JP2018125671A - Communication device and data transmission program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a communication device and a data transmission program which restrain increase in the latency time of data frame transmission in each ONU of lower PON, in a PON system connected in multiple stages, and in which discard of data frame is less likely to occur.SOLUTION: An ONU notifies an OLT of the first time at which a discovery signal, i.e., a response request for an unknown connection apparatus from a master OLT, is received, the OLT starts transmission of a gate signal for permitting transmission of an uplink signal to a slave ONU at a second time when a predetermined adjustment time elapsed from the first time, completes tabulation processing for tabulating the request band values, contained in the uplink signal from the slave ONU, as the total request band value, and band allocation processing for the slave ONU based on the total request band value, at the third time, and establishes synchronization with te master OLT by repeating the processing from the second time to the third time until reception of next discovery signal.SELECTED DRAWING: Figure 3

Description

  The present invention relates to a communication device and a data transmission program, and more particularly, to a communication device and a data transmission program of a multistage PON system in which passive optical networks (PON) are connected in multiple stages.

  2. Description of the Related Art In recent years, a service called FTTH (Fiber To The Home) using an optical fiber as a transmission path has been widespread for the purpose of providing high-speed and broadband broadband services to general private homes. For the provision of broadband services by FTTH, an optical access network called a PON system is often used. In many cases, an Ethernet (registered trademark) frame standardized by IEEE 802.3 is used as a frame in data transmission in the PON system, and the PON system in this case is particularly an E-PON (Ethernet (registered trademark) Passive Optical Network). ) It is called a system.

  In general, an E-PON system includes one intra-station optical line terminator (OLT) and a plurality of optical subscriber line terminators (ONU) using optical fibers and passive components. This is a point-to-multi-point (P to MP, one-to-many) type optical subscriber network system connected by an optical splitter. And the operation | movement in the case of communicating between OLT and ONU is standardized by the nonpatent literature 1, the nonpatent literature 2, etc. In the E-PON system, an FTTH service can be economically provided by sharing an optical fiber, an OLT, and the like with a plurality of subscribers.

  In the E-PON system, an Ethernet (registered trademark) frame is transmitted and received by an optical signal between the OLT and the ONU. Specifically, all connected ONUs are logically identified by LLID (Logical Link ID). Since the downstream (from OLT to ONU) signal reaches all ONUs because the optical signal statically branched by the optical splitter reaches each ONU, the LLID addressed to itself or the LLID addressed to everyone (Broadcast LLID) Receive and discard LLIDs addressed to others.

  The uplink signal (from ONU to OLT) is a multiplexing TDMA (Time Division Multiple Access) system in which a frame transmission time is divided and shared by a plurality of users. This is because the frames transmitted from each ONU are multiplexed by the optical splitter, so that collision between frames occurs unless the transmission timing of the optical signal transmitted from each ONU is controlled.

  More specifically, the following method is employed to avoid a collision. That is, first, in the Discovery process, a 32-bit counter called “Local Time” possessed by the OLT is loaded by the ONU, thereby synchronizing the Local Time between the OLT and the ONU. The Discovery process is a series of procedures for connecting a new ONU to the OLT. Strictly speaking, the OLT and the ONU are loaded with counters that differ by the delay difference of the physical distance. This delay difference is generally called RTT (Round Trip Time).

  Using this synchronized Local Time, the OLT instructs each ONU by a frame called GATE, the transmission timing (Start Time) and the transmission length (Length) from the ONU, and each ONU transmits transmission data according to the instruction. A collision is avoided by sending a REPORT frame including the like.

  Here, in the case of a P to MP type connection in which a plurality of ONUs are connected by branching statically between an OLT which is a station side device and an optical fiber like a PON, every time two branches are made by an optical splitter for branching. A loss (attenuation of about 3 dB) occurs in which the optical transmission power is about half. Also in the optical fiber that is the transmission path, attenuation of optical transmission power occurs in proportion to the wiring length.

  On the other hand, in the optical transceivers provided in the OLT and the ONU, the optical transmission power and the light receiving sensitivity are defined by standards such as Non-Patent Document 1 and Non-Patent Document 2. Therefore, the network composed of the optical splitter and the optical fiber must be designed so that the optical level of each part falls within the dynamic range defined by the optical transmission power and the optical reception sensitivity. As described above, the conventional single-stage PON system has a limit in increasing the number of branches, extending the transmission path, and constructing a long-distance network.

The multistage PON system was conceived. A multistage PON system replaces one or more of the ONUs connected to the first stage OLT with a relay device, and connects a lower ONU under the relay device to connect the second stage PON. It is the system which constituted.
If one or more of the lower ONUs are replaced with relay devices, a third-stage PON can be configured. In the multistage PON system, since the OLT of the first stage PON is the highest-level device, this OLT is hereinafter referred to as “parent OLT”. According to such a multistage PON system, the total number of branches as viewed from the parent OLT is increased, and it is possible to flexibly cope with the installation distance of the user and to increase the distance.

  As a conventional technique of the multistage PON system as described above, a multistage PON system disclosed in Patent Document 1 is known. FIG. 11 shows a configuration of a multistage PON system according to Patent Document 1. As shown in FIG. 11, the multistage PON system according to Patent Document 1 includes an upper PON composed of an OLT (hereinafter, “parent OLT”) and relay nodes # 1 to #m, and relay nodes # 1 to #m. Are connected to ONU # 1-1... # M-n to form a lower level PON. Taking relay node # 1 as an example, relay node # 1 includes ONU # 1 having an ONU function and OLT # 1 having an OLT function. ONU # 1 is opposed to OLT, and OLT # 1 is ONU. It faces # 1 to 1-n.

  More specifically, in the multistage PON system according to Patent Document 1, a relay node is connected to the parent OLT by an optical splitter in addition to the ONU, and an ONU branched by the optical splitter is also connected under the relay node. . According to Patent Document 1, the relay node aggregates the bandwidth request amount indicated in the REPORT frame from each ONU under the relay node (that is, ONU # 1-1 of the lower PON, etc.) and A REPORT frame is transmitted to the parent OLT. The parent OLT calculates the bandwidth request amount of the REPORT frame from each relay node, and instructs each relay node to perform uplink transmission using the GATE frame. The relay node issues a transmission instruction to each ONU of the lower PON via the GATE frame so as to match the upstream transmission band specified by the Start Time and Length of the GATE frame from the parent OLT. As a result, upstream frames from lower ONUs can be transmitted without waiting at the relay node, and the buffer capacity of the relay node can be reduced by not retaining the upstream frames at the relay node. .

Japanese Unexamined Patent Publication No. 2006-082393

IEEE (Institute of Electrical and Electronics Engineers) std 802.3ah-2009 IEEE (Institut of Electrical and Electronics Engineers) std 802.3av-2009

  By the way, in a general PON system, when an OLT receives a bandwidth request by a REPORT frame from each ONU and sends a transmission instruction to each ONU by a GATE frame, the OLT sends a GATE frame to the ONU with the shortest distance that the OLT can take on the system. The ONU that transmits the shortest distance transmits the REPORT frame to the OLT in the shortest time from the reception of the GATE frame, and the time until the OLT receives the REPORT frame (hereinafter, this time is expressed as “ΔGate”). At least needed. Each ONU receives this GATE frame, and transmits the amount of bandwidth accumulated until the REPORT frame is transmitted to the REPORT frame.

  At this time, the OLT aggregates the REPORT frames of all the ONUs having different connection distances from the OLT (that is, different RTTs), so that the REPORT frames can be received together within a certain period of time, and sent to each ONU. An instruction is given by a GATE frame. Therefore, the OLT requires time for REPORT frame aggregation corresponding to the number of ONUs that can be connected to the subordinates, not the number of ONUs currently connected (this time is hereinafter referred to as “ΔRep”). In the system specification of the PON, the time from the transmission of the GATE frame to the aggregation of the REPORT frame is set in consideration of the propagation delay time depending on the connection distance between the OLT and the ONU. As an example, the maximum connection distance between the OLT and the ONU is defined as 20 km in Non-Patent Document 1.

