KR100204598B1 - Error detecting and correcting method - Google Patents

Abstract

The present invention relates to an error detection and error control method of a multi-party communication in a transmission protocol, in which a receiver moves to a standby 2 state and a standby 3 state, detects an error message, transmits data recovered to a reception state, The host moves to the standby 1 state and the transmission restoration state, detects an error message, and transmits the recovered data to the transmission or reception state to perform an error correction method. Thus, reliable transmission of multimedia data in a multi- Disclosed is a method and an error correction method for error detection of a multiparty communication in a transmission protocol capable of minimizing duplication of lost data.

Description

Error Detection and Error Correction in Multipoint Communication in Transport Protocol

In particular, the present invention relates to an error detection method and an error correction method for reliable data transmission required for multimedia application programs such as cooperative operation in a multi-user transmission protocol of multimedia computer communication .

A conventional method of reliably transmitting and receiving data in end-to-end communication between a sender and a recipient in a one-to-one communication manner is such that after the recipient receives the data of the sender, And responding to the sender the status of the received data. Error control is a function to detect and resolve errors that occur during data transmission between these senders and receivers. Errors that can occur in the sending and receiving process include the modification of the content of the transmitted packet, the change of the order in which the packets arrive, the arrival of duplicated packets, and the loss of the packets. This error control function has three functions such as error detection, error reporting, and error recovery (correction).

I. Error detection function

This function checks whether the packets arrive in the order in which they were transmitted without modification of the contents of the packet. In order to detect an error, there is a procedure of receiving a message by using information such as a sequence number, a packet length, and a data checksum, checking errors therefrom, and then receiving a message in the next sequence.

II. Error reporting function

Error reporting is a well known method of transmitting Negative Acknowledgment (NACK) as a function to transmit information about the arrival of a wrong packet or a lost packet to the sender. Generally, the conventional protocol does not have such an error reporting function and detects an error due to arrival of a positive acknowledge (hereinafter referred to as PACK) packet in which a receiver performs a reliable response to a sender within a certain period of time. It causes overload in transmission.

III. Error recovery function

The error recovery (correction) function is a function to retransmit and recover the reported error. It uses a method of retransmitting all packets after the errored packet. This is adopted and used in existing protocols because it is simple to implement and can be easily applied to a communication network having a high error rate. However, a modified method is needed in a high-speed communication network environment where the reliability of the communication network is improved and a large amount of data is transmitted in a short time.

The above functions are merely an error control method occurring in an end-to-end communication composed of a sender and a receiver. In an error control method such as a multimedia environment, many-to-many Lt; / RTI > is an inefficient method for solving concurrently occurring errors between one or several senders and multiple recipients in a multi-party communication.

Accordingly, it is an object of the present invention to satisfy requirements for multi-party communication in a multimedia environment by improving reliable transmission / reception of data by using error control methods required in a high-speed information communication environment based on super-speeding of a transmission network .

According to another aspect of the present invention, there is provided a method of detecting an error in a multi-party communication in a transmission protocol, the method comprising: a receiver in a reception state tracking an order number and detecting an error in the received data; State to a waiting state; moving to a waiting state when the receiver of the waiting state detects an error in another host and waiting for a response of an error message; To a normal reception state when the recovered data is received; and if the receiver of the standby 2 state generates an error message when the timer expires, sending the error message to the multicast group and moving to the standby state, If the recipient of state 3 does not respond to the error message from another host, And generating a message.

According to another aspect of the present invention, there is provided a method of correcting an error in a multi-party communication in a transmission protocol, the method comprising: moving to a standby state when a host in a transmission or reception state receives an error message; The host responding to the error message and generating recovered data and activating a timer; moving to a transmit recovery state while waiting for the host running the timer to respond to recovered data at the other host; Transmitting a restored data from the host to the multicast group, and returning to the transmitting or receiving state when the host in the standby state responds to the data recovered from the opposite host.

FIG. 1 is a block diagram for explaining an error detection method according to the present invention; FIG.

FIG. 2 is a block diagram for explaining an error correction method according to the present invention; FIG.

