JP4273159B2 - Network device and congestion detection method - Google Patents

Description

  The present invention relates to a network device and a congestion detection method for detecting a congestion state of a network through which packet data flows.

  Conventionally, the timing at which packet data is transmitted from the network device to the opposite device between the transmission source network device to which the packet data is transmitted and the opposite device from which the packet data transmitted from the network device is received is A control signal such as back pressure indicating whether or not packet data transmitted from the network device can be received by the opposite device is transmitted from the opposite device to the network device, and is controlled based on the transmitted control signal. In other words, when packet data cannot be received by the opposite device, for example, packet data stays in the opposite device due to a state where transmission to the device provided in the next stage of the opposite device is not possible. In some cases, the network device is determined to be in a congested state and is notified to the network device that reception is impossible.

  Various methods are considered to detect the congestion state of the network through which the packet data flows. As one of them, a method of detecting a congestion state based on the staying state of packet data is used. This is a calculation method using a buffer usage rate indicating how much packet data is retained in a buffer provided in a network device connected to the network. In this method, when packet data exceeding a preset threshold value is stored in a buffer, it is generally used that determines a congestion state and notifies the host device.

  However, in this method, when packet data is transferred in bursts over the network, the amount of packet data received in the network device in the buffer may momentarily exceed the threshold value. It is conceivable and cannot be said to be a suitable method for judging a steady congestion state.

Therefore, the time when the packet data is transmitted from the source node connected to the network is embedded in the packet data as a time stamp, and the time stamp is extracted and extracted at the destination node where the packet data is received via the network. By comparing the time stamp, that is, the transmission time and the current time, that is, the reception time, the time from when the packet data is transmitted from the transmission source node to reception at the transmission destination node is calculated, A method of determining a congestion state of a network has been considered (for example, see Patent Documents 1 and 2).
JP 2005-269364 A JP 05-007224 A

  However, the methods described in Patent Documents 1 and 2 have a problem that the transmission time must be embedded as a time stamp in the packet data transmitted by the transmission source node. There is also a problem that a time stamp must be extracted from packet data received at the destination node. In addition, there is a problem that the time stamp extracted at the transmission destination node must be compared with the reception time. Furthermore, there is a problem that the congestion state of the network cannot be determined unless the source node and destination node having the above-described means are combined.

  The present invention has been made in view of the problems of the conventional techniques as described above, and an object of the present invention is to provide a network device and a congestion detection method that can easily determine the congestion state of the network. To do.

In order to achieve the above object, the present invention provides:
A network device connected to a network and processing packet data flowing over the network,
The received event of the packet data is registered together with an event processing time for executing the event, a difference between the event execution time and the event processing time is calculated as a delay time, and the delay time is calculated. If the delay time calculated last time is greater, it is determined that the network is in a congestion state.

  Also, the received event of the packet data is registered together with an event processing time for transmitting the packet data, and a difference between the time when the packet data is transmitted and the event processing time is calculated as a delay time. It is characterized by.

  In addition, when the delay time is larger than the previously calculated delay time, the counter is incremented, and the network is in a congested state when the counter value exceeds a predetermined threshold. It is judged that there exists.

  Further, the counter is decremented when the delay time is not larger than the previously calculated delay time.

A congestion detection method for detecting a congestion state of a network through which packet data flows,
A process of registering an event of the received packet data together with an event processing time for executing the event;
A process of calculating a difference between the time when the event is executed and the event processing time as a delay time;
And processing for determining that the network is in a congested state when the delay time becomes larger than the previously calculated delay time.

A process of registering an event of the received packet data together with an event processing time for transmitting the packet data;
And a process of calculating a difference between the time when the packet data is transmitted and the event processing time as a delay time.

Further, when the delay time becomes larger than the previously calculated delay time, a process of incrementing a counter for counting the congestion state;
And a process of determining that the network is in a congested state when the value of the counter is equal to or greater than a predetermined threshold value.

  Further, when the delay time is not larger than the previously calculated delay time, the counter is decremented.

  In the present invention configured as described above, the event of the received packet data is registered together with the event processing time for executing the event, and the difference between the time when the event is executed and the event processing time is When the calculated delay time is larger than the previously calculated delay time, it is determined that the network is in a congestion state.

