JP2019208130A - Communication apparatus and queue management method - Google Patents

Abstract

To manage a queue so as not to reduce a bandwidth utilization efficiency of an entire communication system.SOLUTION: In an optical communication system, an OLT communicates with one or more ONUs 100. A lower communication unit 110 of the ONU 100 receives upstream data. A queue 132 temporarily holds the upstream data. An MPCP unit 134 requests bandwidth allocation to the OLT on the basis of a data amount of the queue 132 and receives information on an allocated bandwidth in response to the request. A discard rate calculation unit 141 calculates a discard rate on the basis of the data amount of the queue 132. An adjustment unit 142 determines at least one of necessity of partial discard of upstream data and increase and decrease of the discard rate on the basis of a congestion status of a communication system determined by using the requested bandwidth and the allocated bandwidth. An upstream signal processing unit 131 discards a part of the upstream data without holding it in the queue 132 according to the determination of the adjustment unit 142, and transmits the upstream data held in the queue 132 without being discarded to the OLT according to the allocated bandwidth.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a communication device and a queue management method.

  In a passive optical network (PON), an optical subscriber line termination unit (OLT) and an optical subscriber line network unit (ONU) are arranged in an optical fiber distribution network (ODN). Communicate point-to-point or point-to-multipoint via the Distribution Network. Hereinafter, the direction from the ONU to the OLT is referred to as “up”, and the direction from the OLT to the ONU is referred to as “down”.

  In a PON using time division multiplexing (TDM), a downstream optical signal arrives at the ONU accommodated in the OLT via the ODN. The OLT may add ONU-ID, Alloc-ID, or LLID (Logical Link ID), which is an identifier for identifying the acquisition target, to the downlink signal, and whether the downlink signal can be encoded and decoded. The acquisition target may be indicated by. The ONU may discard the downstream optical signal to which an identifier other than the acquisition target identifier is assigned or cannot be decoded normally, if necessary.

  The OLT cannot correctly decode the upstream optical signals when the upstream optical signals that should arrive after being separated in time from a plurality of ONUs by time division multiplexing overlap with each other in time. For this reason, in PON using time division multiplexing, the transmission timing of the upstream optical signal of each ONU is scheduled according to the procedure of Dynamic Bandwidth Assignment (DBA). In dynamic bandwidth allocation, there is a method in which the OLT transmits a GATE message indicating transmission permission to the ONU according to a REPORT message such as DBRu (Dynamic Bandwidth Report upstream) which is a declaration from the ONU to the OLT.

  As a method for avoiding congestion of an upstream optical signal, there are AQM (Active Queue Management) such as transmission suppression from user equipment by back pressure and tail dropping in a priority queue in the ONU or ONU. Tail Dropping includes EPD (Early Packet Discard) and PPD (Partial Packet Discard). Tail Dropping is a congestion avoidance method that discards frames due to queue overflow when the priority queue is full before being input to the priority queue. Other AQMs include RED (Random Early Detection) and WRED (Weighted RED). RED and WRED are congestion avoidance methods that cause discarding before congestion occurs. Since RED and WRED discard frames that arrive earlier in time than tail dropping, it is compatible with traffic congestion control that decelerates traffic during congestion such as the Transport Layer Transmission Control Protocol (TCP). Highly effective and effective.

Toshiyuki Iwata, Yasuhiko Nakanishi, Yoichi Maeda, "Proposal of traffic control method for PON type architecture access", The Institute of Electronics, Information and Communication Engineers, IEICE Transactions B, 2006, Vol. J89-B, No. 5, p.705-719

  However, when a simple AQM that operates using only the upstream queue of the own device as a trigger is applied to the upstream queue of the ONU, the data is discarded based only on the queue information in the own device. There is a problem that the bandwidth utilization efficiency of the network may be reduced.

  In view of the above circumstances, an object of the present invention is to provide a communication device and a queue management method capable of managing a queue so as not to reduce the bandwidth utilization efficiency of the entire communication system.

  One aspect of the present invention is the communication device in a communication system in which one or more communication devices communicate with a higher-level communication device, the reception unit receiving uplink data to be transmitted to the higher-level communication device, and the reception unit receiving A queue that holds the uplink data, an allocation band receiving unit that receives allocation band information from the higher-level communication device, and a discard rate that is a ratio of discarding the uplink data based on a data amount of the uplink data, or Based on the congestion state of the communication system determined using the data amount and the information on the allocated bandwidth from the higher-level communication device, the calculation unit that calculates the discard amount of the uplink data, the partial discard of the uplink data According to the determination of the adjustment unit and at least one of the discard rate calculated by the calculation unit or the increase / decrease of the discard amount, and the reception unit An upstream signal that performs a process of discarding a part of the upstream data that is not retained in the queue and a process of transmitting the upstream data that is retained in the queue without being discarded to the upper communication apparatus according to the allocated bandwidth A processing unit.

  One aspect of the present invention is the communication apparatus described above, wherein the adjustment unit needs to partially discard uplink data when the time or number of times that the communication system is determined to be congested satisfies a predetermined condition. When the time or number of times determined to be satisfied a predetermined condition, or when the amount of data is greater than a predetermined amount, the discard rate or the discard amount calculated by the calculation unit is increased.

  One aspect of the present invention is the communication device described above, in which the adjustment unit is configured to receive an uplink when a predetermined time has elapsed, a time or number of times when the communication system determines that the communication system is not congested, and a predetermined condition is satisfied. The discard rate calculated by the calculation unit, which has been increased when the time or number of times when it is determined that the partial discard of data satisfies the predetermined condition, or when the data amount is smaller than the predetermined amount Alternatively, the amount of waste is reduced.

  One aspect of the present invention is the communication device described above, wherein the adjustment unit determines that the time or number of times that the communication system is determined not to be congested satisfies a predetermined condition, or the data amount is greater than a predetermined value. If the number is smaller, the discard rate or the discard amount calculated by the calculation unit is decreased.

  One aspect of the present invention is the communication device described above, in which the adjustment unit is configured to receive an uplink when a predetermined time elapses, a time or number of times when the communication system determines that the communication system is congested, or a predetermined condition is satisfied. When the time or number of times when it is determined to partially discard data satisfies a predetermined condition, or when the amount of data is larger than a predetermined value, the discard rate calculated by the calculation unit or the decrease rate is decreased. Increase the amount of waste.