DESCRIPTION OF THE REFERENCE NUMERALS

11: Receive state 12: Wait 2 state

13: wait state 3: state 21: transmit or receive state

22: Wait 1 state 23: Transmission recovery state

A user participating in a communication is composed of a plurality of senders (hereinafter referred to as a sender) and a plurality of receivers (hereinafter referred to as a receiver), and data is transmitted and received between the plurality of senders. During data transmission, there is a case where a lost message is lost, and an error that a connection between a sender and a receiver is disconnected due to a sudden environment change of the system occurs. In order to reliably transmit / receive data, . In the present invention, these processes are classified into an error detection method and an error correction method as follows.

I. Error detection

1) Loss detection method of message

The sender and the receiver recognize the transmitted / received data using the sequence number of the message. Sequence numbers consist of numbers, incremented sequentially for each data packet, and the sender maintains this number independently. The receiver detects the loss by finding a gap between sequence numbers.

The heartbeat message is used to indicate the presence and status of the sender host. The heartbeat message is sent only by the sender host and contains the sequence number of the last transmitted data packet. The heartbeat occurs in the following cases. The heartbeat occurs only when the sender does not transmit data, and its purpose is to let the recipient recognize the error that may have occurred in the data sent by the sender and to indicate the activity of the sender. The usage method is used when the sender pauses for a certain period. When a sender stops data transmission for a certain period of time, the heartbeat is periodically transmitted for a predetermined period until the data transmission is resumed normally.

The means by which the recipient uses the heartbeat is used by the sender to determine that the sender has intentionally inhibited data transmission while the sender has transmitted a heartbeat, while the recipient has a connection with the sender. Although the connection between the sender and the receiver is still set, the time from the transmission of the last data to the start of transmission of the heartbeat message when the sender inhibits the data transmission is determined by the round trip delivery time of the data packet, the amount of NACK, Flow control parameters and constants.

Most of the participants in the conference will send a small amount of messages, which will occasionally transmit during the multicast session, limiting the heartbeat transmission time. These senders must explain that they are no longer the sender after sending N messages for a specific time. Recipients refer to the source reference state internally. If the receiver fails to obtain these messages and requests the sender to improve the state, the sender repeatedly transmits the information. After a lapse of time, the host transitions from sender to receiver, where the receiver's data synchronization is the responsibility of the session and application layer, and the sender no longer sends a heartbeat message.

FIG. 1 is a block diagram illustrating an error detection method according to the present invention. As shown, in the RECEIVING state 11, the receiver tracks the sequence number to detect message delivery errors and invoke a timer. At this time, the detection target is the received data. After the timer expires, the receiver moves to WAIT2 state (12). In this state, the receiver waits in this state for damping to reduce the number of NACKs before sending a NACK to all multicast groups consisting of all senders and receivers for the detected error, . At this time, when the receiver detects an error in another host and receives the transmitted NACK, it moves to the WAIT 3 state 13 and waits for the response message of the error message, that is, the recovery data REPAIR_DATA. Upon receiving the recovery data REPAIR_DATA, To the RECEIVING state (11). However, if the timer expires in the WAIT2 state 12, the receiver generates an error message NACK and sends it to the multicast group and moves to WAIT3 state 13. If there is no response to the error message from the other host (sender or other opponent receiver), i. E. Recovery data (REPAIR_DATA), the receiver goes back to WAIT2 state 12 and re-generates an error message NACK. This forward and back transition between wait 2 (WAIT_2) 12 and wait 3 (WAIT_3) 13 occurs only 5 times. If there is no response to the error message even after the timer of 5 times has elapsed, the host assumes that there is an error between the sender and the receiver or a communication network shared by the host and generates an error condition to the upper application program.

2) Transmission method in case of site failure

The receiver includes the sender's heartbeat message to determine the sender's state if the reception of the data packet is interrupted. To do this, the recipient sends a specific message to the sender, which multicasts the particular message for other hosts that may have the same problem.

If the sender does not detect a retransmission request for a certain period of time, it is necessary to confirm that there are no more recipients to disconnect or listen to the recipients. Therefore, the sender sends a status request to the recipients. One of the receivers performs a response to the status request, and multicasts the response to the entire group on the multicast address, so that the other receivers do not respond.

When the sender receives the status request response once, it determines that at least one receiver exists and receives the data. If the sender does not hear any response, it assumes that the recipient's response is lost or that there are no more recipients in the group.

