JPH05167616A - Communication processing device - Google Patents

Abstract

(57) [Summary] [Object] The present invention uses a communication method such as a packet switching method for dividing a series of data into one or a plurality of units for communication, and makes an effective use of communication resources to create a moving image. It is an object of the present invention to provide a device capable of communicating and processing real-time data such as voice and voice. [Structure] A sequence number management unit 2013 on the transmitting side gives a sequence number to data in a packet to be communicated. The receiving side checks the received sequence number and, if the data is in order, passes this data to the receiver. If the packet in the middle is missing, this data is held in the communication data memory 202, and if the packet in the middle does not arrive after a certain period of time, the data in the complementary data memory 203 is used to drop the data. Is generated, the length and order of the data are restored, and the data held together with the complementary data is passed to the receiver.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention
Data communication processing that requires high-speed and real-time data communication, and the order of data must be correct and the length of data must be the same as before, but communication quality is not so strict. It relates to a device for

[0002]

2. Description of the Related Art There are two requirements for communication quality of moving images and voice. One is the requirement that the recipient must receive exactly the same information that the sender sent. This is a demand for professional use such as music / video production in a studio or the like.

On the other hand, even if the receiver cannot reproduce exactly the same information as the one transmitted by the sender, the receiver should see and hear the information (the quality deterioration due to the change in the information is not noticeable). There is also a demand for good. This is a demand for home use such as music / video viewing at home.

On the other hand, current communication systems are roughly classified into a circuit switching system and a packet switching system. The circuit switching method is a method in which a certain band of a communication network is fixedly assigned to communication between certain senders and receivers, the speed is always stable, and there is little delay in data transmission and no fluctuation. Further, since it is easy to synchronize the time between the transmitting side and the receiving side, it is often used for communication of real-time data such as voice and moving images.

However, even if information is not actually being communicated, for example, in the case of voice, which is in a silent state, the circuit switching system occupies a band and communication resources cannot be effectively used. Further, when the information changes due to noise or the like on the way, it is difficult to correct it. There is also the drawback that storing the information in a form that can be processed later is a bit more cumbersome. Therefore, it is not suitable for the communication of data such as computer data in which the information must always be complete.

In the packet switching system, continuous data is divided into a certain unit length, and a packet is formed by adding information such as a receiver address and a sender address of communication to the packet, and data communication is performed by this. Is.

[0007] According to this method, the communicator basically only needs to generate and communicate packets as needed when necessary, so that there is no wasteful occupation of communication resources, such as circuit switching, and communication is possible. Effective use of resources can be achieved. In addition, because it is easy to detect changes in information due to noise and to easily restore information, like computer data,
It is commonly used for data communication between senders and receivers, where complete restoration of information is essential.

On the other hand, in packet communication, on the contrary, there is no guarantee that the packet will be immediately delivered to the recipient, and the delivery intervals will be different, making it very difficult to synchronize the time between the sender and the receiver. Therefore, it is generally unsuitable for real-time data communication.

In addition, the packet switching system generally has a complicated processing procedure. In particular, the error recovery procedure of data is very complicated, and if data is retransmitted, it becomes more difficult to synchronize the sender and receiver.

On the other hand, there is also a very lightweight protocol that does not perform error recovery, such as a packet communication protocol called UDP. Such a protocol is suitable for high-speed processing, and at first glance, it seems to be a packet communication protocol suitable for transmission of images and voice in situations where communication quality such as home use is not strictly required.

However, since such a protocol does not assume error recovery, there is no concept of a sequence number from the beginning, so that it is impossible to detect a case where a packet is lost or a case where the arrival order of packets is changed. .. In other words, the order of data such as image information and the length of the data are very significant, and a large amount of data that should not be interrupted or the order is exchanged is sent by a protocol without such a concept of a sequence number. That is actually impossible.

In recent years, an asynchronous communication system (hereinafter abbreviated as ATM) has been proposed as a communication system having the features of both the circuit switching system and the packet switching system, and has been attracting attention.

In ATM, continuous data is divided into a fixed length of 48 bytes, and a 5-byte header having information such as a destination address is added to the divided data for communication in units called cells. Unlike packets, cells have a fixed length, so
It is said that the exchange can be easily implemented as hardware and is suitable for speeding up communication.