After aggregating the REPORT frames, the OLT calculates the bandwidth in the GATE frame to be given next from all bandwidth request values, and transmits the GATE frame indicating the permitted bandwidth to each ONU again. The function of calculating the uplink transmission timing and transmission band (that is, the time allocated for transmission) for the next transmission of each ONU is called a DBA (Dynamic Bandwidth Allocation) calculation processing function. Yes. Hereinafter, the time required for the DBA calculation process is referred to as “DBA calculation time”. As a DBA calculation method, calculation is performed by firmware (F / W) while aggregating all REPORT frames, or calculation by FW by converting a part or all of the calculation processing to hardware (H / W). There is a method of speeding up. As described above, a certain period (hereinafter referred to as “Δdba”) is required until the DBA operation is completed after the GATE frame is transmitted, and communication between the OLT and the ONU is performed. In this case, this is realized by repeating the cycle of Δdba. Here, the minimum of Δdba is
Δdba = ΔGate + ΔRep + DBA calculation time. Naturally, the shorter this Δdba, the shorter the waiting time in the ONU.

  However, in the multi-stage PON system according to Patent Document 1, each band request condition and band grant condition of the multi-stage PON is relayed. Therefore, from all distances, operation time, aggregation time, etc. configured in multi-stages, Δdba The period is very large. For this reason, the DATA frame input to the ONU immediately after the transmission of the REPORT frame in the ONU subordinate to the lower PON is not requested for bandwidth, so that the transmission timing of the next REPORT frame is long due to the extremely long Δdba cycle. You will have to wait.

  Therefore, in the multistage PON system according to Patent Document 1, there is an advantage that the buffer of the relay node is reduced. However, for each ONU that has already been installed, the buffer amount corresponding to the waiting time longer than before is increased due to an increase in Δdba. There are disadvantages that are required. That is, in the multistage PON system according to Patent Document 1, since the GATE frame and REPORT frame used in the lower PON are once terminated and transferred to the upper PON, the bandwidth occupied by the REPORT frame is reduced, and the It can be said that the efficiency is improved in the sense that the bandwidth that can be used other than the control frame is increased. However, with respect to Δdba of the multistage PON system according to Patent Document 1, instead of the relay node, the parent OLT and the ONU connected to the lower PON are connected using an optical amplifier that reproduces the optical signal as it is, and the parent OLT It is easily assumed that it is larger than Δdba when band control is performed for all of the upper PON and the lower PON connected via the optical amplifier.

  On the other hand, when considering the clock in the multistage PON system, it is difficult to control the lower PON so that the time and the error indicated by the upper PON do not cause an error if the clock frequency used by the upper PON and the lower PON is different. . In the multi-stage PON system according to Patent Document 1, the lower PON has a mechanism for performing a DBA operation process upon receiving a GATE frame from the upper PON (parent OLT) and giving a bandwidth to the lower PON. On the other hand, the OLT of the PON system transmits a Discovery GATE for connecting a new ONU whose connection distance from the OLT is unknown to the OLT at a predetermined fixed interval in the periodic transmission of the GATE frame. The connected relay node and the ONU of the upper level PON cannot perform any uplink transmission during this period. This also applies to the lower PON. Since the ONU of the lower PON operates based on the GATE frame received from the OLT, the timing at which the OLT transmits the Discovery GATE is not considered in the operation of the ONU.

  Also, when the relay node aggregates the REPORT frames of each ONU under the lower PON and transmits them to the parent OLT, the simplest method is to transmit the sum of the bandwidths of the DATA frames received from each ONU. When the bandwidth request values of the ONUs and relay nodes under the upper PON in the parent OLT are larger than the bandwidth given by the parent OLT, it is not possible to give all of the bandwidth request values sent by the relay node. In this case, the relay node needs to calculate distribution according to the contents of the REPORT frame received from each subordinate ONU, and the waiting time at the lower level PON increases.

  On the other hand, in order not to increase the delay of the DATA frame input after the bandwidth request by the REPORT frame from the ONU under the lower PON due to the increase of the overall Δdba including the upper PON and the lower PON, It is also conceivable to perform band control independently with each PON. In this case, the operation of the PON system is not multistage, and Δdba may be considered in a normal cycle. However, in this case, when the time zone for receiving the DATA frame from the lower PON at the relay node is not in time for the transmission time of the REPORT frame transmitted from the relay node to the parent OLT in the upper PON, the timing until the next Δdba The bandwidth request is awaited, causing a delay in data transmission time.

  FIG. 12 is a timing chart for explaining the above operation. In FIG. 12, the GATE frame issued at time T00 (indicated by a □ including the letter “G” in FIG. 12) is received by relay node # 1 at time T10, and then the DBA operation is started. The DBA operation is finished at T11. Thereafter, the relay node # 1 sends a GATE frame to the subordinate ONU # 1, and upon receiving this, the ONU # 1 sends a REPORT frame (indicated by □ including the letter “R” in FIG. 12) to the relay node. doing. The relay node also aggregates REPORT frames from other ONUs #N, and transmits the REPORT frame to the parent OLT at time T13. The parent OLT that has received the REPORT frame at time T01 aggregates the REPORT frames from other relay nodes #N at time T02, and executes the DBA operation. The parent OLT that has finished the DBA operation at time T03 sends a GATE frame to the relay node. The relay node receives the GATE frame, starts the DBA operation at time T15, and ends the DBA operation at time T16. doing.

  In FIG. 12, from time T00 to time T03 becomes Δdba of the parent OLT, and from time T11 to time T16 becomes Δdba in the relay node. In other words, Δdba is from when a GATE frame is sent until the next GATE frame is sent. In this way, in the operation shown in FIG. 12, the DBA operation is executed independently for the upper PON and the lower PON. However, in the operation shown in FIG. 12, as indicated by the symbol <1>, the relay node # 1 is given a band from the parent OLT in response to the previous transmission request in the REPORT frame, and the relay node # 1 further turns ONU # As a result of the bandwidth request to 1-1, ONU # 1-1 transmits a DATA frame. In other words, the bandwidth request of ONU # 1-1 is a bandwidth given by the parent OLT through the two-stage timing of the lower PON and the upper PON, and ONU # 1-1 has a time DATA frame that is approximately twice as long as when it is not multistage. May be awaited to send. Therefore, in order to cooperate without changing the bandwidth control period between the upper PON and the lower PON, it is necessary to match the operation timings of the upper PON and the lower PON.

  Further, even when the upper PON timing and the lower PON timing are linked and the upper PON and the lower PON are operated independently, all of the bands allocated by the lower PON remain in the relay node from the parent OLT. Not necessarily assigned. If this occurs continuously, the ONU # 1 temporary storage buffer of the relay node may fail, and frame discard may occur at the relay node.

  The present invention has been made in view of the above-described problems. In a PON system connected in multiple stages, the increase in the waiting time of data frame transmission in each ONU of the lower-level PON is suppressed, and the discard of the data frame is performed. It is an object of the present invention to provide a communication device and a data transmission program that are less likely to generate the problem.