  Thus, it is not necessary to provide a special signal or bit for determining a common congestion state between the transmission source device and the transmission destination device.

  As described above, in the present invention, an event of received packet data is registered together with an event processing time for executing the event, and the difference between the time when the event is executed and the event processing time is set as a delay time. When the calculated delay time becomes larger than the previously calculated delay time, the network is determined to be in a congested state, so that the network congestion state can be easily determined.

  Embodiments of the present invention will be described below with reference to the drawings.

  FIG. 1 is a diagram showing an embodiment of a network device of the present invention in a wireless communication system.

  As shown in FIG. 1, the present embodiment includes a core network 101, a radio base station control apparatus 102, and radio base stations 103-1 to 103-n. The radio base stations 103-1 to 103-n transmit and receive radio signals to and from a radio terminal (not shown) possessed by a user who uses radio communication. The radio base station control apparatus 102 controls the radio base stations 103-1 to 103-n. The core network 101 is a network to which the radio base station control device 102 is connected. Here, the network apparatus of the present invention may be the radio base station control apparatus 102, the radio base stations 103-1 to 103-n, or both.

  FIG. 2 is a diagram illustrating a configuration example of the network apparatus according to the present invention.

  As shown in FIG. 2, this configuration includes a time count unit 211, an event management unit 212, a congestion detection unit 213, an event table 214, an event registration unit 215, an event execution unit 216, a reception buffer 217, The network device 201 includes a reception processing unit 218, a transmission processing unit 219, and a transmission buffer 220, and is connected between the signal generation device 202 that generates a clock signal that is an operation source and the network device 201. It is connected to the opposite device 203 that transmits and receives packet data.

  The time count unit 211 receives the clock signal generated by the signal generation device 202 and updates the time Mt_time based on the received clock signal. Here, Mt_time is a periodically updated time used as the master time in the network apparatus 201, and its unit is about 1/1000 [sec].

  The event management unit 212 acquires Mt_time updated by the time counting unit 211 and updates Ev_time, which is the event processing time in the network device 201. If the event cannot be processed at the time when the event should be processed, the Ev_time is updated when the Mt_time is updated while all the registered events have been processed. That is, when a transmission control signal such as back pressure is transmitted from the next-stage device and the transmission control signal is received by the network device 201 (during congestion), the event cannot be processed, and thus Ev_time is not updated. . Also, during non-congestion, since all registered events are completed in time, Mt_time = Ev_time is always established.

  The congestion detection unit 213 compares the Mt_time updated by the time counting unit 211 with the Ev_time updated by the event management unit 212, and detects the congestion state of the network.

  The reception buffer 217 receives the packet data transmitted from the opposite device 203 and temporarily stores it.

  The reception processing unit 218 performs predetermined processing such as error detection on the packet data temporarily stored in the reception buffer 217.

  The transmission processing unit 219 performs transmission processing such as adding a header for transmitting the packet data processed by the reception processing unit 218 to a transmission destination device (not shown).

  The event registration unit 215 registers the transmission event of the packet data processed by the transmission processing unit 219 in the event table 214.

  FIG. 3 is a diagram showing a configuration example of the event table 214 shown in FIG. 2 in which transmission events are registered.

  As shown in FIG. 3, the event table 214 shown in FIG. 2 stores packet data events to be transmitted in association with registration times. The registration time here is Ev_time described above.

  The event execution unit 216 executes the transmission event registered in the event table 214 based on Ev_time updated by the event management unit 212.

  The transmission buffer 220 temporarily stores packet data that has been subjected to transmission processing by the transmission processing unit 219.

  Hereinafter, a congestion detection method in the network apparatus configured as described above will be described.

  FIG. 4 is a flowchart for explaining a congestion detection method in the network device 201 shown in FIG. Here, FIG. 4 is a flowchart for mainly explaining a congestion detection method in the congestion detection unit 213.

  When the packet data transmitted from the opposite device 203 is stored in the reception buffer 217 of the network device 201, the reception processing unit 218 performs predetermined processing such as error detection.