This ATM system is currently under study by CCITT and the like, and its content is not yet fully fixed. However, the current standard allows cells to have sequence numbers in the adaptation layer. The cell error recovery is possible together with the error detection CRC information. Although the actual error recovery procedure has not been studied, it is basically expected to be recovered by retransmission, so that a problem of synchronization between the sender and the receiver also occurs as in the packet switching method.

Further, in ATM, since error recovery is not always an essential function, it is possible not to detect errors at all, but this method cannot recover the same order as in the UDP example shown above. I cannot avoid the problem.

Further, in ATM, a method of notifying only that an error is detected in the cell by the CRC information and passing the error to the receiver while the cell includes the error is also under study. According to this method, only the order of data is guaranteed when the cell reaches the receiving end but has an error. However, ATM is essentially designed for cell discard, and if cell discard occurs somewhere in the relay stage up to the receiving end, some data is lost due to cell discard. There is no way for the receiving end to notify the recipient. Therefore, in this case, the order of the data and the length of the data still collapse, or the synchronization is lost due to the retransmission, which is a problem.

In any case, however, ATM is still a developing technology and has not been put into practical use at present. Therefore, conventionally, circuit switching is generally used for communication of real-time data such as moving images and voice. However, especially when communication quality is strictly required, the method of ignoring the real-time property, treating voice / image information as the same as computer data and communicating by packet switching method, and reproducing after completing data communication is adopted. ing.

[0018]

However, as described above, the circuit switching system occupies a band even when data is not actually being communicated, and is not a preferable system from the viewpoint of effective use of resources. If possible, it is desirable to use a packet-switching method for communication such as video and audio to effectively use resources and reduce communication costs.

In view of the above problems, the present invention does not cause unnecessary occupation of resources unlike the circuit switching system in real-time data communication such as moving images and voices in situations where strict communication quality is not required. An object is to provide a communication processing device.

[0020]

In order to achieve the above object, in the communication processing apparatus of the present invention, as a first means, the sequence number notified from the transmitting end is used to drop or order data. Data order monitoring means for monitoring the exchange of data, missing data complementing means for the receiving end to independently complement the data lost by the loss of data with predetermined complementary data, and And a sequence reconstructing means for restoring the sequence.

As a second means, a data sequence monitoring means for monitoring data loss and order change using a sequence number notified from the transmitting end, and a data sequence number of a part lost due to data loss. Only the missing data sequence number notifying means for notifying the receiver of the data and the order reconstructing means for restoring the order of the data whose order has been changed are adopted.

As a third means, a data sequence monitoring means for monitoring data loss and order change by using a sequence number notified from the transmission end, and a data order monitor means that the order has been changed and the reception end has already received. When the data in the order prior to the order of the data is received, the order inconsistent data discarding means for immediately discarding this data and the data of the part lost due to the loss of the data or the discarding of the data are predetermined. It is configured to have a missing data complementing means for independently complementing the receiving end with existing complementary data.

As a fourth means, a data sequence monitoring means for monitoring data loss and order change using a sequence number notified from the transmission end, and the order has been changed and the reception end has already received. When the data in the order before the data order is received, only the order inconsistent data discarding means for immediately discarding this data and the sequence number of the data of the part lost due to the loss of the data or the discard of the data are provided. And a missing data sequence number notifying means for notifying the receiver.

[0024]

According to the first means, a high-speed packet (or cell, etc.) process can be executed by a lightweight communication protocol that does not perform retransmission processing and its processing device, and almost real-time data by the packet (or cell, etc.) can be executed. Communication can be realized, the lost part of the data can be recognized by the sequence number that is notified even when the packet is lost, and the required amount of data can be complemented by the predetermined complementing data, and the packet (or Even if the order of arrival of cells is changed, the sequence number can rearrange the data in the correct sequence, and the receiver can obtain the data in which the sequence and the length of the data are completely restored. ..

By the second means, high-speed packet (or cell etc.) processing can be executed by a lightweight communication protocol and its processing device which do not perform retransmission processing, and almost real-time data by packet (or cell etc.) Communication can be realized, and even if a packet is lost, the sequence number that is notified notifies the receiver of the lost part of the data, and the receiver creates the lost part by an arbitrary method. Complementary data can be used to supplement the required amount of data, and even if the arrival order of packets (or cells) is changed, the data can be rearranged in the correct order by the sequence number, which ensures that the data order and length are complete. The recipient will be able to obtain the restored data.