  In order to achieve the above-described object, a communication device according to the present invention is a communication device that connects a higher-order PON connected in the upstream direction and a lower-order PON connected in the downstream direction, and uses the OLT of the higher-order PON. An ONU unit connected to a parent OLT, and an OLT unit connected to the child ONU that is connected to the ONU unit and is an ONU of the lower-level PON. The ONU unit is unknown from the parent OLT The OLT unit is notified of a first time at which a discovery signal, which is a response request to the connected device, is received, and the OLT unit receives a second adjustment time after a predetermined adjustment time has elapsed from the first time. The transmission of the gate signal that permits the transmission of the upstream signal to the child ONU is started, and the requested bandwidth value included in the upstream signal from the child ONU is aggregated as the total requested bandwidth value And the bandwidth allocation processing for the child ONU based on the total requested bandwidth value is completed at the third time, and the next discovery signal is received from the second time to the third time. By repeating this until the synchronization with the parent OLT is established.

  In order to achieve the above-described object, the data transmission program according to the present invention connects an upper PON connected in the upstream direction and a lower PON connected in the downstream direction, and is a parent OLT that is an OLT of the upper PON. A program for transmitting data using a communication device having an ONU unit connected to the ONU unit and an OLT unit connected to a child ONU that is connected to the ONU unit and is an ONU of the lower level PON, A first control means for notifying the OLT unit from the ONU unit of a first time at which the ONU unit has received a discovery signal that is a response request for an unknown connected device from the parent OLT; A gate signal that permits transmission of an upstream signal from the OLT unit to the child ONU at a second time after a predetermined adjustment time has elapsed from the time. Transmission processing, summing up the requested bandwidth value included in the uplink signal from the child ONU as a total requested bandwidth value by the OLT unit, and the child ONU based on the total requested bandwidth value The bandwidth allocation process is completed at a third time, and the process from the second time to the third time is repeated until the next discovery signal is received, thereby establishing synchronization with the parent OLT. It functions as the second control means.

  According to the communication apparatus and the data transmission program according to the present invention, in a PON system connected in multiple stages, an increase in waiting time of data transmission in each ONU of a lower level PON is suppressed, and discarding of data is difficult to occur. A communication device and a data transmission program can be provided.

It is a block diagram which shows an example of a structure of the PON system which concerns on embodiment. It is a block diagram which shows an example of a structure of the relay apparatus which concerns on 1st Embodiment. It is a timing chart which shows the procedure of the timing synchronization of the PON system which concerns on 1st Embodiment. It is a timing chart which shows the procedure of the zone | band request | requirement process of the PON system which concerns on 1st Embodiment. It is a figure which shows operation | movement of the up buffer which concerns on 1st Embodiment. It is a flowchart which shows the transmission procedure of the uplink buffer information which concerns on 1st Embodiment. It is a timing chart which shows the DBA arithmetic processing procedure of the PON system which concerns on 1st Embodiment. It is a block diagram which shows an example of a structure of the relay apparatus which concerns on 2nd Embodiment. It is a timing chart which shows the procedure of the timing synchronization of the PON system which concerns on 2nd Embodiment. It is a timing chart which shows the procedure of the zone | band request | requirement process of the PON system which concerns on 2nd Embodiment. It is a block diagram which shows the structure of the PON system which concerns on a prior art. It is a timing chart which shows the packet transmission procedure of the PON system which concerns on a prior art.

  DESCRIPTION OF EMBODIMENTS Hereinafter, embodiments for carrying out the present invention will be described in detail with reference to the drawings. In the following description, a mode using E-PON as an example of the PON system will be described as an example.

[First Embodiment]
A communication device and a data transmission program according to the present embodiment will be described with reference to FIGS.

  FIG. 1 shows an example of the overall configuration of a PON system 1 including a relay device 30 as a communication device according to the present embodiment. As shown in FIG. 1, the PON system 1 is a multistage PON system configured to include an upper PON 100 and a lower PON 102.

  The host PON 100 includes an OLT 10 as a parent OLT connected to an NNI (Network Node Interface), U_ONUs 20-1, 20-2, and 20-3 (hereinafter, collectively referred to as “U_ONU20”) that are subordinates of the OLT 10. And an optical splitter 50. The OLT 10 and the optical splitter 50 are connected by an optical transmission path using a single optical fiber or the like, and the optical transmission path branched from the optical splitter 50 is connected between the optical splitter 50 and each of the relay device 30 and the U_ONU 20. Is connected.

  The lower level PON 102 includes the relay device 30, L_ONUs 22-1, 22-2, 22-3, and 22-4 (hereinafter collectively referred to as “L_ONU 22”) that are ONUs under the relay device 30, and an optical splitter 52. ing. The relay device 30 and the optical splitter 52 are connected by a single optical transmission line, and the optical splitter 52 and each of the L_ONUs 22 are connected by an optical transmission line branched from the optical splitter 52.

  FIG. 2 is a block diagram illustrating an example of the configuration of the relay device 30 according to the present embodiment. As illustrated in FIG. 2A, the relay device 30 includes an R_ONU 31 having a function as an ONU and an R_OLT 32 having a function as an OLT.

  As shown in FIG. 2A, the R_ONU 31 includes a PON optical transmission / reception unit 301, an ONU-MAC (Media Access Control) unit 302, a UI (User Interface) side transmission / reception unit 303, an upstream buffer 307, and a PLL 324. Has been.

  The PON optical transmission / reception unit 301 faces the OLT 10 via the optical splitter 50, converts an optical signal in the Ethernet (registered trademark) format transmitted from the OLT 10 into an electrical signal, and extracts a clock signal from the converted electrical signal. I do. The PON optical transmission / reception unit 301 converts an electrical signal in the Ethernet (registered trademark) frame format to be transmitted to the OLT 10 into an optical signal. In the present embodiment, an Ethernet (registered trademark) frame is an MPCP (Multi Point Control Protocol) frame, an OAM (Operations, Administration, and Maintenance) frame, or a user data frame that is transmitted in accordance with the E-PON protocol. And so on.

  The ONU-MAC unit 302 receives an electrical signal from the PON optical transmission / reception unit 301, terminates the Ethernet (registered trademark) frame from the OLT 10, and transmits the Ethernet (registered trademark) frame to be transmitted to the OLT 10 to the PON optical transmission / reception unit 301. At the same time, the status of the R_ONU 31 is monitored. The UI-side transmitting / receiving unit 303 is an interface unit that transmits / receives an Ethernet (registered trademark) frame that is user data from the UNI (User Network Interface) to the R_OLT 32. The upstream buffer 307 temporarily stores upstream frames transmitted from the bridge unit 315 that is an internal circuit of the ONU-MAC unit 302. The PLL 324 removes the jitter and wander of the clock extracted by the PON optical transmission / reception unit 301, and uses the clock synchronized with the OLT 10 as a clock for the entire relay device 30 centering on the ONU-MAC unit 302, the OLT-MAC unit 305, and the like. Distribute as

  As shown in FIG. 2A, the R_OLT 32 includes a PON optical transmission / reception unit 306, an OLT-MAC unit 305, and an NI (Network Interface) side transmission / reception unit 304.

  The PON optical transmission / reception unit 306 faces the L_ONU 22 via the optical splitter 52 and converts an optical signal in the Ethernet (registered trademark) format transmitted from the L_ONU 22 into an electrical signal. The PON optical transmission / reception unit 306 converts an Ethernet (registered trademark) frame format electrical signal to be transmitted to the L_ONU 22 into an optical signal.

  The OLT-MAC unit 305 receives an electrical signal from the PON optical transmission / reception unit 306, terminates an Ethernet (registered trademark) frame from the L_ONU 22, generates an Ethernet (registered trademark) frame to be transmitted to the L_ONU 22, and generates a PON optical transmission / reception unit At the same time, the status of the R_OLT 32 is monitored. The NI side transmission / reception unit 304 is an interface unit that transmits / receives an Ethernet (registered trademark) frame to / from the R_ONU 31.

  Next, the details of the ONU-MAC unit 302 and the OLT-MAC unit 305 will be described with reference to FIGS.