In order to eliminate the possibility of loss of the response message transmitted by the receiver, the sender repeatedly transmits this status request several times before concluding that there is no receiver in the group.

If the sender recovers from a site failure, the sender may have a new sequence number and the receiver may have an old sequence number, requiring re-synchronization of the sequence number. If the sender sends a new START message with a new sequence number, the receiver receives the START message and forcibly resynchronizes. The sender repeats START 3 times during data transmission, so that the entire receiver receives START. Recipients look at the START message in the header, and when an START message is lost, an error is signaled to the application program.

II. Error correction

Transmission protocols detect and correct errors that occur during data transmission to ensure reliability. This provides reliable data delivery to each participant and each recipient is responsible for predicting loss and destruction of data.

If the sender does not hear anything from the receiver, it assumes that the receiver has received no data. When a receiver is disconnected, the sender can not take any action to correct the error even if there is a retransmission method. The countermeasure is that the recipient resets the "connectivity" to the group and retransmits the data. Since the sender recognizes the lost data only when the receiver requests retransmission of the lost data, error detection is solely the responsibility of the receiver.

The receiver can request retransmission of the necessary data if it finds and confirms one or more damaged data in the message and uses NACK. NACK is transmitted through the IP-multicast group, and the other host sees the request and avoids the retransmission request for the same request using the damping technique.

If the sender can not retransmit, the other host will retransmit the data. A host capable of retransmission must have enough buffer for data to be able to perform the retransmission process.

The use of damping or slotting techniques reduces the likelihood of message congestion occurring as many recipients respond simultaneously and avoids sending a NACK immediately upon detection of lost data in order to reduce congestion, Listen to NACKs. If the receiver detects a NACK (retransmission request), it does not need to send its own NACK. However, if no other NACK is received, or if the missing data complement is insufficient, the local receiver needs to send a NACK. If all the recipients wait for the same period, the damping technique is meaningless. The time intervals recipients wait for are generally different. Each recipient computes a random time value for listening prior to performing the retransmission request. Advantages of damping techniques reduce packet duplication and limit packet saturation of senders and networks.

The data is partially lost or destroyed during the data transfer process. The receiver requests retransmissions for them, and either the sender or the peer-hosts resend them. When the sender does not transmit data for a certain period of time, it sends a heartbeat message instead.

The two models for data retransmission are as follows.

A method by which the sender takes responsibility for retransmission, responding to all retransmissions of data. If for some reason the sender is not able to retransmit, the peer-hosts are responsible for the retransmission.

As a way for all hosts to be responsible for retransmissions, hosts are responsible for retransmitting by listening to all senders, including the sender itself.

Excessive retransmission response has two solutions.

The first method is performed by the sender himself, but the host is responsible for the retransmission request and the peer host does not respond. However, this method should be used selectively. For example, if the sender host is overloaded with other things, it should be used selectively because retransmission delivery is impossible in a timely manner.

The second method is performed by the sender's agent. In this method, the retransmission process is delayed. If the sender has a connectivity problem, the retransmission response of the peer-hosts is delayed. The advantage of having all the hosts with potential retransmission points is that the sender's overload reduction can be handled. This method can provide efficiency to other hosts that handle fast response responses. No other host is overloaded. The retransmission carried out by the host is when the sender is overloaded more than the host. This will be clear from the performance tests and simulations.

If the sender-preferred model is used, the retransmission request will be satisfied even if the sender's retransmission capacity is insufficient. Another way is to let other hosts resend the sender. This host must be the receiver of the sender data to provide this capability and this host can respond to the retransmission request in the same way as the sender host. The sender first provides a retransmission response because it can not avoid a response to retransmission of that data. If for some reason the sender can not respond to the retransmission request, it is done by another host. The question in this case is to include the sender and respond to the retransmission request at the same time. It uses the damping technique used in retransmissions. However, this solves the following questions.

The host can resolve better in the retransmission handling than the sender, and the other peer-hosts can determine a sufficient time to respond to the retransmission request. This can be done with a specific number in a timer or retransmission request.

Even after the peer-hosts retransmit the requested data, many request responses can be solved using damping techniques in a way that the peer-hosts can block them. All peer-hosts that request a retransmission request response wait for some time to listen to respond to requests from other hosts.