  One aspect of the present invention is the above-described communication device, in which the adjustment unit calculates whether the upstream data is partially discarded based on the congestion state statistics determined multiple times. At least one of the discard rate and the increase / decrease in the discard amount is determined.

  One aspect of the present invention is the communication apparatus described above, wherein the adjustment unit determines a congestion state of the communication system using the statistics of the data amount and the statistics of the allocated bandwidth from the higher-level communication apparatus. .

  One aspect of the present invention is a queue management method for the communication device in a communication system in which one or more communication devices communicate with a higher-level communication device, the reception step receiving uplink data to be transmitted to the higher-level communication device, Based on the queuing step for temporarily holding the uplink data received in the receiving step in a queue, the allocated band receiving step for receiving allocated band information from the higher-level communication device, and the data amount of the uplink data, The calculation step for calculating the discard rate that is the rate of discarding uplink data or the discard amount of the uplink data, and the congestion of the communication system determined using the data amount and the information on the allocated bandwidth from the host communication device Necessity of partial discard of upstream data based on the state, the discard rate calculated in the calculation step or the discard An adjustment step for determining at least one of the increase and decrease, a discard step for discarding a part of the uplink data received in the reception step without holding it in the queue according to the determination in the adjustment step, and without discarding A transmission step of transmitting the uplink data held in the queue to the higher-level communication device according to the allocated band.

  According to the present invention, it is possible to manage a queue of a communication system so as not to reduce the bandwidth utilization efficiency of the entire communication system.

1 is a configuration diagram of an optical communication system according to an embodiment of the present invention. It is a functional block diagram of ONU by this embodiment. It is a figure which shows the increase method of the discard rate by 1st Embodiment. It is a functional block diagram of ONU relevant to this embodiment.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
FIG. 1 is a configuration diagram of an optical communication system 1 according to an embodiment of the present invention. In the present embodiment, a case where the optical communication system 1 is a PON (Passive Optical Network) will be described as an example. The optical communication system 1 includes an ONU 100, an OLT 200, and an ODN 300.

  The ONU 100 is an optical network termination device that communicates using optical signals. The ONU 100 temporarily holds upstream data input from lower devices such as the user device 400 in a queue, and outputs the upstream data as an upstream optical signal to the OLT 200 via the ODN 300 in response to transmission permission from the OLT 200. Further, the ONU 100 receives a downstream optical signal from the OLT 200 via the ODN 300, and outputs it as downstream data to a lower-level device such as the user apparatus 400. The ONU 100 may be a part of an apparatus such as FTTB (Fiber to the building).

  The OLT 200 is an optical line termination device. The OLT 200 receives an upstream optical signal from one or more ONUs 100 via the ODN 300, and outputs upstream data included in the upstream optical signal to a host device (not shown). The OLT 200 transmits downlink data input from a host device (not shown) to the ONU 100 via the ODN 300 as a downlink optical signal.

  The ODN 300 transmits an upstream optical signal and a downstream optical signal. When the optical splitter is provided in the ODN 300, the optical splitter joins the upstream optical signals transmitted from each of the one or more ONUs 100 and transmits them to the OLT 200, and branches the downstream optical signal received from the OLT 200 to transmit to each ONU 100. . For example, the optical splitter multiplexes uplink signals transmitted from a plurality of ONUs 100 by a TDMA (Time Division Multiple Access) method and transmits the multiplexed signals to the OLT 200.

  The ONU 100 has an AQM function for an upstream data queue or a priority queue. The AQM function is based on the information of the congestion avoidance parameters such as the queue length, and the data discard rate, for example, the discard rate in each Ethernet (registered trademark) frame, the discard threshold, etc. (hereinafter collectively referred to as “discard rate”). Congestion is avoided by discarding data according to the discard rate.

  Here, the ONU related to the present embodiment will be described. FIG. 4 is a functional block diagram showing the configuration of the ONU 900 related to the present embodiment. The ONU 900 is used in place of, for example, the ONU 100 shown in FIG. The ONU 900 performs congestion avoidance by AQM based only on the queue information in its own device.

  The uplink signal processing unit 931 holds (queues) the frame of the uplink signal output from the UNI (User-Network Interface) port 910 in the queue 932. The uplink signal processing unit 931 outputs queue length information indicating the length of the frame queued in the queue 932 to the MPCP unit 934 and the AQM unit 940. The discard rate calculation unit 941 of the AQM unit 940 outputs the discard rate information indicating the discard rate determined based on the queue length information to the uplink signal processing unit 931. Based on the discard rate information, the upstream signal processing unit 931 discards part of the upstream frame data output from the UNI port 910 without queuing.

  In the SR mode in which SR (Status Reporting) DBA is performed, an MPCP (Multi-Point Control Protocol) unit 934 is based on queue length information indicating the amount of uplink data (hereinafter also referred to as “uplink data amount”). The amount of data for requesting bandwidth allocation is determined and set in the bandwidth request information of the REPORT message. Bandwidth allocation is processing for allocating bandwidth to ONUs. A bandwidth allocated to the ONU by bandwidth allocation is referred to as an allocated bandwidth. Requesting bandwidth allocation is described as a bandwidth request, and a bandwidth requesting allocation is described as a requested bandwidth. The bandwidth request information is information indicating the requested bandwidth. The REPORT message is output from the upstream signal processing unit 931 to the PON-IF (Interface) port 920, converted into an optical signal at the PON-IF port 920, and transmitted to the OLT 200. The OLT 200 transmits a GATE message in response to the reception of the REPORT message, the result of traffic monitoring such as the history of the usage amount or the usage rate of the allocated bandwidth of each ONU 900, or a notification from another device. The GATE message transmitted from the OLT 200 is converted from an optical signal to an electrical signal at the PON-IF port 920 and output from the downstream signal processing unit 933 to the MPCP unit 934. The MPCP unit 934 acquires band allocation information from the GATE message and outputs it to the uplink signal processing unit 931. The bandwidth allocation information is information indicating the bandwidth allocated to the ONU. The upstream signal processing unit 931 reads the upstream signal frame queued in the queue 932 and converts it into a PON frame according to the allocated bandwidth of the ONU 900 indicated by the bandwidth allocation information, and the PON-IF port 920 converts the PON frame into the PON frame. It transmits to OLT200 with an optical signal.