  As shown in FIG. 2B, the ONU-MAC unit 302 according to the present embodiment includes a PON transmission unit 308, a PON reception unit 309, a bridge unit 315, an OAM transmission / reception unit 310, an MPCP transmission / reception unit 311, a CPU (Central A processing unit (Processing Unit) 312, a RAM (Random Access Memory) 313, and a ROM (Read Only Memory) 314 are included.

  The bridge unit 315 performs buffering of Ethernet (registered trademark) frames to be transferred to the upper PON 100, granting and deleting of VLAN-Tags for identifying VLANs (Virtual Local Area Networks), address learning, priority processing, and the like, and uplink frames The downstream frame continuity is monitored.

  The OAM transmission / reception unit 310 and the MPCP transmission / reception unit 311 control the PON interface. The PON transmission unit 308 transmits a PON control frame such as a user data frame, an MPCP frame, or an OAM frame to the PON optical transmission / reception unit 301. The PON reception unit 309 receives and identifies user data and PON control frames from the PON optical transmission / reception unit 301 and distributes them to the MPCP transmission / reception unit 311, the OAM transmission / reception unit 310, and the bridge unit 315.

  The CPU 312 is connected to each unit of R_ONU and performs various settings, readings, and the like. Also, Ethernet (registered trademark) frames input from the UNI via the UI-side transmitting / receiving unit 303 are temporarily stored (stored) in the upstream buffer 307 connected to the bridge unit 315. It has a function of reading out as an accumulated amount and notifying the OLT 10 via the MPCP transmission / reception unit 311 using a REPORT frame. A RAM 313 performs development, temporary storage, and the like of a program executed by the CPU 312. The ROM 314 stores programs and various setting values.

  As shown in FIG. 2C, the OLT-MAC unit 305 according to the present embodiment includes a PON transmission unit 323, a PON reception unit 322, a bridge unit 316, an OAM transmission / reception unit 317, an MPCP transmission / reception unit 318, a CPU 319, and a RAM 320. , And a ROM 321.

  The bridge unit 316 determines an L_ONU 22 to which the Ethernet (registered trademark) frame from the R_ONU 31 is transferred, an address learning function for determining a transfer destination from information such as a MAC address and a VLAN, and upper and lower frames for each LLID. A function for performing continuity monitoring, VLAN-Tag assignment and deletion for VLAN identification, buffering for priority processing and LLID identification, and the like are performed.

  The MPCP transmission / reception unit 318 transmits / receives a control frame necessary for link establishment of the PON interface. The OAM transmission / reception unit 317 transmits an OAM frame for performing various settings necessary for the operation of the L_ONU 22 and receives the OAM frame from which the state of the L_ONU 22 is read. The PON transmission unit 323 controls the user data from the bridge unit 316, the OAM frame from the OAM transmission / reception unit 317, the MPCP frame from the MPCP transmission / reception unit 318, and the like, and sends them to the PON optical transmission / reception unit 306. The PON reception unit 322 receives and identifies user data, PON control frames, and the like from the PON optical transmission / reception unit 306 and distributes them to the MPCP transmission / reception unit 318, the OAM transmission / reception unit 317, and the bridge unit 316.

  The CPU 319 performs various settings and readings. More specifically, the CPU 319 is connected to each unit including an OAM transmission / reception unit 317, an MPCP transmission / reception unit 318, and a bridge unit 316. When each L_ONU 22 transmits an upstream frame, the frame accumulation information is notified to the R_OLT 32 using a REPORT frame. However, the CPU 319 collects the frame accumulation information notified by each L_ONU 22 via the MPCP transmission / reception unit 318, and next time. In this REPORT frame, each L_ONU 22 has a DBA calculation processing function (dynamic band control function) for calculating an uplink transmission timing and a transmission band (transmission length) for transmitting a data frame. The RAM 320 has a function of a development area when the CPU 319 executes a program, a function of temporary storage in DBA arithmetic processing, and the like. The ROM 321 stores programs such as a startup program for the relay device 30, a DBA processing program, and setting values.

  In this embodiment, the UI side transceiver unit 303 and the NI side transceiver unit 304 use a general PHY (physical layer) that performs physical interface conversion for media conversion. However, the present invention is not limited to this, and these functions may be realized by a logic circuit connected in the same substrate. In this case, the UI side transceiver unit 303 and the NI side transceiver unit 304 are combined into an FPGA ( You may comprise with logic circuits, such as a Field-Programmable Gate Array (CPLD) and CPLD (Complex Programmable Logic Device).

  Further, the ONU-MAC unit 302 receives a LOAD signal (a signal denoted as “LOAD” in FIG. 2) to be transmitted to the OLT-MAC unit 305 at the timing of receiving the above-described Discovery GATE (that is, the timing of loading the Local Time). ), And the OLT-MAC unit 305 can know the timing when the R_ONU 31 receives the Discovery GATE using the LOAD signal as a trigger.

  As described above, the relay device 30 includes the R_ONU 31 connected to the OLT 10 and the R_OLT 32 connected to the L_ONU 22 subordinate to the relay device 30, and the user interface side of the R_ONU unit 31, that is, the UI side transmission / reception unit 303 is The standard Ethernet (registered trademark) frame is transmitted and received, and the network side interface of the R_OLT 32, that is, the NI side transmitting / receiving unit 304 is configured to transmit and receive the standard Ethernet (registered trademark) frame. Therefore, the R_ONU 31 and the R_OLT 32 can be connected without converting to a special interface.

  Next, the operation of the PON system 1 will be described with reference to FIG. 3 and FIG. In the present embodiment, the operation timings are made to coincide with each other without changing the bandwidth control periods of the upper PON 100 and the lower PON 102. In the operation of the PON system 1 described below, it is assumed that the R_ONU 31 of the relay device 30 is already connected to the OLT 10 that is the parent OLT.

  Here, as described above, the E-PON has a Discovery process for connecting a new ONU to the OLT. The OLT transmits a Discovery GATE to the ONU whose connection distance within the system specification range is unknown, and only the ONU that has not been connected performs the connection process in response to the Discovery GATE (for example, Non-patent document 1). Since the OLT itself receives an ONU frame whose connection distance is unknown (that is, the RTT is unknown), the OLT sets a time for a response of a new ONU called Discovery Window. During this Discovery Window period, uplink transmission is not assigned to the already connected ONU.

  The OLT thus has a time for the Discovery process (hereinafter referred to as “Δdba_dis”) that is different from the time for normal allocation (hereinafter referred to as “Δdba”), and is normally continuous. A Discovery GATE is periodically transmitted (for example, once every 100 times or once every 200 times) during allocation of GATE frames. Δdba_dis, Δdba, and the time ΔRep for REPORT frame aggregation (the end timing of ΔRep may be used as the timing for starting the DBA calculation) can be uniquely determined by the number of clocks in the PON system. In E-PON, the above-mentioned Local Time (unit is TQ (Time Quanta), specifically 16 ns) is generally used as a unit of these periods.

  As described above, in the PON system, each time is fixedly determined and operated as a system. However, even in the upper PON 100 and the lower PON 102 of the PON system 1 (multistage PON system) according to the present embodiment, this fixed time is determined. It is set in advance so as to operate with sharing conditions.

  The R_ONU 31 of the relay device 30 extracts a clock from a signal received from the OLT 10 when connected to the OLT 10 that is the parent OLT, distributes it as a clock used by the R_OLT 32 of the relay device 30, and performs frequency synchronization with the OLT 10. The R_ONU 31 distributes the RTT value obtained when the connection with the OLT 10 is completed to each L_ONU 22 using the extended OAM frame, and then distributes it at an arbitrary cycle. Although the RTT value is usually dominated by the distance between the OLT and the ONU, the R_ONU 31 is distributed so that it can cope with long-term changes such as fluctuations in delay time due to fluctuations in optical fiber length due to temperature, etc. To do.