According to the third means, not only the retransmission processing is not performed but also the rearrangement processing associated with the change of the arrival order of the packets is not performed, so that the processing is further simplified.
Real-time data communication can be realized, and even if data is lost due to discarding a packet that arrived late due to packet loss or packet order change, the sequence number notified Recognizing the dropout part, it is possible to complement the required amount of data by the predetermined complementing data, and thus the recipient can obtain the data in which the order and length of the data are completely restored. ..

According to the fourth means, not only the retransmission processing is not performed, but also the rearrangement processing accompanying the change of the arrival order of the packets is not performed, so that the processing is further simplified.
It is possible to realize more real-time data communication, and the sequence number to be notified even if data is lost due to the loss of packets or the discarding of packets that arrived late due to the switching of the packet order. Recognizes the lost part and notifies it to the receiver, and the receiver can complement the required amount with the complementary data created by any method, and the order and length of the data can be improved. The fully restored data will be available to the recipient.

[0028]

FIG. 1 shows a packet format used in the first embodiment of the present invention. This embodiment will be described on the premise of the packet switching system. In FIG. 1, the packet is composed of a header section 11 containing various information for communication processing and a body section 12 containing data to be actually communicated.

The header section 11 is further divided into several fields. That is, a 16-bit sender address field 111, a 16-bit receiver address field 112, and a 16-bit overall packet length field 113, 16b.
It's control flag field 114 (of which, bit 1141 is a sync flag, bit 1142 is an end flag, the remaining 14 bits are unused), a 32-bit sequence number field 115, a 16-bit unused field 116, and a 16-bit header checksum field 117. Is divided into

A 16-bit value uniquely assigned to each correspondent is described in the address fields 111 and 112. The length field 113 of the entire packet describes how many bytes this packet has from the beginning of the packet header to the end of the data body. The synchronization flag 1141 is a flag indicating that the sender is the first packet to be sent in a certain series of communication, and communication is started from the packet in which this flag is set. The end flag 1142 is a flag that indicates that the series of communication that has been performed up to now by the packet is ended, and the packet with this flag set is the last packet to be sent. A packet in which neither of these flags is set indicates that the packet is in the middle of a series of communications. In addition, a packet in which both flags are set indicates that communication is completed with only that packet. A sequence number described later is described in the sequence number field 115. Information for detecting whether or not there is a communication error in the header information is described in the header checksum field 117. Although there are various algorithms for calculating the header checksum (for example, a method of simply adding each byte and using the lower 2 bytes of the answer as the checksum value), the details are essential in this embodiment. Omitted because it is not relevant. In any case, it is possible to know from the value described in this field whether or not a communication error has occurred in the information in the header.

FIG. 2 is a block diagram of the communication processing apparatus in the first embodiment of the present invention. In FIG. 2, the communication processing device is a protocol processing device 201, a communication data memory 202, a complementary data memory 203, a network interface 204,
It is composed of a host interface 205 and a context memory 206, and these respective constituent elements are bus-connected. The protocol processing device 201 further internally includes an address management unit 2011, a context management unit 2012, and a sequence number management unit 20.
13, data division reconstruction processing unit 2014, checksum operation unit
Have 2015.

In the communication context managed by the context management unit 2012, the sender's address, the receiver's address, the sequence number of the data to be transmitted next, the sequence number of the data to be received next, and the arrival delay due to some reason. It contains the first sequence number of the data being stored, the length of that data, a timer, etc. This context information is stored in the context memory 206 when communication is started between a sender and a receiver.
It is recorded on the above, the contents are updated as needed during the communication thereafter, and erased when the communication is completed.

Note that the context information actually includes, for example, other information for flow control for managing the amount of transmission / reception buffers, but since flow control is not essential in the present invention, it is always present. Since sufficient transmission / reception buffers are prepared, the description will be omitted from the context information in this embodiment.

FIG. 3 shows a generation processing algorithm of the data body portion 12 in the transmission processing of the first embodiment of the present invention. The data body part 12 generation processing at the transmitting end will be described with reference to FIGS. 2 and 3.

At the transmitting end, the data to be transmitted is received from the data sender through the host interface 205 and stored in the communication data memory 202 (step 3).
01).

The data division / reconstruction processing unit 2014 checks the length of this data and confirms whether or not it is necessary to divide the data in order to send it to the network (step 301). If the division is necessary, divide it to an appropriate length (step 303) and create one or more data bodies.