  As described above, the ONU 900 discards the upstream data based only on the queue information of the ONU 900 regardless of the PON congestion state. Therefore, when the PON is not congested, the efficiency of the entire PON may be reduced. is there. Therefore, the ONU 100 according to the present embodiment is included in the bandwidth request information included in the REPORT message or the like transmitted from the ONU 100 to the OLT 200, or the amount of uplink data in the ONU 100, and the GATE message transmitted from the OLT 200 to the ONU 100. Based on the bandwidth allocation information, the upstream congestion state of the entire PON (optical communication system 1) is estimated. The ONU 100 reflects the estimation result in the control of execution or stop of AQM, the control of execution or stop of discard of uplink data, and the determination of the discard rate of the queue, thereby reducing the discard at the time of non-congestion. Improve overall bandwidth utilization efficiency.

  The DBA includes an NSR (Non Status Reporting) DBA and a TM (Traffic monitoring) DBA in addition to the SR DBA in which the ONU notifies the OLT of a necessary bandwidth by a REPORT message or the like. In TM DBA or the like, the upstream data amount of each ONU is grasped by monitoring the upstream data received by the OLT from the ONU. When the PON is in the SR mode, the ONU 100 can use the bandwidth request information included in the REPORT message or the like for the estimation of the upstream congestion state of the entire PON. However, the ONU 100 is a TM mode (NSR mode) that performs TM DBA, When the REPORT message to be notified to the OLT 200 is not generated, the uplink data amount in the ONU 100 is used. Even when a REPORT message or the like to be notified to the OLT 200 in the SR mode or the like is generated, the ONU 100 may use the uplink data amount in the ONU 100.

  When the ONU 100 uses the upstream data amount in the ONU 100 for estimating the upstream congestion state of the entire PON, it is not necessary to wait until the bandwidth request information is generated, so that the processing can be speeded up. Further, there is an advantage that it is possible to use other than the SR DBA that makes a bandwidth request from the ONU 100. The ONU 100 may use the value at the time of bandwidth allocation as the amount of uplink data used for estimation. However, in the case of SR DBA, it is desirable to use the value at the time corresponding to bandwidth allocation. Further, in the case of TM DBA or the like, it is better to use the history and statistical values of the past uplink data amount used for determining the allocated bandwidth.

  Further, when using the bandwidth request information, if the requested bandwidth and the allocated bandwidth generally correspond, it is convenient to use the corresponding bandwidth. Even when bandwidth request information is used, when TM DBA is used together for bandwidth allocation, or when history or statistical values of uplink data amount are used, it is necessary to use a value corresponding to the accuracy of estimating the PON congestion state. It is desirable from the viewpoint of raising

  FIG. 2 is a functional block diagram showing a configuration of the ONU 100 according to the present embodiment, and only functional blocks related to the present embodiment are extracted and shown. The ONU 100 includes a lower communication unit 110, an upper communication unit 120, a PON signal processing unit 130, and a queue management unit 140.

  The lower communication unit 110 transmits / receives data to / from a lower communication device such as the user device 400. The lower communication unit 110 is, for example, the UNI port 910 illustrated in FIG. The lower communication unit 110 outputs the uplink signal received from the user device 400 to the PON signal processing unit 130. Further, the lower-level communication unit 110 transmits the downlink signal output from the PON signal processing unit 130 to the user apparatus 400.

  The upper communication unit 120 transmits and receives optical signals to and from the OLT 200 via the ODN 300. The upper communication unit 120 is, for example, the PON-IF port 920 shown in FIG. The upper communication unit 120 converts the upstream signal output from the PON signal processing unit 130 from an electrical signal to an optical signal, and transmits the optical signal to the OLT 200 via the ODN 300. The upper communication unit 120 converts the downstream optical signal received via the ODN 300 into an electrical signal and outputs the electrical signal to the PON signal processing unit 130.

  The PON signal processing unit 130 includes an upstream signal processing unit 131, a downstream signal processing unit 133, and an MPCP unit 134. The upstream signal processing unit 131 includes a queue 132 that temporarily holds (queues) a frame of the upstream signal. The upstream signal processing unit 131 may include a plurality of queues 132 according to priority. Note that the PON signal processing unit 130 may include a queue 132 outside the upstream signal processing unit 131. The uplink signal processing unit 131 outputs to the MPCP unit 134 and the queue management unit 140 the queue length information indicating the length of the frame queued in the queue 132, that is, the data amount of the queued uplink data. When the queue management unit 140 uses the bandwidth request information output from the MPCP unit 134 as information on the amount of uplink data, the uplink signal processing unit 131 does not output queue length information to the queue management unit 140. Also good. The uplink signal processing unit 131 sets a control signal such as a REPORT message output from the MPCP unit 134 in an uplink signal frame, and outputs the frame to the upper communication unit 120. Further, the uplink signal processing unit 131 discards a part of the frame of the uplink signal output from the lower communication unit 110 according to the final discard rate information instructed from the adjustment unit 142 of the queue management unit 140 described later. The upstream signal processing unit 131 reads the frame queued in the queue 132 according to the allocated bandwidth of the ONU 100 indicated by the bandwidth allocation information notified from the MPCP unit 134, converts it into a PON frame, and outputs it to the higher-level communication unit 120. To do.

  The downlink signal processing unit 133 has the same function as the downlink signal processing unit 933 shown in FIG. The downlink signal processing unit 133 receives the downlink signal transmitted from the OLT 200 from the higher-level communication unit 120, and based on whether the downlink signal can be decoded based on identification information such as Alloc-ID given to the downlink signal, or the like. Get the downstream signal addressed to it. The downlink signal processing unit 133 outputs a GATE message of the control signal set in the acquired downlink signal to the MPCP unit 134. ITU ATM (Automatic Teller Machine) -PON, Broadband PON (BPON), Gigabit Passive Optical Network (G-PON), XG (10Gigabit Capable) -PON, XGS (10Gigabit-capable Symmetric) -PON NG (Next Generation) -PON2 or the like does not receive a frame for each ONU. In the case of the latter four methods, the band allocation information is set to BWMap. Further, the downlink signal processing unit 133 converts the downlink signal addressed to the user apparatus 400 into a frame with the lower apparatus and outputs the frame to the lower communication unit 110.