  A method of timing synchronization between the OLT 10 and the relay device 30 (R_OLT 32) will be described with reference to FIG. The OLT 10 transmits a Discovery GATE (indicated by a □ including the letter “D” in FIG. 3) indicating the start of the allocation cycle at an arbitrary time T000. Here, the Discovery GATE transmitted by the OLT 10 is transmitted at the beginning of the discovery process at the beginning of the allocation sequence managed by the OLT 10, as shown in FIG. The reason why the Discovery GATE is used as a timing synchronization reference in the present embodiment is that the Discovery GATE is addressed to all of the devices under the OLT (in this embodiment, the U_ONU 20 and the relay device 30).

  The relay device 30 (R_ONU 31) loads this Local Time, T000. On the other hand, the R_OLT 32 operating using an arbitrary Local Time acquires the Local Time (T200 in this example) of the R_OLT 32 synchronized with the OLT 10 at the same timing as when the R_ONU 31 loads the Local Time T000. In the present embodiment, an example in which the local time of the R_OLT 32 and the local time of the R_ONU 31 are not aligned will be described as an example. However, the present invention is not limited to this, and the same local time as the R_ONU 31 may be loaded by the R_OLT 32. Good.

Since the local time T200 received by the R_OLT 32 is a local time shifted by the amount of the OLT 10 and the RTT, in order to synchronize the R_OLT 32 with the allocation period of the OLT 10, in this embodiment, only Δdba_dis− (2 × RTT + α) Shifted time,
T201 = T200 + Δdba_dis− (2 × RTT + α) (Expression 1)
A GATE frame is transmitted from.
Here, α represents the deadline time required for calculating the bandwidth request value of the REPORT frame transmitted from the R_ONU 31 to the OLT 10 (indicated by the symbol <1> in FIG. 3. X> indicates the corresponding part in the figure). The bandwidth calculation deadline α is a fixed time used in the R_ONU 31.

When (Formula 1) is transformed,
Tadj = T201−T200 = Δdba_dis− (2 × RTT + α) (Expression 2) This Tadj is the adjustment time from when the relay device 30 (R_OLT 32) acquires the Local Time synchronized with the OLT 10 until the GATE frame is transmitted, that is, until the operation synchronized with the OLT 10 is started. In that sense, Tadj may be referred to as “synchronization adjustment time”.

When the R_OLT 32 issues a GATE frame to each L_ONU 22 at T201, the L_ONU 22 returns a REPORT frame. The period from when the GATE frame is sent until the REPORT frame is returned is ΔGate. The R_OLT 32 aggregates REPORT frames from each L_ONU 22, and this aggregation time is ΔRep. The R_OLT 32 performs a DBA operation after the aggregation of each REPORT frame. In the example of FIG. 3, the DBA operation is completed at time T202. Therefore, Δdba in the relay device 30 is indicated by an arrow with a period “A” in FIG.
Δdba = ΔGate + ΔRep + DBA calculation time (Expression 3)
It can be expressed as In this case, the time T202 can be expressed by T202 = T201 + Δdba.

Thereafter, the processing represented by (Equation 3) is periodically executed until the next period of Δdba_dis. When the next period of Δdba_dis arrives, the local time of the OLT 10 is loaded again by the relay device 30 (R_ONU 31), and the above operation is repeated.
As a result, it is possible to establish a bandwidth allocation period synchronized with the OLT 10. In the present embodiment, by using the synchronization establishment technique based on the above principle, the DATA frame allocated by the lower level PON 102 is necessarily a bandwidth request target in the relay device 30.

  Next, the bandwidth request process will be described in more detail with reference to FIG. FIG. 4 shows a diagram after N allocation cycles have occurred for one allocation cycle A shown in FIG. That is, in FIG. 4, the DATA frame is transmitted in the L_ONU 22 (any one of the L_ONUs 22-1, 22-2, 22-3, and 22-4) subordinate to the lower level PON 102 in the period immediately before the (A + N) th period. Received (<1>).

The L_ONU 22 makes a bandwidth request to the relay device 30 (R_OLT 32) at the timing of the next REPORT frame (<2>). The relay device 30 (R_OLT 32) performs a DBA allocation operation on all the L_ONUs 22 of the lower level PON 102 from the REPORT frame of each subordinate L_ONU 22 received including the other U_ONUs 22 to give a bandwidth (<3>). Each L_ONU 22 under the relay device 30 (R_OLT 32) transmits a DATA frame in the next REPORT frame cycle (<4>).

  Here, since the timing of the relay device 30 (R_OLT 32) is synchronized with the OLT 10, all the L_ONUs 22 subordinate to the relay device 30 (R_OLT 32) transmitted to the relay device 30 (R_ONU 31) in this allocation cycle. The bandwidth is requested from the relay device 30 (R_ONU 31) to the OLT 10 (<5>).

  Here, in the present embodiment, the timings of the upper PON 100 and the lower PON 102 are linked as described above. However, in order to suppress an increase in the allocation period in each of the upper PON 100 and the lower PON 102, each operates independently. I am doing. Therefore, not all of the bandwidths assigned by the lower-level PON 102 are assigned from the OLT 10 to the relay device 30. If this occurs continuously, the upstream buffer 307 for temporary storage of the relay device 30 (R_ONU 31) may break down, and the relay device 30 may discard a frame. Therefore, in the present embodiment, the amount of buffer remaining in the upstream buffer 307 of the relay device 30 (R_ONU 31) is fed back to the DBA operation of the lower PON 102, thereby dynamically adjusting the allocation amount in the lower PON 102, thereby repeating the relay device. Frame discard at 30 is suppressed.

  Here, the frame discard in the relay device 30 is the following phenomenon. That is, the relay device 30 performs band allocation for each of the L_ONUs 22-1, 22-2, 22-3, and 22-4 in this order, for example, and receives each upstream burst frame. The received burst frame is temporarily stored in the upstream buffer 307 of the relay device 30. However, if the state in which the bandwidth output from the relay device 30 to the OLT 10 is smaller than the requested bandwidth from each L_ONU 22 continues, the upstream buffer 307 may overflow.

  Here, also in L_ONU 22, a buffer amount several times as large as the frame amount accumulated in the period of Δdba is generally secured in consideration of the allocation time from the bandwidth request and the time when transmission cannot be performed by the Discovery process. Is. Therefore, in the present embodiment, an upstream buffer 307 having a capacity equivalent to that of the L_ONU 22 is provided as a buffer of the relay device 30 (R_ONU 31) instead of a particularly large buffer. Furthermore, in the present embodiment, a threshold value is set in the upstream buffer 307 taking into account the maximum buffer amount that can be received during the period of Δdba.

For example, assuming four threshold values TH1, TH2, TH3, TH4 (TH1 <TH2 <TH3 <TH4) as threshold values taking into consideration the maximum buffer amount, the remaining data amount Dr represents the amount of still remaining frames in the upstream buffer 307. And In this case, for each threshold value, the relay device 30 (R_OLT 32) allocates a band as follows in the DBA calculation for the subordinate L_ONU 22.
(Case 1) When Dr <TH1 Allocate a band corresponding to all Δdba time.
(Case 2) When TH1 ≦ Dr <TH2, a band corresponding to 3/4 of Δdba time is allocated.
(Case 3) When TH2 ≦ Dr <TH3: A band corresponding to ½ of Δdba time is allocated.
(Case 4) When TH3 ≦ Dr <TH4: A band corresponding to ¼ of Δdba time is allocated.
(Case 5) When Dr ≧ TH4 Bandwidth allocation in that cycle is not performed.
Note that the number of thresholds and the amount of bandwidth allocation are examples, and the number of thresholds and the amount of bandwidth allocation may be arbitrarily set in consideration of the period of Δdba and system requirements.