A header is added to the data body created here and the packet is transmitted as one packet.

FIG. 4 shows a generation processing algorithm of the header section 11 in the transmission processing of the first embodiment of the present invention. The header section 11 generation processing at the transmitting end will be described with reference to FIGS. 2 and 4.

First, it is confirmed whether the data body to be transmitted exists in the communication data memory 202 (step 40).
1). If there is a data body part to be transmitted, the protocol processing part 201 reserves an area for creating a header in the communication data memory 202 (step 402).

The address management unit 2011 stores the own address in the sender address field 111 of the secured header area and the destination address in the receiver address field 112.
Is described (step 403), the packet division and reconstruction processing unit 2014 calculates and obtains the length of the packet, and the result is described in the length field 113 (step 404).

Next, the context management unit 2012 confirms whether or not the communication between the sender and the receiver has already started (step 405). This is confirmed by comparing the sender / receiver of each piece of context information already recorded in the context memory 206 with the sender / receiver who is about to communicate.

If the communication has not been started yet (that is, if the matching context cannot be found), the context management unit 2012 sets the synchronization flag 1141 in the control flag area 114 (step 406) and stores the context information at the same time. An area is secured in the context memory 206 and the sender / receiver address is recorded there (step 40
7). Sync flag 114 if communication has already started
1 cannot stand.

Further, the context management unit 2012 confirms whether or not the currently created packet is the last one of the data requested by the sender to be transmitted (step 40).
8). If it is the last one, the context management unit 2012 sets the end flag 1142 in the control flag area 114 (step 409). After that, the context information relating to the current communication is deleted from the context memory 206 (step 410).

Next, a sequence number is assigned to this data by the sequence number management unit 2103 and written in the sequence number field 115 (step 411). This sequence number allocation method will be described later.

Finally, the checksum calculation unit 2015 obtains the checksum of the header, and the obtained value is described in the checksum field 117 (step 412).

Since the header is completed as described above, the header and the data body portion are combined, passed to the network interface 204, and transmitted to the network (step 413).

The above process is repeated while there is data to be transmitted, and if there is an instruction from the sender to end the transmission, the transmission is terminated.

FIG. 5 shows an algorithm for the sequence number management unit 2103 to generate a sequence number in the transmission processing of the first embodiment of the present invention.

In this embodiment, it is assumed that a sequence number is assigned for each 1-byte data. That is, it is confirmed whether or not the packet currently being created is the first data to be communicated in the current communication context (step 501), and if so, the initial value of the sequence number is set to the sequence number management unit 2013. Are generated (step 502). Otherwise, the sequence number of the data to be transmitted next recorded by the context management unit 2012 is taken out (step 503). The sequence number thus obtained is described in the sequence number area 115 of the header (step 504). Then, the context management unit 2012 updates the current context information with a value obtained by adding the number of bytes of the data body part to be transmitted to this sequence number as the sequence number of the data to be transmitted next (step 505).

When a sequence number is assigned in byte units,
Since it is too wasteful to describe the sequence numbers of all the data contained in a packet in the packet, as shown in the above example, generally only the sequence number of the first data contained in the packet is written in the packet. Describe.

FIG. 6 shows an algorithm of the receiving process in the first embodiment of the present invention. The reception process will be described with reference to FIGS. 2 and 6.

Upon receiving the packet from the network, the network interface 204 stores it in the communication data memory 202 (step 601).

The checksum calculation unit 2015 calculates the checksum of the header of the stored packet, and confirms whether this value matches the value described in the checksum field 117 of the packet header (step 60).
2). If they do not match, the header of this packet has an error and cannot be processed, so this packet is discarded (step 603).

Next, the address management unit 2011 extracts the sender address and the recipient address from the packet header and notifies the context management unit 2012 (step 60).
Four). On the other hand, the sequence number management unit 2013 extracts the sequence number from the sequence number field 115 of the packet header and notifies it to the context management unit (step 605). Further, the total length of the packet is taken out from the length field 113, the length of the header is subtracted from this to obtain the length of the data body part, and the result is notified to the context management part 2012 (step 606).

Next, the context management unit 2011 checks the synchronization flag 1141 in the control flag field 114 (step 6).
07).

If the synchronization flag 1141 is set, it means that a new context is started, and therefore an area for new context information is secured in the context memory (step 60).
8), the sender / receiver address notified from the address management unit 2011 is recorded (step 609). Then, the sequence number management unit 2013 is notified that the new context is started (step 610).