  The MPCP unit 134 has an MPU function on the ONU side. When requesting a bandwidth, the MPCP unit 134 determines a data amount for requesting bandwidth allocation to the OLT 200 based on the queue length information received from the uplink signal processing unit 131. The amount of data can be converted into the required bandwidth amount using the communication speed, and the data amount can be regarded as the required bandwidth from the ONU 100, that is, the bandwidth that the ONU 100 requests for allocation. The MPCP unit 134 sets bandwidth request information indicating the determined data amount in the REPORT message, and outputs it to the uplink signal processing unit 131. In GPON, XG-PON, XGS-PON, NG-PON2, etc., the bandwidth request information is set to DBRu. Further, the MPCP unit 134 receives a GATE message of the control signal from the downlink signal processing unit 133. The MPCP unit 134 acquires band allocation information indicating a band allocated to the ONU 100 by the OTL 200 from the GATE message and the like, and outputs the band allocation information to the uplink signal processing unit 131. Further, the MPCP unit 134 sends the bandwidth request information set in the REPORT message and the bandwidth allocation information acquired from the GATE message received in response to the REPORT message to the adjustment unit 142 of the queue management unit 140 described later. Output.

  The queue management unit 140 determines the congestion state of the entire PON at every predetermined timing, and manages the queue 132 based on the determination result. The predetermined timing may be, for example, a predetermined cycle or reception of one or more predetermined GATE messages. The queue management unit 140 includes a discard rate calculation unit 141 and an adjustment unit 142.

  The discard rate calculation unit 141 calculates the discard rate by AQM such as RED using the queue length information received from the upstream signal processing unit 131. For example, the discard rate calculator 141 calculates the discard rate in the same manner as the discard rate calculator 941 shown in FIG. The discard rate calculation unit 141 outputs discard rate information indicating the calculated discard rate to the adjustment unit 142.

  Based on the presence or absence of congestion in the entire PON, the adjustment unit 142 determines at least whether or not AQM is to be executed, whether or not AQM is to be discarded, and whether the discard rate or discard amount is calculated by the discard rate calculation unit 141. Decide either. The adjusting unit 142 is based on the discard rate information calculated by the discard rate calculating unit 141 and the bandwidth allocation information from the MPCP unit 134 (a functional unit corresponding to IEEE802.3 or corresponding to ITU-T). These processes may be performed, or may be performed based on the bandwidth allocation information and the bandwidth request information from the MPCP unit 134 (in the case of conforming to IEEE802.3, in the case of conforming to ITU-T, the corresponding functional unit). Good. Use of the bandwidth allocation information and bandwidth request information from the MPCP unit 134 is particularly suitable for determining whether or not AQM execution is necessary based on the congestion state of the entire PON.

  If the adjustment unit 142 determines that the AQM needs to be executed, the adjustment unit 142 executes the AQM for the queue 132. When executing the AQM, the adjustment unit 142 outputs the final discard rate information indicating the result of the increase / decrease to the discard rate calculated by the discard rate calculation unit 141 to the uplink signal processing unit 131 to execute the discard of the uplink frame. . If the discard rate is not increased or decreased, the adjustment unit 142 outputs the discard rate information calculated by the discard rate calculation unit 141 to the upstream signal processing unit 131 as final discard rate information. If the adjustment unit 142 determines that the AQM is not required to be executed, the adjustment unit 142 stops the AQM for the queue 132. If the adjustment unit 142 determines that the AQM discard is unnecessary, the adjustment unit 142 stops the AQM discard for the queue 132.

  Note that when the AQM is stopped, the adjustment unit 142 may stop the AQM itself in the discard rate calculation unit 141 (AQM itself is OFF). The discarding process may be stopped by stopping the output of the final discard rate to the processing unit 131 (AQM discarding process OFF). Further, when executing the AQM, the adjustment unit 142 may start executing the AQM itself that has been stopped by the discard rate calculating unit 141 (AQM itself is ON), and the discard rate calculating unit 141 executes the AQM itself. However, the state in which the output of the final discard rate information to the upstream signal processing unit 131 is stopped may be canceled, and the final discard rate information may be output to the upstream signal processing unit 131 to execute the discarding process. (AQM discard processing ON).

  The adjustment unit 142 may further determine whether or not it is necessary to execute AQM using information indicating the congestion state in the ONU 100. For example, even if the entire PON is not congested, the adjustment unit 142 may determine that AQM needs to be performed when the ONU 100 is congested. The adjustment unit 142 may use, for example, the discard rate information acquired from the discard rate calculation unit 141 or the queue length information as the information indicating the congestion state of the ONU 100.

  With the above configuration, the ONU 100 operates as follows. The upstream signal processing unit 131 holds (queues) the upstream signal frame output from the lower communication unit 110 in the queue 132. The uplink signal processing unit 131 outputs the queue length information to the MPCP unit 134 and the discard rate calculation unit 141. The discard rate calculation unit 141 determines the frame discard rate based on the queue length information, and outputs the discard rate information indicating the determined discard rate information to the adjustment unit 142.

  The MPCP unit 134 determines the data amount for requesting bandwidth allocation based on the queue length information, sets the bandwidth request information indicating the determined data amount in the REPORT message, and outputs it to the uplink signal processing unit 131. The uplink signal processing unit 131 outputs the REPORT message to the higher-level communication unit 120. The upper communication unit 120 transmits a REPORT message to the OLT 200 using an optical signal.

  The upper communication unit 120 receives a GATE message from the OLT 200 in response to the REPORT message, converts the optical signal into an electrical signal, and outputs the electrical signal to the downlink signal processing unit 133. The downlink signal processing unit 133 acquires a GATE message addressed to the own ONU 100 and outputs it to the MPCP unit 134. The MPCP unit 134 acquires the bandwidth allocation information set in the GATE message. The MPCP unit 134 outputs the band allocation information to the uplink signal processing unit 131, and outputs the band request information and the band allocation information to the adjustment unit 142.