  The above processing occurs when the OLT 10 which is the parent OLT is in a congestion state in bandwidth allocation. Even when there is no allocation to the lower level PON 102, DATA frames to be transmitted are accumulated in the relay device 30. Therefore, the accumulation in the relay device 30 can be regarded as a process equivalent to the case where the ONU under the lower PON accumulates the DATA frame according to the allocation state from the parent OLT in the multistage PON system according to the related art. (In other words, a particularly large amount of data is not accumulated in the PON system 1 according to the present embodiment).

  Next, with reference to FIG. 5 and FIG. 6, a specific method for establishing synchronization when the relay device 30 performs a bandwidth allocation operation in cooperation with the OLT 10 and a bandwidth allocation method for the relay device 30 of the OLT 10 will be described. . The OLT 10 according to the present embodiment dynamically changes the bandwidth grant amount to the L_ONU 22 under the relay device 30 according to the accumulation amount of the uplink buffer 307 of the relay device 30 (R_ONU 31). Although the discarding of the upstream frame does not occur in the relay device 30 due to the dynamic change of the bandwidth provision amount, the bandwidth control method in this case will be specifically described.

  FIG. 5 shows a part of the relay device 30 shown in FIG. 2, and FIG. 6 shows the connection between the CPU 312 in the ONU-MAC unit 302 and the CPU 319 in the OLT-MAC unit 305. Shows a flow for transmitting and receiving buffer information. The ONU-MAC unit 302 grasps the storage state of the upstream buffer 307 that temporarily stores upstream user data frames (that is, upstream user frames collected from each of the L_ONUs 22) connected to the bridge unit 315. Based on this storage state, the R_ONU 31 notifies the OLT 10 of the uplink frame accumulation amount it owns via the REPORT frame. For this reason, this storage state is shared by the CPU 312 as buffer information of the upstream buffer 307.

  As shown in FIG. 5, the user frame from each L_ONU 22 is temporarily stored in the upstream buffer 307 via the OLT-MAC unit 305-NI side transceiver unit 304-UI side transceiver unit 303-ONU-MAC unit 302. . In the upstream buffer 307, threshold values TH1 to TH4 corresponding to the remaining data amount Dr are set. On the other hand, the CPU 312 acquires buffer information from the bridge unit 315 and transmits the buffer information to the CPU 319 via the bridge unit 315 -UI side transmission / reception unit 303 -NI side transmission / reception unit 304 -bridge unit 316. When the CPU 312 and the CPU 319 are provided on the same base in the relay device 30, a memory that can be commonly used by the CPU 312 and the CPU 319 is separately provided, and the buffer information is directly transferred using the memory. Good. In this case, for example, Dual Port Memory (DPM) can be adopted as the memory. When DPM is used, the CPU 312 and the CPU 319 in FIG. 5 may be connected by a bus line, and the DPM may be connected to the bus line so that both the CPU 312 and the CPU 319 can access the DPM.

  With reference to FIG. 6, first, in step S <b> 1, the CPU 312 acquires buffer information of the upstream buffer 307 or threshold excess information. In step S <b> 2, the CPU 312 transmits the acquired buffer information of the upstream buffer 307 to the CPU 319. Next, returning to step S1, the CPU 312 periodically executes this transmission.

  Since the CPU 312 and the CPU 319 each have a MAC address, the CPU 319 receives the buffer information frame of the upstream buffer 307 by filtering its own MAC address from the main signal in step S3.

In the next step S4, the CPU 319 reflects the received buffer information in the DBA calculation process. That is, based on the received buffer information, the total bandwidth allocated to each L_ONU 22 under the subordinate PON 102 is assigned to a bandwidth allocation method corresponding to Δdba based on the threshold values TH1 to TH4 (that is, the above (Case 1) to (Case 1)). 4)) Assign according to. Thereafter, the CPU 319 returns to step S3 and repeats reception of buffer information.
As described above, in this embodiment, the CPU 312 and the CPU 319 cooperate to control the flow so that no overflow occurs in the upstream buffer 307 of the R_ONU 31.

  Here, the CPU 312 of the R_ONU 31 determines whether or not each of a plurality of threshold values provided for all the buffer sizes of the upstream buffer 307 of the R_ONU 31 is exceeded at a constant cycle. The buffer information of the upstream buffer 307 includes this determination result. That is, when the result of confirmation at a certain period is the case 1 described above, “0” is included in the buffer information and transmitted to the CPU 319 of the R_OLT 32. Similarly, in case 2 above, buffer information is “1”, in case 3 is “2”, in case 4 is “3”, and in case 5 is “4”. And transmitted to the CPU 319 of the R_OLT 32.

  With reference to FIG. 7, a method in which the CPU 319 of the R_OLT 32 receives the buffer information frame and performs the DBA calculation process will be described. First, in the DBA cycle N, a GATE frame is transmitted from the R_OLT 32 to the L_ONUs 22-1 to 22-4 under the relay apparatus 30 (<1>). Each L_ONU 22 transmits a REPORT frame including the requested bandwidth as a response to the GATE frame (<2>). For the requested bandwidth amount from each L_ONU, the CPU 319 of the R_OLT 32 executes DBA calculation processing (<3>). Here, the effective grant bandwidth for each L_ONU 22 is determined by a DBA computation processing algorithm that assigns a bandwidth according to the requested amount of each L_ONU 22 with Δdba, which is the period of the DBA computation processing, being the maximum time for total allocation. In the following description, the term “granted bandwidth to each L_ONU 22” refers to the total amount of bandwidth granted to all L_ONUs 22, and the total amount of bandwidth granted may be referred to as “Tgmax”.

  FIG. 7 shows a case where the threshold value of the upstream buffer 307 received from the CPU 312 of the R_ONU 31 does not exceed TH1 in the DBA cycle N (<4>). Therefore, in the DBA cycle N, the DBA calculation process is executed with the band granting range for each L_ONU 22 set to Tgmax = Δdba, so the bandwidth assigned to the L_ONU 22 in the DBA cycle N + 1 is Δdba (maximum bandwidth) (<5>).

  Also in the DBA cycle N + 1, the GATE frame and the REPORT frame are exchanged between the R_OLT 32 and each L_ONU 22 (<6>). In the DBA cycle N + 1, the threshold information of the upstream buffer 307 received from the CPU 312 of the R_ONU 31 is that the threshold value TH2 is exceeded within a range that does not exceed the threshold value TH3 (<7>).

  In response to this, in the DBA cycle N + 1, the bandwidth request information of the REPORT (22-1 to 22-4) frame received this time and the TH2 excess information received via the buffer information frame of the upstream buffer 307 are added, and the next The bandwidth grant amount Tgmax to each L_ONU 22 given in the DBA cycle N + 2 is determined. That is, the bandwidth calculation is performed by setting the maximum allocation time Tgmax in the DBA calculation process to Tgmax = 1 / 2Δdba (<8>), and as a result, in the DBA cycle N + 2, ½ bandwidth amount of Δdba is allocated to the L_ONU 22 ( <9>). In this case, it is preferable to execute the DBA calculation process in consideration of the REPORT request content from each L_ONU 22.

  As described above, according to the present embodiment, according to the buffer information frame of the upstream buffer 307 from the CPU 312, the total bandwidth allocated to each L_ONU 22 under the relay device 30 is controlled in the DBA processing of the relay device 30 (R_OLT 32). The overflow of the upstream buffer 307 is avoided.