If the synchronization flag 1141 is not set, this packet is a packet related to any of the existing contexts. Therefore, the context management unit 2012 uses the sender / receiver address notified from the address management unit 2011 to obtain the context information. A search is performed to find the relevant context (step 611). The sequence number to be received next is extracted from the obtained corresponding context information (step
612), and notifies the sequence number management unit 2013 (step 61).
3).

The sequence number management unit 2013 compares the sequence number expected to be received next, with the sequence number in the sequence number field 115 in the packet header (step 614).
However, if it is the start of a new context, this comparison is not performed, and the two processes are unconditionally determined to be the same and the process proceeds to the next step.

If both match, the data contained in this packet is continuous with the data received immediately before in this context, so the data part is taken out and sent to the receiver via the host interface 205. Pass the data (step 615). Then, a value obtained by adding the length of the data to the sequence number of the packet just received is obtained as the sequence number to be received next, and this is notified to the context management unit 2012 (step 616). The context management unit 2012 records the notified sequence number to be received next in the context information (step 617).

On the other hand, if they do not match, the following processing is performed. Packet sequence number field 11
Check if the value of 5 is greater than the next sequence number to be received in the context (step
618). If it is large, it means that the packet on the way is delayed or dropped. Therefore,
In this case, the received data cannot be passed to the receiver yet, so it is held in the communication data memory (step 619). Then, the product obtained by adding the length of the data part to the sequence number of the packet just received is notified to the context management unit 2012 as the sequence number to be received next (step 610). At the same time, the first sequence number of the data in the “hole” portion of the data that has not been received and the length of the hole are obtained and similarly notified to the context management unit 2012 (step 611). The context management unit 2012 records these values as context information (step 612).
On the other hand, if the value of the sequence number field 115 of the packet is smaller than the sequence number to be received next, it is checked whether or not there is a "hole" in the data (step 62).
3). If there is no hole, the packet is duplicated. Therefore,
Such packets are discarded (step 624).

Similarly, the sequence number field 115 of the packet.
Is smaller than the sequence number to be received next, but if there is a "hole" in the data, it is judged whether or not the data just received is contained in that hole (step 62).
Five). If not included, the packet is discarded because it is data duplication (step 624).

If it is included in the hole, the hole processing described later is performed (step 625). Next, it is checked whether or not the end flag 1142 is set (step 627), and if it is set, it is confirmed whether or not the processing for all remaining holes is completed (step 628). If the processing of the hole is not completed, there may be a packet that arrives late, so the process returns to the beginning of the packet reception processing.

When the hole processing is completed, the post context management unit 2012 deletes the context information from the context memory 206 (step 629).

FIG. 7 shows a processing algorithm at the receiving end side in the first embodiment of the present invention when data corresponding to the aforementioned hole is received. This process will be described with reference to FIGS. 2 and 7.

It is determined whether the data in the received packet is contained in a hole, which results in a continuation of the data already passed to the recipient (step 70).
1). If there is a continuous part, the part is given to the recipient (step 702).

After that, since the state of the hole (the initial sequence number and size) has changed depending on the data received,
The context management unit 2012 is notified of the new hole information (step 703) and the context information is updated (step 704).

As a result, it is confirmed whether or not the hole has disappeared (step 705). If the hole has disappeared, the processing of the hole portion ends.

If holes still remain, the timer added to the first hole is confirmed (step 706). This timer is used to determine if the packet in the hole is late or lost and will not be reached (there are a wide variety of algorithms to determine what the initial value of the timer should be). However, since the algorithm itself is not essential in this embodiment, detailed description thereof will be omitted).
The value of the timer is decremented at a constant pace until the packet containing the data for filling the hole arrives, and the packet of the hole portion which becomes 0 is judged to be lost.

If this timer is not 0, the processing of the hole ends. If the timer is set to 0, the data division reconstruction processing unit 2014 extracts the complementary data from the complementary data memory 203 (step 707), and based on this, creates complementary data corresponding to the length of the first hole. (Step 70
8).

As a result, the first hole portion is filled, so that the data immediately following this hole is received until the end of the data currently held in the communication data memory 202 or the start of the next hole. Hand over to person (Step 71
0).

Then, the context management unit 2013 is notified of new information regarding the hole changed by the complementary work (step 710), and the context information is updated (step 71).
1).