When the sender or receiver uses a non-procedural method under special conditions, the receiver does not receive the retransmission after the fifth try, and if the receiver does not hear the heartbeat from the sender during that period, the receiver determines that the sender host is no longer available. That is, the sender is considered to have left the group. In both cases where the recipient does not know the status of the sender to leave the group or the connection to the sender is lost, there is no recipient meeting the retransmission request, or the recipient does not have the desired data. At this time, the transmission protocol notifies the application layer of the occurrence of a reliability error. The application layer decides what action to take. If the recipient has a lost period during that point, the sender leaves the group. The recipient does not pay attention to the withdrawal of the sender. That is, the receiver is arranged to receive data or heartbeat regularly from the sender.

If no message is received from the sender for a specific time, the receiver determines the sender's status. This is important because the recipient must be receiving the most recent data transmitted by the sender at the last listening time from the sender. At this point, it is not known whether or not the sender transmits other data.

The sender's state is obtained by an explicit request. If the sender is still in the multicast group but is a recipient, the last sender host should acknowledge (ACK) that it is no longer the sender. The old sender can no longer hold the last segment of the data, but sends the sender the specified sequence number of the last packet. The receiver then determines the presence or absence of the lost data. When the receiver notifies the application layer, the processing by the sender transition is completed when the data of the sender is no longer valid. This discards the data from all receivers. If the sender is no longer a multicast group (upon withdrawal), the status requester responds to the other host. This assumes that the other hosts still have information about the sender.

A receiver with a data transmission and a sender heartbeat message does not request retransmission termination. This is due to a network error returned to the sender. In this case, the recipient of the opposite party can perform the retransmission from the sender's point of view.

If there is no other recipient listening to the retransmission request, the receiver terminates the error condition and hands it to the application layer.

FIG. 2 is a block diagram illustrating an error correction method according to the present invention.

As shown, the host in the SENDING or RECEIVING state 21 responds to the error request via the other party's host. When receiving an error message, the host goes into the WAIT_1 (WAIT_1) state (22), responds to the error message and provides the correct recovered data (REPAIR_DATA). The WAIT_1 (WAIT_1) state 22 is used by the opposing host for damping to respond to an error message. The host generates the recovered data (REPAIR_DATA) and activates the timer. The host moves to the SEND_REPAIR state 23 in the timer period in which the host waits for the host to respond to the recovered data REPAIR_DATA and then transmits the recovered data REPAIR_DATA to the multicast group do. However, the host returns to the original transmitting or receiving state (SENDING or RECEIVING) state 21 when the opposite host responds with the recovered data (REPAIR_DATA) in the state of WAIT_1 (22).

As described above, according to the present invention, it is possible to transmit reliable multimedia data in a multi-party transmission protocol of computer communication and to minimize the loss of data.

Claims (5)

A first step in which a receiver in a reception state tracks an order number and detects an error in the received data and calls a timer; a second step in which a receiver terminating the called timer moves to a standby state; A third step of moving to a wait state and waiting for a response of an error message when a receiver of the state detects an error in another host; and a third step of, if receiving the recovered data, The receiver of the standby 2 state generates an error message when the timer expires and sends the error message to the multicast group and moves to the standby 3 state; and if the receiver of the standby 3 state receives an error message from another host And a sixth step of moving to the standby 2 state and generating an error message if there is no response Error detecting method of tracking communications. 2. The method of claim 1, wherein the receiver moves to the standby 2 state of the second step detects an error and transmits a passive confirmation message to the multicast group. 2. The method of claim 1, wherein when the receiver moves to the standby 2 state in the second step, the host in the multicast group listens to the passive confirmation messages when waiting in the standby 2 state. Error detection method. The method as claimed in claim 1, wherein the sixth step generates an error message in the host if there is no response to the error message even after a predetermined number of forward / backward transitions have been completed between the standby 2 state and the standby 3 state An error detection method for multi-party communication. A first step of moving to a standby state when a host in a transmitting or receiving state receives an error message; and a second step of responding to an error message, generating a recovered data and activating a timer, A third step in which the host operating the timer moves to a transmission restoration state while waiting for responding to data recovered from the other host; a third step of transmitting data restored to the multicast group from the host in the transmission restoration state; And a fifth step of the host in the standby state returning to the transmitting or receiving state when responding to the data recovered from the other host.