  The packet data subjected to predetermined processing such as error detection in the reception processing unit 218 is subjected to transmission processing such as header addition in the transmission processing unit 219, and at the same time, the event registration unit 215 performs a transmission event of the packet data. Is registered in the event table 214. Here, as in the example shown in FIG. 3, events A to D are registered at time “100”, events E to G are registered at time “101”, and events H to K are further registered at time “100”. 102 ”. The relationship between the event and the registration time is based on a conventional technique, and is set based on, for example, a preset transmission speed. As a further specific example, if a transmission rate for transmitting a predetermined number of packet data, for example, three packet data, is set at a certain time interval, for example, 2 ms, the packet data is registered in the same unit in units of three. It is time.

  Thereafter, the congestion detection unit 213 compares Mt_time updated by the time counting unit 211 and Ev_time of the event registered in the event table 214 (step S1).

  If it is determined that Mt_time is not greater than Ev_time, “delay value” (described later) is cleared (step S2).

  FIG. 5 is a diagram illustrating a transition state between Mt_time and Ev_time when it is determined that Mt_time is not larger than Ev_time.

  As shown in FIG. 5, since all the events to be executed at each time are completed within the time, there is an idle time that is not processed at each time. Since the update of Mt_time is in the idle time, Ev_time is updated simultaneously with the update of Mt_time. In this way, Mt_time = Ev_time is always satisfied when there is no congestion.

  FIG. 6 is a diagram illustrating the relationship between the execution of each event and the time during non-congestion.

  As shown in FIG. 6, the events A to D registered with Ev_time “100” are executed when the Mt_time is “100”, and the events E to G registered with the Ev_time “101” are executed. Are executed when Mt_time is “101”, and events H to K registered when Ev_time is “102” are executed when Mt_time is “102”, and always Mt_time = Ev_time. An event is being executed.

  On the other hand, when it is determined that Mt_time is larger than Ev_time, the delay of Ev_time is calculated (step S3). This is a state in which events are not processed in the registered Ev_time. When this delay time is “delay”, it is calculated by delay = Mt_time−Ev_time.

  7 and 8 are diagrams illustrating a transition state between Mt_time and Ev_time when it is determined that Mt_time is larger than Ev_time, that is, when an event execution delay occurs.

  As shown in FIG. 7, since the event has not been processed in time at time 101, the execution time Ev_time of the event at time 102 is delayed with respect to Mt_time. In FIG. 7, since the event processing at time 102 and time 103 is completed within the unit time, Ev_time catches up with Mt_time at time 105.

  On the other hand, as shown in FIG. 8, when event processing cannot keep up, that is, in a congested state, a delay of Ev_time occurs at time 101, and event processing at time 102 is also performed within a unit time. Since it has not been completed, the delay at time 103 is large. Furthermore, since the processing is not in time after time 104, the difference between Mt_time and Ev_time increases with the passage of time. Here, when the difference between Mt_time and Ev_time increases as Mt_time elapses, it is determined that the congestion state is present.

  FIG. 9 is a diagram showing the relationship between the execution of each event and the time at the time of congestion.

  As shown in FIG. 9, among the events A to H registered with Ev_time “100”, the events F to H cannot be executed when the Mt_time is “100”, and the next event execution time Mt_time is It is executed at the time of “101”. The events I to K registered with Ev_time “101” are affected by the execution delay of the events F to H, and the event K cannot be executed when the Mt_time is “101”. It is executed when the execution time Mt_time is “102”.

  When the delay of Ev_time is calculated in step S3, it is determined whether or not the delay of Ev_time is larger than the “delay value” (step S4). The “delayed value of delay” is a delay value calculated last time. That is, it is determined whether or not the delay of Ev_time is larger than the past by accumulating the delay of Ev_time.

  If it is determined that the Ev_time delay is greater than the “delay stored value”, the current Ev_time delay is updated and stored as the “delay stored value” (step S5). As a result, the latest “delay stored value” is used in the next comparison.

  Then, a congestion counter that counts the congestion state is incremented (step S6). This congestion counter is a counter that increases one by one when congestion occurs, and decreases one by one with a minimum of 0 when processing in time of non-congestion is in time. It may be provided in the detection unit 213.