  The adjustment unit 142 determines whether or not to execute AQM based on the presence or absence of congestion of the entire PON obtained from the bandwidth request information and the bandwidth allocation information. The adjustment unit 142 performs ON / OFF of the AQM itself or ON / OFF of the discard process by the AQM according to the determination result. When executing the AQM, the adjustment unit 142 increases / decreases the discard rate calculated by the discard rate calculation unit 141 based on the presence / absence of congestion of the entire PON, and sends final discard rate information indicating the increased / decreased discard rate to the upstream signal processing unit It outputs to 131. Alternatively, the adjustment unit 142 may use the discard rate calculated by the discard rate calculation unit 141 as it is as the final discard rate information. Further, the adjustment unit 142 may increase or decrease the discard rate calculated by the discard rate calculation unit 141 based on the presence or absence of congestion of the entire PON without determining whether or not AQM is executed.

The upstream signal processing unit 131 discards a part of the upstream signal frame input from the lower communication unit 110 without queuing the queue 132 according to the final discard rate information, and queues the frame that has not been discarded to the queue 132. To do. The upstream signal processing unit 131 reads the upstream signal frame queued in the queue 132 according to the allocated bandwidth of the ONU 100 indicated by the bandwidth allocation information, converts it into a PON frame, and outputs the frame to the higher-level communication unit 120. The upper communication unit 120 transmits the PON frame to the OLT 200 using an optical signal.
Hereinafter, a specific queue management method by the queue management unit 140 will be described.

(First embodiment)
This embodiment is queue management that combines an increase in the discard rate and AQM stoppage.
Based on the upstream data amount of the ONU 100 set in the bandwidth request information transmitted from the ONU 100 or notified from the upstream signal processing unit 131 and the allocated bandwidth allocated to the ONU 100, the adjustment unit 142 of the ONU 100 Estimate the total required bandwidth. The adjustment unit 142 determines that the entire PON is congested when the estimated required bandwidth of the entire PON exceeds the total of the upstream bandwidths that can be allocated to the ONU 100, and the discard rate calculation unit 141 performs the discard calculated by the AQM. Instruct the upstream signal processing unit 131 to increase the rate. If the requested bandwidth of the entire PON is equal to or less than the sum of the upstream bandwidths that can be allocated to each ONU 100, the adjustment unit 142 determines that the AQM is not necessary and stops the AQM.

  For example, consider a case where the total uplink bandwidth that can be allocated to the ONU 100 is 10 Gbit / s (gigabit per second), and ten ONUs 100 are connected to the OLT 200. Assume that the requested bandwidth of the ONU 100 is 1 Gbit / s, and the allocated bandwidth assigned to the ONU 100 from the OLT 200 is 500 Mbit / s. Further, it is assumed that the DBA mechanism of the OLT 200 is proportional distribution according to the requested bandwidth of each ONU 100. In this case, the adjustment unit 142 finds that 1 Gbit / s, which is the required bandwidth, is (500 Mbit / s) ÷ (10 Gbit / s) = 1/20 of the required bandwidth of the entire PON. It can be estimated that it was 20 Gbit / s. The adjustment unit 142 finds that the entire PON is congested because the requested bandwidth of the entire PON exceeds the total 10 Gbit / s of the upstream bandwidth that can be allocated to the ONU 100. Judgment and increase the discard rate.

  In the REPORT message, the amount of data is set in the bandwidth request information indicating the requested bandwidth, and in the GATE message, a time during which the ONU 100 can communicate (for example, an assigned time slot) is set as the bandwidth allocation information. Since the speed of the upstream optical signal transmitted by the ONU 100 via the ODN 300 is determined in advance, the data amount and the communicable time are converted into the data amount in unit time in the above. The same applies to the following embodiments.

  FIG. 3 is a diagram illustrating a method of increasing the discard rate. 3A to 3D, L1 is the discard rate calculated by the discard rate calculating unit 141, and L2 is the increased discard rate in the adjusting unit 142. In the example shown in the figure, the discard rate calculation unit 141 increases the discard rate as the queue length increases until the discard rate reaches the upper limit value.

  As illustrated in FIGS. 3A and 3B, the adjustment unit 142 may increase the discard rate increment with respect to the queue length change so that the increment increases as the queue length increases. Further, as shown in FIGS. 3C and 3D, the adjustment unit 142 may shift the discard rate so as to increase by a certain amount while maintaining the gradient of the discard rate increase with respect to the queue length. . Further, the adjustment unit 142 may set the upper limit of the discard rate to be the same as the upper limit before the increase of the discard rate as shown in FIGS. 3B and 3D. As shown in 3 (c), it may be set higher than the upper limit before the discard rate is increased. However, the discard rate ranges from 0 to 1.

  The adjustment unit 142 may return or reduce the increased discard rate to a predetermined value, for example, the value before the increase, when any of the following conditions (1a) to (8a) is satisfied. May be. That is, (1a) a predetermined time has elapsed, (2a) it is determined that there is no congestion, (3a) it is determined that there is no congestion for a predetermined time or a predetermined number of times, and (4a) in the determination of a past predetermined time More than a predetermined time or determined in the past a predetermined number of times, it is determined that congestion is not more than a predetermined number of times, (5a) AQM is determined not to be executed, (6a) a predetermined time or a predetermined number of times continuously It is determined that AQM is not required to be executed. (7a) It is determined that AQM is not required to be executed for a predetermined time or more in a predetermined predetermined time in the past or a predetermined number of times in the past. (8a) The condition is that the upstream data amount in the ONU 100 is larger than a predetermined amount.

(Second Embodiment)
In the present embodiment, queue management is a combination of a decrease in the discard rate and AQM start.
The adjustment unit 142 of the ONU 100 estimates the requested bandwidth of the entire PON based on the requested bandwidth set in the bandwidth request information transmitted by the ONU 100 and the allocated bandwidth of the ONU 100 indicated by the bandwidth allocation information. The adjustment unit 142 determines that there is congestion when the estimated required bandwidth of the entire PON exceeds the total uplink bandwidth that can be allocated to the ONU 100, and executes AQM by RED or the like. The adjustment unit 142 sets the discard rate calculated by the discard rate calculation unit 141 by AQM as the discard rate instructing the upstream signal processing unit 131 as it is. The adjustment unit 142 reduces the discard rate calculated by the discard rate calculation unit 141 using the AQM to the uplink signal processing unit 131 if the estimated required bandwidth of the entire PON is equal to or less than the total uplink bandwidth that can be allocated to the ONU 100. Instruct.