  As described above in detail, in the communication device and the data transmission program according to the present embodiment, bandwidth control in which R_ONU 31 and R_OLT 32 cooperate with each other is performed in relay device 30 as a communication device. For this reason, the DBA operation cycle in the OLT and the DBA operation cycle in the relay node, which have been a problem in the multistage PON system according to the prior art, are increased, and the waiting time of the ONUs under the lower PON is increased. It is possible to suppress an increase in delay amount and an increase in required buffer capacity. Furthermore, according to the communication device and the data transmission program according to the present embodiment, even if the OLT 10 and the relay device 30 independently perform DBA operation control in a short cycle, the DATA frame may be discarded by the relay device 30. Therefore, it is possible to prevent the bandwidth from becoming unfair, and the fairness of the granted bandwidth between the L_ONUs 22 connected to the subordinates of the lower level PON 102 is ensured.

[Second Embodiment]
The communication apparatus and data transmission program according to the present embodiment will be described with reference to FIGS. FIG. 8 is a block diagram showing an example of the configuration of a relay device 30a as a communication device according to the present embodiment. FIG. 8A shows the overall configuration of the relay device 30a, and FIG. 8B shows the relay device. FIG. 8C is a block diagram showing details of the OLT-MAC unit 305 of the R_OLT 32 configuring the relay apparatus 30a. FIG. 8C is a block diagram showing details of the ONU-MAC unit 302 of the R_ONU 31 configuring 30a. As shown in FIG. 8, the relay device 30a differs from the relay device 30 according to the above embodiment only in that the total bandwidth request amount REPORT__Sum is transmitted from the OLT-MAC unit 305 to the ONU-MAC unit 302. Other configurations are the same. Therefore, the same code | symbol is attached | subjected to the same structure and detailed description is abbreviate | omitted.

The total bandwidth request amount REPORT_Sum is the sum of the bandwidth request amounts of the REPORT frames of all L_ONUs 22 connected under the R_ONU 31. A specific transmission method of the total bandwidth request amount REPORT_Sum is performed by, for example, transferring the information itself from the CPU 319 of the R_OLT 32 to the CPU 312 of the R_ONU 31 using memory transfer or the like.
This transfer may be executed by giving a trigger signal as required.

  Next, a timing synchronization method between the OLT 10 and the relay device 30a will be described with reference to FIG. Similarly to the relay device 30 according to the above embodiment, the relay device 30a also receives the Discovery_GATE indicating the start of the OLT 10 period (<1>) in order to acquire the timing, and based on this timing, the relay device 30a The DBA cycle is synchronized with the timing generated by the OLT 10.

  On the other hand, the relay device 30a (R_OLT 32) determines the allocation start timing as follows. That is, as shown in FIG. 9, when a Discovery_GATE frame (indicated by a square including the letter “D” in FIG. 9) is transmitted from OLT 10 at Local_Time = T000, the relay device 30a (R_ONU 31) transmits this Local_Time. = T000 is received (<1>). In this example, the relay device 30a (R_OLT31) receives Local_TimeT000 with its own Local_Time = T200.

The relay device 30a (R_OLT32) starts transmission of Discovery_GATE based on the information of Local_Time = T200 (<2>). The transmission start time of Discovery_GATE is T210. At this time, in relay device 30a according to the present embodiment, time T210 is determined by performing an offset calculation represented by the following (formula 4).
T210 = T200 + ΔGate + ΔRep + β− (2 × RTT) (Formula 4)
Here, β is a time required for the relay device 30a (R_OLT 32) to receive all the REPORT frames from all the subordinate L_ONUs 22 and pass the total required bandwidth to the relay device 30a (R_ONU 31) (<3>). ), A value that can be uniquely determined as a system.

Based on the same idea as the synchronization adjustment time Tadj in the relay device 30 according to the above embodiment, Tadj ′ shown in the following (Equation 5) is the synchronization adjustment time in the relay device 30a according to the present embodiment.
Tadj ′ = T210−T200 = ΔGate + ΔRep + β− (2 × RTT) (Formula 5)

  At time T210, the relay device 30a (R_OLT31) starts the Discovery process (<2>), and from T211 (<4>) after the lapse of Δdba_dis, performs bandwidth allocation at a period of Δdba similarly to the relay device 30 (< 5>) and thereafter, this bandwidth allocation is repeatedly executed until the next Discovery process cycle. As shown in FIG. 3, although the synchronization adjustment time is defined by Δdba_dis in the relay device 30, there is a difference that it is defined by ΔGate and ΔRep in the relay device 30a. However, the timing is set for each Discovery process that occurs repeatedly. Is the same as recalculating (re-synchronizing).

  In this embodiment, the Local_Time of the relay device 30a (R_ONU31) and the Local_Time of the relay device 30a (R_OLT32) are illustrated and described as being independent, but the present invention is not limited to this, and the relay device 30a (R_ONU31) The local_time of the relay device 30a (R_OLT32) may be loaded into the local_time so that the local_time of both are synchronized.

  Here, when the relay device 30a (R_ONU 31) according to the present embodiment receives all REPORT frames from all the subordinate L_ONUs 22, the total value of the requested bandwidth and the upstream buffer 307 of the relay device 30a (R_ONU 31) are accumulated. The total sum of the bands (that is, the band of the remaining data amount Dr. hereinafter referred to as “band of remaining frame”) is reported to the OLT 10 (<6>). The relay device 30a (R_ONU 31) reports the remaining frame bandwidth to the OLT 10 in addition to the total requested bandwidth of all L_ONUs 22 for the following reason.

That is, the relay device 30a (R_ONU 31) according to the present embodiment reports the sum of bandwidth request values from all L_ONUs 22 to the OLT 10, while the relay device 30a (R_OLT 32) receives the bandwidth request value and buffer information (upstream buffer 307). The bandwidth is assigned to each L_ONU 22 based on the accumulated amount in the At this time, if the upper PON 100 is congested and the OLT 10 does not give the relay device 30a a bandwidth as requested by the relay device 30a, a frame remains in the upstream buffer 307 of the relay device 30a (R_ONU 31). If there is no bandwidth request from each L_ONU 22 with a frame remaining in the upstream buffer 307, the bandwidth request for the OLT 10 from the relay device 30a is also eliminated.
In this case, since the frame remaining in the upstream buffer 307 is not transmitted to the OLT 10, the frame continues to remain in the upstream buffer 307.

  Therefore, the relay device 30a (R_ONU 31) according to the present embodiment sets the sum of the bandwidth request values of all L_ONUs 22 under the lower PON 102 and the remaining frame bandwidth in the upstream buffer 307 of the relay device 30a (R_ONU 31) as the upper level. Report to OLT10 of PON100. As a result, if there is room in the bandwidth that the OLT 10 gives to the lower-level PON 102 (relay device 30a), the frames stored in the upstream buffer 307 and each L_ONU 22 at the time when the OLT 10 grants the bandwidth to the relay device 30a. Can be transmitted to the OLT 10 (predicted) from the relay device 30a. As a result, even if there is no bandwidth request from the L_ONU 22, the frame does not remain in the upstream buffer 307 of the relay device 30a (R_ONU 31).

  Next, a band allocation method in the relay device 30a will be described with reference to FIG. FIG. 10 shows the state of the Nth cycle when the one allocation cycle Δdba shown in FIG. First, the L_ONU 22 (any one of L_ONUs 22-1, 22-2, 22-3, and 22-4) under the relay device 30a receives the DATA frame (U1) in the period immediately before the A + N period (<1>). In the A + N cycle, the L_ONU 22 makes a bandwidth request to the relay device 30a (R_OLT 32) via the REPORT frame (<2>). The relay device 30a (R_OLT 32) receives all REPORT frames from all subordinate L_ONUs 22 and sends the total bandwidth request received after ΔRep from the start of reception to the relay device 30a (R_ONU 31) (<3>).