  Thereafter, it is determined whether or not the incremented congestion counter value is equal to or greater than a preset threshold value (step S7). This threshold value can be arbitrarily set.

  When it is determined that the value of the congestion counter is equal to or greater than the threshold value, the host device (not shown) and the opposite device 203 are notified of the congestion state (step S8).

  On the other hand, if it is determined that the value of the congestion counter is not greater than or equal to the threshold value, no processing is performed.

  If it is determined in step S4 that the delay of Ev_time is not greater than the “delay stored value”, it is determined whether or not the value of the congestion counter is greater than 0 (step S9).

  If it is determined that the congestion counter value is greater than 0, the congestion counter value is decremented (step S10).

  On the other hand, if it is determined that the value of the congestion counter is not greater than 0, that is, 0, no processing is performed.

  As described above, the congestion state of the network connected to the next stage can be recognized based on the delay state of the processing time of the event to be processed. Furthermore, by using a congestion counter, it is possible to recognize not only a congestion state that occurs in a burst manner but also a steady congestion state.

It is a figure which shows one Embodiment of the network apparatus of this invention in a radio | wireless communications system. It is a figure which shows the example of 1 structure of the network apparatus of this invention. It is a figure which shows one structural example of the event table shown in FIG. 2 with which the transmission event was registered. 3 is a flowchart for explaining a congestion detection method in the network apparatus shown in FIG. 2. It is a figure which shows the mode of a transition between Mt_time and Ev_time when it is judged that Mt_time is not larger than Ev_time. It is a figure which shows the relationship between each event execution at the time of non-congestion, and time. It is a figure which shows the mode of a transition between Mt_time and Ev_time when it is judged that Mt_time is larger than Ev_time, ie, when the execution delay of an event generate | occur | produced. It is a figure which shows the mode of a transition between Mt_time and Ev_time when it is judged that Mt_time is larger than Ev_time, ie, when the execution delay of an event generate | occur | produced. It is a figure which shows the relationship between each event execution at the time of congestion, and time.

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 Core network 102 Wireless base station control apparatus 103-1 to 103-n Wireless base station 201 Network apparatus 202 Signal generator 203 Opposite apparatus 211 Time count part 212 Event management part 213 Congestion detection part 214 Event table 215 Event registration part 216 Event Execution unit 217 reception buffer 218 reception processing unit 219 transmission processing unit 220 transmission buffer

Claims (8)

A network device connected to a network and processing packet data flowing over the network,
The received event of the packet data is registered together with an event processing time for executing the event, a difference between the event execution time and the event processing time is calculated as a delay time, and the delay time is calculated. A network device that determines that the network is congested when the delay time calculated last time is greater. The network device according to claim 1,
The received event of the packet data is registered together with an event processing time for transmitting the packet data, and a difference between the time when the packet data is transmitted and the event processing time is calculated as a delay time. Network device. The network device according to claim 1 or 2,
The counter is incremented when the delay time is greater than the previously calculated delay time, and the network is in a congested state when the counter value exceeds a predetermined threshold A network device characterized by determining. The network device according to claim 3, wherein
The network device, wherein the counter is decremented when the delay time is not greater than the previously calculated delay time. A congestion detection method for detecting a congestion state of a network through which packet data flows,
A process of registering an event of the received packet data together with an event processing time for executing the event;
A process of calculating a difference between the time when the event is executed and the event processing time as a delay time;
A congestion detection method comprising: determining that the network is in a congestion state when the delay time is greater than a previously calculated delay time. The congestion detection method according to claim 5,
Processing for registering an event of the received packet data together with an event processing time for transmitting the packet data;
A congestion detection method comprising: calculating a difference between a time at which the packet data was transmitted and the event processing time as a delay time. In the congestion detection method according to claim 5 or 6,
When the delay time becomes larger than the previously calculated delay time, a process of incrementing a counter that counts the congestion state;
And a process of determining that the network is in a congested state when the value of the counter is equal to or greater than a predetermined threshold value. The congestion detection method according to claim 7,
And a process of decrementing the counter when the delay time is not greater than the previously calculated delay time.

Images