  For example, consider a case where the total uplink bandwidth that can be allocated to the ONU 100 is 10 Gbit / s, and ten ONUs 100 are connected to the OLT 200. Assume that the requested bandwidth of the ONU 100 is 1 Gbit / s, and the allocated bandwidth assigned to the ONU 100 from the OLT 200 is 2 Gbit / s. Further, it is assumed that the DBA mechanism of the OLT 200 is proportional distribution according to the requested bandwidth of each ONU 100. In this case, the adjustment unit 142 finds that 1 Gbit / s, which is the requested bandwidth, is (2 Gbit / s) ÷ (10 Gbit / s) = 1/5 of the requested bandwidth of the entire PON. It can be estimated that it was 5 Gbit / s. The adjustment unit 142 finds that the entire PON is not congested because the requested bandwidth of the entire PON is 10 Gbit / s or less of the total uplink bandwidth that can be allocated to the ONU 100. In this case, the adjustment unit 142 subtracts an arbitrary value from the discard rate calculated by the discard rate calculation unit 141 using the AQM, and finally sets the discard rate to instruct the upstream signal processing unit 131 to execute. For example, if the discard rate calculated by the discard rate calculation unit 141 based on the queue length is 0.2, the adjustment unit 142 decreases the value by 0.1 and sets the final discard rate to 0.1. Can do. However, the discard rate ranges from 0 to 1. Further, the adjustment unit 142 can determine the amount of decrease in the discard rate according to the ratio between the requested bandwidth of the ONU 100 and the allocated bandwidth, or the ratio between the requested bandwidth of the ONU 100 and the requested bandwidth of the entire PON.

  A method for reducing the discard rate will be described with reference to FIG. When the discard rate is decreased, L2 is the discard rate calculated by the discard rate calculation unit 141, and L1 is the discard rate decreased by the adjustment unit 142. As shown in FIGS. 3A and 3B, the adjustment unit 142 may decrease the increase in the discard rate with respect to the change in the queue length as the queue length becomes longer. Further, as shown in FIGS. 3C and 3D, the adjustment unit 142 may shift the discard rate so as to decrease by a certain amount while maintaining the gradient of the discard rate increase with respect to the queue length. . Further, the adjustment unit 142 may set the upper limit of the discard rate after the decrease to the same as the upper limit before the decrease of the discard rate as shown in FIGS. 3B and 3D. ) And FIG. 3C, it may be set lower than the upper limit before the discard rate is decreased.

  In addition, the adjustment unit 142 may return the decreased discard rate to a predetermined value, for example, a value before the decrease, or increase when the following conditions (1b) to (8b) are satisfied. You may let them. That is, (1b) a predetermined time has elapsed, (2b) it has been determined that congestion has occurred, (3b) it has been determined that congestion has continued for a predetermined time or a predetermined number of times, and (4b) it has been determined in advance based on the determination of a predetermined time in the past (5b) AQM is determined to be necessary for execution, or (6b) AQM is required to be executed for a predetermined time or continuously for a predetermined number of times. (7b) It is determined that AQM needs to be executed for a predetermined time or more in the past predetermined time determination, or for a predetermined number of times in the past, or (8b) Uplink data in the ONU 100 The condition is that the amount is larger than a predetermined amount.

(Third embodiment)
The present embodiment is queue management for executing and stopping AQM.
The adjustment unit 142 of the ONU 100 compares the requested bandwidth set in the bandwidth request information transmitted by the ONU 100 with the allocated bandwidth of the ONU 100 indicated by the bandwidth allocation information. The adjustment unit 142 determines that the entire PON is congested if the requested bandwidth exceeds the allocated bandwidth, and executes AQM by RED or the like. The adjustment unit 142 sets the discard rate calculated by the discard rate calculation unit 141 by AQM as the discard rate instructing the upstream signal processing unit 131 as it is. If the requested bandwidth is equal to or less than the allocated bandwidth, the adjustment unit 142 determines that the AQM is not necessary because the entire PON is not congested, and stops the AQM.

  For example, consider a case where the total uplink bandwidth that can be allocated to the ONU 100 is 10 Gbit / s, and ten ONUs 100 are connected to the OLT 200. Assume that the requested bandwidth of the ONU 100 is 1 Gbit / s, and the allocated bandwidth assigned to the ONU 100 from the OLT 200 is 500 Mbit / s. In this case, since the requested bandwidth exceeds the allocated bandwidth, the adjustment unit 142 determines that the entire PON is congested and executes AQM. On the other hand, it is assumed that the requested bandwidth of the ONU 100 is 1 Gbit / s and the allocated bandwidth allocated from the OLT 200 is 2 Gbit / s. In this case, the adjustment unit 142 determines that the entire PON is not congested because the requested bandwidth is equal to or less than the allocated bandwidth, and stops AQM.

(Fourth embodiment)
This embodiment is queue management that increases and decreases the discard rate.
The adjustment unit 142 of the ONU 100 compares the requested bandwidth set in the bandwidth request information transmitted by the ONU 100 with the allocated bandwidth of the ONU 100 indicated by the bandwidth allocation information. The adjustment unit 142 determines that the entire PON is congested if the requested bandwidth exceeds the allocated bandwidth, and increases the discard rate calculated by the discard rate calculation unit 141 using AQM such as RED. If the requested bandwidth is equal to or less than the allocated bandwidth, the adjustment unit 142 determines that the entire PON is not congested, and decreases the discard rate calculated by the discard rate calculation unit 141 using the AQM.