  At the same time, the relay device 30a (R_OLT 32) also performs DBA operation processing on the subordinate L_ONU 22 (<4>), and performs bandwidth allocation via the GATE frame in the next A + N + 1 cycle (<5>). On the other hand, the relay device 30a (R_ONU 31) transmits a REPORT frame based on the requested bandwidth value received from the relay device 30a (R_OLT 32) and the remaining frame bandwidth (<6>).

  The same process is performed in the A + N + 1 period following the A + N period. That is, the L_ONU 22 that receives the DATA frame (U2) in the A + N cycle (<7>) makes a bandwidth request via the REPORT frame to the relay device 30a (R_OLT 31) in the A + N + 1 cycle (<8>). The relay device 30a (R_OLT 32) sends the sum of the bandwidth request values received from all the subordinate L_ONUs 22 to the relay device 30a (R_ONU 31). Further, the relay device 30a (R_ONU 31) declares the sum of the bandwidth request values received from all the L_ONUs 22 to the OLT 10, but at this time, as a result of the allocation of the OLT 10 in the A + N cycle, a frame remains in the upstream buffer 307. In addition, the remaining bandwidth is also added and reported to the OLT 10 (<9>).

  On the other hand, the relay device 30a (R_OLT32) receives buffer information indicating the relationship between the frames remaining in the upstream buffer 307 of the relay device 30a (R_ONU31) and the thresholds TH1 to TH4 as a result of the allocation of the OLT 10 in the A + N cycle. (<10>). For this reason, the DBA calculation process in the A + N + 1 cycle is executed after setting the maximum band to be allocated according to the above-described cases 1 to 4 based on the buffer information (<11>). Thereafter, the above operation is repeated, and the bandwidth allocation according to the present embodiment is executed.

  As described above in detail, according to the communication device and the data transmission program according to the present embodiment, it is possible to perform band allocation in each of the upper PON 100 and the lower PON 102 without increasing the band allocation period Δdba. Also, by quickly transmitting the sum of the bandwidth requests of the L_ONUs 22 under the subordinate PON 102 to the OLT 10, one allocation period is about twice that of a normal PON system as in the conventional multi-stage PON system. As a result, the occurrence of the problem that the DATA frame that is not in time for the bandwidth request is kept waiting for a long time is suppressed. Furthermore, according to the communication apparatus and the data transmission program according to the present embodiment, it is possible to make a bandwidth request for all L_ONUs 22 as early as one Δdba cycle. Further, the DBA calculation processing of the lower PON 102 and the upper PON 100 is executed separately, but by feeding back the allocation state of the upper PON 100 to the allocation state of the lower PON 102, frame loss due to buffer overflow in the relay device 30a is eliminated. It is possible to suppress an increase in buffer capacity in the relay device 30a.

1 PON system 10 OLT
20, 20-1, 20-2, 20-3 U_ONU
22, 22-1, 22-2, 22-3, 22-4 L_ONU
30, 30a Relay device 31 R_ONU
32 R_OLT
50, 52 Optical splitter 100 Upper PON
102 Lower PON
301, 306 PON optical transmission / reception unit 302 ONU-MAC unit 303 UI side transmission / reception unit 304 NI side transmission / reception unit 305 OLT-MAC unit 307 Upstream buffer 308, 323 PON transmission unit 309, 322 PON reception unit 310, 317 OAM transmission / reception unit 311 318 MPCP transmission / reception units 312, 319 CPU
313, 320 RAM
314, 321 ROM
315, 316 Bridge part,
324 PLL

Claims (9)

A communication device that connects a higher order PON connected in the upstream direction and a lower order PON connected in the downstream direction,
An ONU unit connected to a parent OLT that is an OLT of the upper PON;
An OLT unit connected to the ONU unit and connected to a child ONU that is an ONU of the lower level PON,
The ONU unit notifies the OLT unit of a first time at which a discovery signal that is a response request for an unknown connection device is received from the parent OLT,
The OLT unit starts transmission of a gate signal that permits transmission of an upstream signal to the child ONU at a second time after a predetermined adjustment time has elapsed from the first time, and the child ONU And summing up the requested bandwidth value included in the uplink signal from the total requested bandwidth value, and completing the bandwidth allocation processing for the child ONU based on the aggregate requested bandwidth value at a third time, The communication apparatus is characterized in that synchronization with the parent OLT is established by repeating the processing from the time until the third time until the next discovery signal is received. The ONU unit includes a buffer unit that temporarily holds data received from the child ONU, and transmits buffer information indicating a remaining state of data in the buffer unit to the OLT unit.
The communication apparatus according to claim 1, wherein the OLT unit performs the bandwidth allocation process using the total requested bandwidth value and the buffer information. The buffer information is determined in advance according to an occupation ratio in the buffer unit of data remaining in the buffer unit, with Δdba being a time difference between the second time and the third time as a maximum bandwidth that can be allocated. The communication apparatus according to claim 2, wherein the communication apparatus is information indicating a ratio of a band that can be provided to the maximum band. The time required for the OLT unit to perform discovery processing for the child ONU is Δdba_dis, the delay time when a signal is transmitted from the parent OLT to the ONU unit is RTT, and the OLT unit transmits a gate signal to the child ONU When the delay time from the possible time to the time when the ONU unit transmits the bandwidth request signal to the parent OLT is α, the adjustment time is:
Δdba_dis− (2 × RTT + α)
The communication device according to any one of claims 1 to 3, wherein the communication device is indicated by. The timing at which the OLT unit completes the transmission of the gate signal and the timing at which the ONU unit makes a bandwidth request to the parent OLT are synchronized with each other. The communication apparatus according to claim 1. The OLT unit transmits the gate signal to the child ONU, the time until the OLT unit receives the requested bandwidth value from the child ONU is ΔGate, and the time for summing the bandwidth request value from the child ONU is ΔRep , Β is a delay time until the OLT unit aggregates the total required bandwidth value until transmission to the ONU unit, RTT is a delay time when a signal is transmitted from the parent OLT to the ONU unit, and the OLT unit is the child When the time required for performing the discovery process for the ONU is Δdba_dis, the adjustment time is
ΔGate + ΔRep + β− (2 × RTT) + Δdba_dis
The communication device according to any one of claims 1 to 3, wherein the communication device is indicated by. The timing at which the OLT unit completes reception of a bandwidth request value from the child ONU and the timing at which the ONU unit makes a bandwidth request to the parent OLT are synchronized. The communication apparatus according to any one of claims 3 and 6. The ONU unit includes a buffer unit that temporarily holds data received from the child ONU, and the total required bandwidth value and data remaining in the buffer unit in response to a gate signal transmitted by the parent OLT The communication apparatus according to claim 6 or 7, wherein a bandwidth value corresponding to the amount is reported. An upper PON connected in the upstream direction and a lower PON connected in the downstream direction and connected to a parent OLT that is an OLT of the upper PON, and connected to the ONU unit and the lower PON A program for transmitting data using a communication device having an OLT unit connected to a child ONU that is an ONU of a PON,
Computer
First control means for causing the ONU unit to notify the OLT unit of a first time at which the ONU unit receives a discovery signal that is a response request for an unknown connection device from the parent OLT;
At a second time when a predetermined adjustment time has elapsed from the first time, the OLT unit starts transmission of a gate signal that permits transmission of an upstream signal to the child ONU, and the OLT unit Completes at a third time the totaling processing for summing up the requested bandwidth value included in the uplink signal from the child ONU as the total requested bandwidth value and the bandwidth allocation processing for the child ONU based on the total requested bandwidth value And second control means for establishing synchronization with the parent OLT by repeating the processing from the second time to the third time until the next discovery signal is received,
A data transmission program characterized by functioning as: