistical functionality |
US9985883B2 (en) | 2016-02-26 | 2018-05-29 | 128 Technology, Inc. | Name-based routing system and method |
US10205651B2 (en) | 2016-05-13 | 2019-02-12 | 128 Technology, Inc. | Apparatus and method of selecting next hops for a session |
US10298616B2 (en) | 2016-05-26 | 2019-05-21 | 128 Technology, Inc. | Apparatus and method of securing network communications |
US10200264B2 (en) | 2016-05-31 | 2019-02-05 | 128 Technology, Inc. | Link status monitoring based on packet loss detection |
US10091099B2 (en) | 2016-05-31 | 2018-10-02 | 128 Technology, Inc. | Session continuity in the presence of network address translation |
US11075836B2 (en) | 2016-05-31 | 2021-07-27 | 128 Technology, Inc. | Reverse forwarding information base enforcement |
US10257061B2 (en) | 2016-05-31 | 2019-04-09 | 128 Technology, Inc. | Detecting source network address translation in a communication system |
US10841206B2 (en) | 2016-05-31 | 2020-11-17 | 128 Technology, Inc. | Flow modification including shared context |
US9832072B1 (en) | 2016-05-31 | 2017-11-28 | 128 Technology, Inc. | Self-configuring computer network router |
US10009282B2 (en) | 2016-06-06 | 2018-06-26 | 128 Technology, Inc. | Self-protecting computer network router with queue resource manager |
US9985872B2 (en) | 2016-10-03 | 2018-05-29 | 128 Technology, Inc. | Router with bilateral TCP session monitoring |
US10425511B2 (en) | 2017-01-30 | 2019-09-24 | 128 Technology, Inc. | Method and apparatus for managing routing disruptions in a computer network |
WO2018165182A1 (en) | 2017-03-07 | 2018-09-13 | 128 Technology, Inc. | Router device using flow duplication |
US10432519B2 (en) | 2017-05-26 | 2019-10-01 | 128 Technology, Inc. | Packet redirecting router |
KR101983088B1 (ko) * | 2017-06-23 | 2019-05-31 | (주)넷비젼텔레콤 | 다중 경로 환경에서의 udp 패킷 처리 방법 |
US11165863B1 (en) | 2017-08-04 | 2021-11-02 | 128 Technology, Inc. | Network neighborhoods for establishing communication relationships between communication interfaces in an administrative domain |
KR20200053620A (ko) * | 2017-10-05 | 2020-05-18 | 지멘스 악티엔게젤샤프트 | 네트워크 및 통신 네트워크를 구성하기 위한 방법 및 장치 |
US20190253341A1 (en) | 2018-02-15 | 2019-08-15 | 128 Technology, Inc. | Service Related Routing Method and Apparatus |
GB2575509A (en) * | 2018-07-13 | 2020-01-15 | Nokia Technologies Oy | Spatial audio capture, transmission and reproduction |
GB2575511A (en) | 2018-07-13 | 2020-01-15 | Nokia Technologies Oy | Spatial audio Augmentation |
WO2021217070A1 (en) | 2020-04-23 | 2021-10-28 | Juniper Networks, Inc. | Session monitoring using metrics of session establishment |
KR102502758B1 (ko) | 2022-11-24 | 2023-02-23 | 펌킨네트웍스(주) | 멀티코어 환경의 운영체제에 적합한 패킷 전송 장치 |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH08223217A (ja) * | 1995-02-08 | 1996-08-30 | Nippon Telegr & Teleph Corp <Ntt> | 無線パケット多重方法 |
JP2004088208A (ja) * | 2002-08-23 | 2004-03-18 | Sony Corp | データ伝送システム及びデータ伝送方法 |
WO2005067261A1 (ja) * | 2004-01-09 | 2005-07-21 | Nec Corporation | 通信方法 |
Family Cites Families (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH01101758A (ja) * | 1987-10-15 | 1989-04-19 | Oki Electric Ind Co Ltd | 固定長パケット通信におけるパケット分解・組立および粉失検出方法 |
US6021433A (en) * | 1996-01-26 | 2000-02-01 | Wireless Internet, Inc. | System and method for transmission of data |
JP3587352B2 (ja) * | 1999-02-04 | 2004-11-10 | 富士通株式会社 | ネットワーク通信性能測定方法及び装置並びにネットワーク通信性能測定プログラムを格納したコンピュータ読取り可能な記録媒体 |
US20020031103A1 (en) * | 2000-05-02 | 2002-03-14 | Globalstar L.P. | User terminal employing quality of service path determination and bandwidth saving mode for a satellite ISP system using non-geosynchronous orbit satellites |
JP3816314B2 (ja) * | 2000-07-11 | 2006-08-30 | 三菱電機株式会社 | パケット交換装置 |
EP1175034A2 (en) * | 2000-07-18 | 2002-01-23 | Eastman Kodak Company | A packet data transmission system with adaptive packet size |
US20020073217A1 (en) * | 2000-12-08 | 2002-06-13 | Ma David Yin-Shur | Method and apparatus for facilitating communication between a wireless device and disparate devices or systems |
JP4187940B2 (ja) * | 2001-03-06 | 2008-11-26 | 株式会社エヌ・ティ・ティ・ドコモ | パケット伝送方法及びシステム、並びにパケット送信装置、受信装置、及び送受信装置 |
US20020131103A1 (en) * | 2001-03-16 | 2002-09-19 | Nicholas Bambos | Method and system for reconfiguring a network element such as an optical network element |
US7386000B2 (en) * | 2001-04-17 | 2008-06-10 | Nokia Corporation | Packet mode speech communication |
US8098648B2 (en) * | 2004-01-09 | 2012-01-17 | Nec Corporation | Load distributing method |
US7778179B2 (en) * | 2005-11-23 | 2010-08-17 | Telefonaktiebolaget L M Ericsson (Publ) | Using filtering and active probing to evaluate a data transfer path |
-
2007
- 2007-01-18 EP EP07706964A patent/EP1981220A4/en not_active Withdrawn
- 2007-01-18 CN CNA2007800028984A patent/CN101379781A/zh active Pending
- 2007-01-18 JP JP2007554937A patent/JP4780343B2/ja not_active Expired - Fee Related
- 2007-01-18 KR KR20087018142A patent/KR20080079335A/ko not_active Application Discontinuation
- 2007-01-18 WO PCT/JP2007/050660 patent/WO2007083687A1/ja active Search and Examination
- 2007-01-18 US US12/161,922 patent/US20090059958A1/en not_active Abandoned
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH08223217A (ja) * | 1995-02-08 | 1996-08-30 | Nippon Telegr & Teleph Corp <Ntt> | 無線パケット多重方法 |
JP2004088208A (ja) * | 2002-08-23 | 2004-03-18 | Sony Corp | データ伝送システム及びデータ伝送方法 |
WO2005067261A1 (ja) * | 2004-01-09 | 2005-07-21 | Nec Corporation | 通信方法 |
Non-Patent Citations (11)
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2009253663A (ja) * | 2008-04-07 | 2009-10-29 | Mitsubishi Electric Corp | モバイルルータアドホックネットワーク通信システム |
WO2011040006A1 (ja) * | 2009-09-30 | 2011-04-07 | パナソニック株式会社 | 送信装置、受信装置、送受信システム、及び送受信方法 |
US8593985B2 (en) | 2009-09-30 | 2013-11-26 | Panasonic Corporation | Transmitting apparatus, receiving apparatus, transmitting and receiving system, and transmitting and receiving method |
JP2011160039A (ja) * | 2010-01-29 | 2011-08-18 | Brother Industries Ltd | パケット送信装置、パケット送信方法及びパケット送信プログラム |
US8514858B2 (en) | 2010-01-29 | 2013-08-20 | Brother Kogyo Kabushiki Kaisha | Packet transmission device, packet transmission method and computer program product stored on non-transitory computer-readable medium |
Also Published As
Publication number | Publication date |
---|---|
EP1981220A4 (en) | 2011-04-13 |
CN101379781A (zh) | 2009-03-04 |
US20090059958A1 (en) | 2009-03-05 |
KR20080079335A (ko) | 2008-08-29 |
JP4780343B2 (ja) | 2011-09-28 |
EP1981220A1 (en) | 2008-10-15 |
JPWO2007083687A1 (ja) | 2009-06-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP4780343B2 (ja) | 通信方法、通信システム、ノードおよびプログラム | |
JP4394541B2 (ja) | 通信装置、データ通信方法およびプログラム | |
EP1705845B1 (en) | Load distributing method | |
EP2479941B1 (en) | Communication apparatus and communication system for enhancing speed of communications between terminals | |
US11451481B2 (en) | Network control apparatus and network control method | |
US6788686B1 (en) | Method of maintaining packet order in multipath transmission systems having non-uniform traffic splitting | |
EP2195980B1 (en) | Method, system and computer program product for adaptive congestion control on virtual lanes for data centre ethernet architecture | |
JP4547341B2 (ja) | 通信品質制御機能を備えるパケット中継装置 | |
US6072797A (en) | Methods, apparatus and computer program products for aggregated transmission groups in high speed networks | |
JP4547339B2 (ja) | 送信制御機能を備えるパケット中継装置 | |
US20050185621A1 (en) | Systems and methods for parallel communication | |
JP2008518552A (ja) | 粗細試験期間を使用したネットワーク・パケットの経験的スケジューリング法 | |
WO2002087276A2 (en) | Method and device for robust real-time estimation of bottleneck bandwidth | |
CN107770085B (zh) | 一种网络负载均衡方法、设备及系统 | |
CN108540380A (zh) | 多子流网络传输方法及装置 | |
JP4488256B2 (ja) | 通信方法、ノード及び制御プログラム | |
WO2005086436A1 (ja) | パケット転送装置、パケット転送ネットワークシステム、および、端末装置 | |
JP7151537B2 (ja) | ネットワークコントローラ装置、ネットワーク制御システム、通信ネットワークの制御方法及びプログラム | |
JP2009105662A (ja) | マルチホップ通信システム、マルチホップ通信方法、端末装置および中継装置 | |
JP4505575B2 (ja) | 通信システム、ゲートウェイ送信装置、ゲートウェイ受信装置、送信方法、受信方法および情報記録媒体 | |
US7088738B1 (en) | Dynamic fragmentation of information | |
Jacquet et al. | Striping over wireless links: effects on transmission delays | |
Jacquet et al. | A method for increasing throughput based on packet striping | |
JP2000236352A (ja) | データ転送システム及びその時間演算方法並びにその制御プログラムを記録した記録媒体 | |
CN115733755A (zh) | 一种可填充网络带宽的数据中心传输控制系统及方法 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
DPE1 | Request for preliminary examination filed after expiration of 19th month from priority date (pct application filed from 20040101) | ||
WWE | Wipo information: entry into national phase |
Ref document number: 2007554937 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 200780002898.4 Country of ref document: CN Ref document number: 12161922 Country of ref document: US Ref document number: 1020087018142 Country of ref document: KR Ref document number: 2007706964 Country of ref document: EP |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
DPE1 | Request for preliminary examination filed after expiration of 19th month from priority date (pct application filed from 20040101) |