  For example, consider a case where the total uplink bandwidth that can be allocated to the ONU 100 is 10 Gbit / s, and ten ONUs 100 are connected to the OLT 200. Assume that the requested bandwidth of the ONU 100 is 1 Gbit / s, and the allocated bandwidth assigned to the ONU 100 from the OLT 200 is 500 Mbit / s. The adjustment unit 142 knows that the entire PON is congested because the requested bandwidth exceeds the allocated bandwidth. In this case, the adjustment unit 142 adds an arbitrary value to the discard rate calculated by the discard rate calculation unit 141 by AQM, and finally sets the discard rate to instruct the upstream signal processing unit 131 to execute. For example, if the discard rate calculated by the discard rate calculation unit 141 based on the queue length is 0.1, the adjustment unit 142 can increase the value by 0.1 to 0.2. Alternatively, it is assumed that the requested bandwidth of the ONU 100 is 1 Gbit / s and the allocated bandwidth allocated from the OLT 200 is 2 Gbit / s. In this case, the adjustment unit 142 knows that the entire PON is not congested because the requested bandwidth is equal to or less than the allocated bandwidth. In this case, the adjustment unit 142 subtracts an arbitrary value from the discard rate calculated by the discard rate calculation unit 141 using the AQM, and finally sets the discard rate to instruct the upstream signal processing unit 131 to execute. However, the discard rate ranges from 0 to 1. Further, the increment or decrease of the discard rate may be determined according to the ratio of the requested bandwidth to the allocated bandwidth of the ONU 100 or the ratio of the requested bandwidth of the ONU 100 to the requested bandwidth of the entire PON.

  The adjustment unit 142 may increase / decrease the increase in the discard rate with respect to the change in the queue length when increasing / decreasing the discard rate, or may shift the increase / decrease while maintaining the inclination of the increase in the discard rate with respect to the queue length. In addition, an upper limit or a lower limit of the discard rate may be set. For example, the adjustment unit 142 may determine the increased discard rate as in the first embodiment, or may determine the decreased discard rate as in the second embodiment.

  The adjustment unit 142 may return the increased discard rate to a predetermined value, for example, a value before increasing, when any of the conditions (1a) to (8a) is satisfied as in the first embodiment. , May be reduced. In addition, the adjustment unit 142 returns the decreased discard rate to a predetermined value, for example, the value before the decrease, when any of the conditions (1b) to (8b) is satisfied as in the second embodiment. Or may be increased.

(Fifth embodiment)
In this embodiment, statistical processing such as averaging is added to the determination in the first to fourth embodiments. Specifically, the adjustment unit 142 performs AQM execution or stop processing when the amount of allocation for a request is a predetermined time or a predetermined number of times in a predetermined number of times in the past or the most recent determination, Increase or decrease the disposal rate. The adjustment unit 142 uses a statistical value of the requested bandwidth and a statistical processing value of the allocated bandwidth instead of the requested bandwidth and the allocated bandwidth in the first to fourth embodiments. For example, a moving average or an exponential average can be used as the statistical value. The statistical value of the requested bandwidth and the statistical value of the allocated bandwidth may be converted into a data amount or may be converted into a data amount per unit time. Also, when comparing by the number of times, a moving average determined by the sum of values of a predetermined number of times may be used, or a sum of exponential averages that are added with a smaller weight as the past is used.

  In any of the embodiments described above, the bit rate unit is used for each of the requested bandwidth and the allocated bandwidth, but may be a byte unit or a bit unit, or an arbitrary data unit such as an Ethernet (registered trademark) frame unit or a packet unit. Good. Further, the adjustment unit 142 can determine whether or not the AQM is executed and whether the discard rate is increased or decreased for each priority queue. In addition, a requested bandwidth or an allocated bandwidth in a bit unit or a byte unit in a certain time section can be converted into a bit rate or the like.

  The PON to which the above-described embodiment is applied may be Ethernet (registered trademark) PON (EPON), for example, ATM-PON, BPON, G-PON, NG-PON2, XG-PON, Alternatively, XGS-PON may be used. It should be noted that a functional unit conforming to EPON such as an MPCP unit may be a functional unit having the same function in another PON. In the above-described embodiment, one or a plurality of ONUs have been described as an example. However, another network termination device for PON, for example, an ONT (Optical Network Unit) can be used. Furthermore, although the present embodiment has been described by taking RED as an example, other congestion avoidance schemes may be used. Although the ONU 100 has been described as an example including one Alloc-ID, a plurality of ONUs 100 may be provided. In addition, the format of each information when it is set in each message and each message transmitted / received between each function unit in the ONU 100 such as bandwidth request information and bandwidth allocation information can be appropriately changed.

  According to the embodiment described above, in the communication system, one or more communication devices communicate with the upper communication device. For example, the communication system is the optical communication system 1, the upper communication apparatus is the OLT 200, and the communication apparatus is the ONU 100. The communication device includes a reception unit, a queue, an allocated band reception unit, a calculation unit, an adjustment unit, and an uplink signal processing unit. The receiving unit receives uplink data to be transmitted to the higher-level communication device. The receiving unit is, for example, the lower communication unit 110. The queue holds uplink data received by the receiving unit. The allocated band receiving unit receives allocated band information from the host communication device. The allocated bandwidth receiving unit is, for example, the MPCP unit 134. The calculation unit calculates a discard rate, which is a rate of discarding uplink data, or an discard amount of uplink data (for example, a discard threshold) based on the data amount of uplink data held in the queue. The adjustment unit calculates whether the uplink data is partially discarded based on the congestion state of the communication system determined using the data amount of the uplink data and the allocated bandwidth information from the higher-level communication device. At least one of the discard rate and the increase / decrease of the discard amount is determined. The uplink signal processing unit discards a part of the uplink data received by the reception unit without holding it in the queue according to the determination of the adjustment unit. Further, the uplink signal processing unit transmits the uplink data held in the queue without being discarded to the upper communication apparatus according to the allocated band.

  The adjustment unit, when the time or number of times when the communication system is determined to be congested satisfies a predetermined condition, when the time or number of times when it is determined that partial discard of uplink data is necessary satisfies a predetermined condition, or When the data amount of the uplink data is larger than the predetermined amount, the discard rate or the discard amount calculated by the calculation unit is increased. For example, the adjustment unit may increase the discard rate or the discard amount calculated by the calculation unit by a certain amount, or may increase the increase rate as the amount of queued data increases. When the predetermined time has elapsed after increasing the discard rate, the adjustment unit determines that it is not necessary to partially discard uplink data when the time or number of times that the communication system determines that the communication system is not congested satisfies a predetermined condition. When the determined time or number of times satisfies a predetermined condition, or when the amount of uplink data is smaller than a predetermined amount, the discard rate or discard amount calculated by the calculation unit is decreased.