Then, it is confirmed whether or not there are any holes remaining (step 712). If any holes remain, the timer of the first hole among the remaining holes is initialized and started (step 713).

FIG. 8 shows a processing algorithm when the data corresponding to the above-mentioned hole is received on the receiving end side in the second embodiment of the present invention. This process will be described with reference to FIGS. 2 and 8. The configuration of the apparatus of this embodiment is the same as the apparatus of the first embodiment except that the complementary data memory 203 is removed, and the processing other than the processing of the hole shown below is the same as that of the first embodiment. ..

It is checked whether the data in the received packet is contained in a hole, which causes a continuous part with the data already passed to the receiver (step 80).
1). If there is a continuous part, the part is passed to the receiver (step 802).

After that, since the state of the hole (the initial sequence number and size) has changed depending on the data received,
The context management unit 2012 is notified of the new hole information (step 803) and the context information is updated (step 804).

As a result, it is confirmed whether or not the hole is eliminated (step 805), and if it is eliminated, the processing of the hole portion is finished.

If holes still remain, the timer added to the first hole is confirmed (step 806).

If this timer is not 0, the processing of the hole ends. If the timer is set to 0, the data division reconstruction processing unit 2014 notifies the receiver of the first sequence number of the first hole and the length of the hole (step 807). From the notified hole sequence number and length, the recipient can use data before and after the hole to generate any appropriate supplementary data as necessary and supplement it.

The data following the notified hole is passed to the receiver until the end of the data currently held in the communication data memory 202 or the start of the next hole (step 808).

Then, the context management unit 2013 is notified of new information regarding the hole changed by the above work (step 809), and the context information is updated (step 81).
0).

Then, it is confirmed whether or not any holes remain (step 811). If any holes remain, the timer of the first hole among the remaining holes is initialized and started (step 812).

FIG. 9 shows an algorithm of the receiving process in the third embodiment of the present invention. The configuration of the apparatus of this embodiment is the same as that of the first embodiment, and the transmission processing is also the same as that of the first and second embodiments. Here figure
The reception process will be described with reference to 2 and FIG.

Upon receiving the packet from the network, the network interface 204 stores it in the communication data memory 202 (step 901).

The checksum calculation unit 2015 calculates the checksum of the stored packet header, and confirms whether this value matches the value described in the checksum field 117 of the packet header (step 90).
2). If they do not match, the header of this packet has an error and cannot be processed, so this packet is discarded (step 903).

Next, the address management unit 2011 extracts the sender address and the recipient address from the packet header and notifies the context management unit 2012 (step 90).
Four). On the other hand, the sequence number management unit 2013 extracts the sequence number from the sequence number field 115 of the packet header and notifies it to the context management unit (step 905). Also, the total length of the packet is extracted from the length field 113, the length of the header is subtracted from this to obtain the length of the data body part, and the result is notified to the context management part 2012 (step 906).

Next, the context management unit 2011 checks the synchronization flag 1141 in the control flag field 114 (step 9
07).

If the synchronization flag 1141 is set, it means that a new context has started, and therefore an area for new context information is secured in the context memory (step 90
8), the sender / receiver address notified from the address management unit 2011 is recorded (step 909). Then, the sequence number management unit 2013 is notified that the new context is started (step 910).

If the synchronization flag 1141 is not set, this packet is a packet related to one of the existing contexts. Therefore, the context management unit 2012 uses the sender / receiver address notified from the address management unit 2011 to obtain the context information. Search and find the relevant context (step 911). The sequence number to be received next is extracted from the obtained corresponding context information (step
912), and notifies the sequence number management unit 2013 (step 91).
3).

The sequence number management unit 2013 compares the sequence number expected to be received next, with the sequence number in the sequence number field 115 in the packet header (step 914).
However, if it is the start of a new context, this comparison is not performed, and the two processes are unconditionally determined to be the same and the process proceeds to the next step.

If they match, the data contained in this packet is continuous with the data received immediately before in this context, so the data part is taken out and sent to the receiver via the host interface 205. Pass the data (step 915). Then, a value obtained by adding the length of the data to the sequence number of the packet just received is obtained as the sequence number to be received next, and this is notified to the context management unit 2012 (step 916). The context management unit 2012 records the notified sequence number to be received next in the context information (step 917).

On the other hand, if they do not match, the following processing is performed. Packet sequence number field 11
Check if the value of 5 is greater than the next sequence number to be received in the context (step
918). If it is large, it means that the packet on the way is delayed or lost. Therefore,
In this case, the data division reconstruction processing unit 2014 extracts the complementary data from the complementary data memory 203 for the data that complements the data of the missing part, and generates the complementary data corresponding to the length of the missing part based on this. Step 919). Then, the generated complementary data and the data of the packet just received are passed to the receiver via the host interface 205 (step 920).

The sequence number management unit 2013 obtains a new sequence number to be received next by adding the length of the data body portion of the packet to the sequence number of the packet just received, and notifies the context management unit 2012 (step 921). The context management unit 2012 records the notified new sequence number to be received next in the context information (step 92).
2).

Therefore, in this embodiment, there is no "hole" in the data as seen in the first and second embodiments.

On the other hand, the packet sequence number field 115
If the value of is smaller than the sequence number to be received next, it means that the packet is duplicated and such packet is discarded (step 923). Eventually, not only packets lost completely on the way but also packets delayed in order by one packet will be discarded. Therefore, in this embodiment, hole management and packet delay are delayed. Since it is not necessary to manage the timer for determining whether it has been lost or lost, the processing speed can be further increased. In addition, the fact that there is no time to wait for late arrival packets also leads to higher speed. However, there is a possibility that the part that uses the complementary data may increase, which may be a disadvantage in terms of communication quality. Therefore, this embodiment is particularly effective on a highly reliable network.

FIG. 10 shows the algorithm of the receiving process in the fourth embodiment of the present invention. The reception process will be described with reference to FIGS. 2 and 10.

Upon receiving the packet from the network, the network interface 204 stores it in the communication data memory 202 (step 1001).

The checksum calculator 2015 calculates the checksum of the stored packet header, and checks whether this value matches the value described in the checksum field 117 of the packet header (step 100).
2). If they do not match, the header of this packet has an error and cannot be processed, so this packet is discarded (step 1003).

Then, the address management unit 2011 extracts the sender address and the recipient address from the packet header and notifies the context management unit 2012 (step 100).
Four). On the other hand, the sequence number management unit 2013 extracts the sequence number from the sequence number field 115 of the packet header and notifies the context management unit (step 1005). Also, the total length of the packet is extracted from the length field 113, the length of the header is subtracted from this to obtain the length of the data body portion, and the result is notified to the context management unit 2012 (step 1006).
Next, the context management unit 2011 displays the control flag field.
The sync flag 1141 of 114 is checked (step 1007).

If the synchronization flag 1141 is set, it means that a new context is started, and therefore an area for new context information is secured in the context memory (step 100
8), the sender / receiver address notified from the address management unit 2011 is recorded (step 1009). Then, the sequence number management unit 2013 is notified that it is the start of a new context (step 1010).

If the synchronization flag 1141 is not set, this packet is a packet related to any of the existing contexts. Therefore, the context management unit 2012 uses the sender / receiver address notified from the address management unit 2011 to obtain the context information. Search and find the relevant context (step 1011). The sequence number to be received next is extracted from the obtained corresponding context information (step
1012), and notifies the sequence number management unit 2013 (step 101).
3).

The sequence number management unit 2013 compares the sequence number that is expected to be received next, with the sequence number of the sequence number field 115 in the packet header (step 101).
Four). However, if it is the start of a new context, this comparison is not performed, and the two processes are unconditionally determined to be the same and the process proceeds to the next step.

If they match, the data contained in this packet is continuous with the data received immediately before in this context, so the data part is taken out and sent to the receiver via the host interface 205. Pass the data (step 1015). Then, a value obtained by adding the length of the data to the sequence number of the packet just received is obtained as the sequence number to be received next, and this is notified to the context management unit 2012 (step 1016). The context management unit 2012 records the notified sequence number to be received next in the context information (step 1017).

On the other hand, if they do not match, the following processing is performed. Packet sequence number field 11
Check if the value of 5 is greater than the next sequence number to be received in the context (step
1018). If it is large, it means that the packet on the way is delayed or lost. Therefore,
In this case, the data division / reconstruction processing unit 2014 obtains the first sequence number of the data of the missing part and the length of the part (step 1019), and notifies this value to the receiver (step 10).
20). Then, the data body part of the packet just received is passed to the receiver (step 1021).

The sequence number management unit 2013 adds the length of the data body part of the packet to the sequence number of the packet just received to obtain a new sequence number to be received next, and notifies the context management unit 2012 (step 1022). The context management unit 2012 records the notified new order number to be received next in the context information (step 102).
3).

Therefore, in this embodiment, there is no "hole" in the data as in the third embodiment. On the other hand, if the value of the sequence number field 115 of the packet is smaller than the sequence number to be received next, it means that the packet is a duplicate and such packet is discarded (step 1024).

[0106]

As described above, according to the present invention,
While using communication efficiently using packets and cells without occupying communication resources, high-speed communication can be realized by lightweight protocol processing and the complementary function of lost data and late arrival data. Real-time data communication such as video and audio can be efficiently realized in situations where high quality is not required.

[Brief description of drawings]

FIG. 1 is a diagram showing a packet format according to first to fourth embodiments of the present invention.

FIG. 2 is a configuration diagram of a communication processing device according to first and third embodiments of the present invention.

FIG. 3 is a diagram showing a generation processing algorithm of a data body part in the first to fourth embodiments of the present invention.

FIG. 4 is a diagram showing a header part generation processing algorithm in the first to fourth embodiments of the present invention.

FIG. 5 is a diagram showing a sequence number generation algorithm in the first to fourth embodiments of the present invention.

FIG. 6 is a diagram showing a reception processing algorithm in the first embodiment of the present invention.

FIG. 7 is a diagram showing a reception processing algorithm of hole data according to the first embodiment of the present invention.

FIG. 8 is a diagram showing a reception processing algorithm of hole data according to the second embodiment of the present invention.

FIG. 9 is a diagram showing a reception processing algorithm in the third embodiment of the present invention.

FIG. 10 is a diagram showing a reception processing algorithm in the fourth embodiment of the present invention.

[Explanation of symbols]

 11 Header part 12 Body part 111 Sender address field 112 Recipient address field 113 Packet length field 114 Control flag field 1141 Sync flag 1142 End flag 115 Sequence number field 116 Unused field 117 Checksum field 201 Protocol processor 2011 address Management unit 2012 Context management unit 2013 Sequence number management unit 2014 Data division reconfiguration processing unit 2015 Checksum operation unit 202 Communication data memory 203 Complementary data memory 204 Network interface 205 Host interface 206 Context memory

Claims (4)

[Claims] 1. In communication of continuous data, a sequence number is given to each data or a group of data having a certain unit size, and a transmitting device notifies the receiving end of this sequence number. Data sequence monitoring means for monitoring data loss and order change using the sequence number notified from the end, and the part of the data lost due to the data loss is received at the receiving end with predetermined complementary data. The communication processing device is characterized by having a missing data complementing means for independently complementing and an order reconstructing means for restoring the order of the data whose order has been changed. 2. A communication device in which a sequence number is given to each data or a group of data having a certain unit size in the communication of continuous data, and the transmitting end notifies the receiving end of this sequence number. , A data sequence monitoring means that uses the sequence number notified from the sending end to monitor for data loss and order change, and a missing number that notifies the receiver of only the sequence number of the data lost due to data loss A communication processing apparatus comprising: a data sequence number notifying means; and an order reconstructing means for restoring the order of data whose order has been changed. 3. A communication device in which a sequence number is given to each data or a group of data having a certain unit size when communicating continuous data, and the transmitting end notifies the receiving end of this sequence number. , A data sequence monitoring means for monitoring data loss or order change using the sequence number notified from the sending end, and a sequence prior to the order of the data already received by the receiving end due to the order change. When the data is received, the inconsistent order data discarding means for immediately discarding this data, and the part of the data lost due to the loss of data or the discarding of the data is received at the receiving end by the predetermined supplementary data. Has a missing data complementing means for independently complementing the communication processing equipment. 4. A communication device in which a sequence number is given to each data or a group of data having a certain unit size when communicating continuous data, and the transmitting end notifies the receiving end of this sequence number. , A data sequence monitoring means for monitoring data loss or order change using the sequence number notified from the sending end, and a sequence prior to the order of the data already received by the receiving end due to the order change. When the data of is received, this data is immediately discarded, and the inconsistent data discarding means and the missing data sequence that notifies the receiver of only the sequence number of the data lost due to the loss of data or the discarding of the data A communication processing device, comprising: a number notifying unit.