  In addition, the adjustment unit, when the time or number of times that the communication system is determined not to be congested satisfies a predetermined condition, or when the amount of uplink data is less than a predetermined amount, Reduce waste. For example, the adjustment unit may decrease the discard rate or the discard amount by a certain amount, or may decrease the decrease amount as the queued data amount increases. The coordinator determines that the upstream data will be partially discarded if the predetermined time has elapsed since the drop rate was decreased, or when the time or number of times the communication system is determined to be congested satisfies the predetermined condition When the time or number of times satisfies a predetermined condition, or when the amount of uplink data is larger than a predetermined amount, the discard rate or discard amount calculated by the calculation unit is increased. The predetermined condition of the number of times is, for example, one time, a plurality of times, or a predetermined number of times of a plurality of times of determination.

  Further, the adjustment unit determines at least one of the necessity of partial discard of uplink data and the increase / decrease of the discard rate or discard amount calculated by the calculation unit based on the congestion status statistics of the communication system determined multiple times. You may decide. Further, the adjustment unit may determine the congestion state of the communication system by using the data amount statistics of the uplink data and the allocated bandwidth statistics from the higher-level communication device.

  According to the present embodiment, it is possible to improve the bandwidth utilization efficiency of the entire PON system without modifying control messages such as REPORT messages and GATE messages.

  All or some of the functions of the PON signal processing unit 130 and the queue management unit 140 of the ONU 100 use hardware such as an application specific integrated circuit (ASIC), a programmable logic device (PLD), and a field programmable gate array (FPGA). Realized. The ONU 100 includes a central processing unit (CPU), a memory, an auxiliary storage device, and the like connected by a bus, and can implement some functions of the PON signal processing unit 130 and the queue management unit 140 by executing a program. Good. In this case, the program may be recorded on a computer-readable recording medium. The computer-readable recording medium is, for example, a portable medium such as a flexible disk, a magneto-optical disk, a ROM, or a CD-ROM, and a storage device such as a hard disk built in the computer system. The program may be transmitted via a telecommunication line.

  The embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and includes designs and the like that do not depart from the gist of the present invention.

  It can be used for a communication system that performs one-to-many communication.

DESCRIPTION OF SYMBOLS 1 ... Optical communication system, 100 ... ONU, 110 ... Lower communication part, 120 ... Upper communication part, 130 ... PON signal processing part, 131 ... Up signal processing part, 132 ... Queue, 133 ... Down signal processing part, 134 ... MPCP , 140 ... Queue management unit, 141 ... Discard rate calculation unit, 142 ... Adjustment unit, 200 ... OLT, 300 ... ODN

Claims (8)

The communication device in a communication system in which one or more communication devices communicate with a host communication device,
A receiving unit for receiving uplink data to be transmitted to the higher-level communication device;
A queue that holds the uplink data received by the receiver;
An allocated band receiving unit for receiving allocated band information from the higher-level communication device;
A calculation unit that calculates a discard rate, which is a ratio of discarding the uplink data, or a discard amount of the uplink data, based on a data amount of the uplink data;
Based on the congestion state of the communication system determined using the data amount and the information on the allocated bandwidth from the higher-level communication device, whether or not to discard a part of uplink data and the discard calculated by the calculation unit An adjustment unit for determining at least one of rate and increase / decrease of the waste amount,
According to the determination by the adjustment unit, a process of discarding a part of the uplink data received by the reception unit without being held in the queue, and the uplink data held in the queue without being discarded according to the allocated band An upstream signal processing unit for performing processing to be transmitted to the host communication device;
A communication device comprising: The adjustment unit, when the time or number of times that the communication system is congested satisfies a predetermined condition, or when the time or number of times that it is determined that partial discard of uplink data is necessary satisfies a predetermined condition Alternatively, when the data amount is larger than a predetermined amount, the discard rate or the discard amount calculated by the calculation unit is increased.
The communication apparatus according to claim 1. When the predetermined time has elapsed, the adjustment unit determines that it is not necessary to discard part of the uplink data when the time or number determined that the communication system is not congested satisfies a predetermined condition. Decreases the discard rate or the discard amount calculated by the calculation unit, when the predetermined condition is satisfied, or when the data amount is smaller than the predetermined amount,
The communication apparatus according to claim 2. When the time or number of times that the communication system determines that the communication system is not congested satisfies a predetermined condition, or when the amount of data is less than a predetermined value, the adjustment unit calculates the discard rate or Reducing the amount of waste,
The communication apparatus according to any one of claims 1 to 3. When the predetermined time has elapsed, the adjustment unit determines that the time or number of times that the communication system has been congested satisfies a predetermined condition, and that the time or number of times that the uplink data is partially discarded is predetermined. When the condition is satisfied, or when the data amount is larger than a predetermined amount, the discard rate or the discard amount calculated by the calculation unit is decreased,
The communication apparatus according to claim 4. The adjustment unit determines at least one of necessity / unnecessity of partial discard of uplink data and increase / decrease of the discard rate or the discard amount calculated by the calculation unit based on the congestion state statistics determined a plurality of times. decide,
The communication apparatus as described in any one of Claims 1-5. The adjustment unit determines a congestion state of the communication system using the statistics of the data amount and the statistics of the allocated bandwidth from the higher-level communication device.
The communication apparatus as described in any one of Claims 1-6. A queue management method for the communication device in a communication system in which one or more communication devices communicate with a higher-level communication device,
A reception step of receiving uplink data to be transmitted to the higher-level communication device;
A queuing step for temporarily holding the uplink data received in the receiving step in a queue;
An allocated band receiving step of receiving allocated band information from the higher-level communication device;
A calculation step of calculating a discard rate, which is a rate of discarding the uplink data, or a discard amount of the uplink data, based on a data amount of the uplink data;
Based on the congestion state of the communication system determined using the data amount and the information on the allocated bandwidth from the higher-level communication device, whether or not part of the uplink data needs to be discarded, and the calculation calculated in the calculation step An adjustment step for determining at least one of a discard rate or an increase / decrease in the discard amount;
According to the determination in the adjustment step, a discarding step of discarding a part of the uplink data received in the reception step without holding it in the queue;
A transmission step of transmitting the uplink data held in the queue without being discarded to the upper communication device according to the allocated bandwidth;
A queue management method.

Images