MXPA06004386A - Dynamic bandwidth distribution method in passive optical network - Google Patents

Abstract

The present invention provides a dynamic bandwidth distribution method in passive optical network, which includes:Classifying the service related to the communication process between ONU and OLT according to the different transmission requirements, and assigning the differenl priority to every classified service;Performing the service data transmission authorization for the service port of every classified service in accordance with the order from high priority to low priority, and recording the authorization message to the vMAC authorization message list;Reading the authorization message on each un-authorized port of the identical ONU from vMAC authorization message list;Determining the transmission starting time of authorization data, generating the downlink authorization information involved the authorization message on every authorized transmission port of current ONU and the authorization data transmission starting time, and transmitting it to current ONU. The bandwidth distribution method of the present invention adapts different type service requirements, improves the bandwidth usage, and realizes the equitable bandwidth distribution .

Description

METHOD FOR ALLOCATION OF DYNAMIC BAND WIDTH IN PASSIVE OPTICAL NETWORK FIELD OF THE INVENTION The present invention relates to an optical communication field, especially to a method for assigning dynamic bandwidth in Passive Optical Network (PON, for its acronym in English).

BACKGROUND OF THE INVENTION As a new broadband access fiber optic technology that covers the "last mile", the PON does not require an active elbow and only a simple optical splitter needs to be installed. Therefore, the PON has the advantage of saving resources of optical cable, sharing resources of width of walk, saving in the investment of the space for the equipment, being a device of high security with rapidity of network of configuration low cost of the construction of the network, etc. As a major type of the present optical access network the PON has been widely applied. The PON takes the structure of master-slave control point-to-multipoint. With respect to Figure 1, an Optical Line Terminal (OLT) is a master node device, which is connected to a plurality of Optical Network Units (ONUs). through an optical splitter, while the UN registration procedures, authorization of the data waiting to be transmitted, the interaction of messages from the Media Access Control (MAC) sublayer, and so on, they are made in the OLT. The ONU is a slave node device, which responds to the OLT discovery message, which reports the sizes of data waiting to be transmitted, transmits the data of each of the ports in accordance with the granting information, as well as a plurality of client terminals are entered into the UN to support multiservice access. With this existing frame structure in the optical access network, the problem of how to reasonably allocate bandwidth and effectively use the resources to satisfy the service requirements, is of great importance for the PON MAC scheme. In the optical access network there are two ways to allocate bandwidth: Static Bandwidth Allocation (SBA) and Dynamic Bandwidth Allocation (DBA, for its acronym in English). In relation to the SBA, the ONUs or their port are assigned with fixed bandwidth and the unused time slot can not be occupied arbitrarily. The disadvantage of the SBA is the low proportion of bandwidth use and the poor adaptive faculty in relation to the auto-probability service with a high proportion of bursts. The DBA algorithm is a mechanism or algorithm for quickly reallocating bandwidth based on the current customer service requirements, in relation to the DBA, the PON bandwidth use ratio can be increased by dynamically adjusting the bandwidth, with reference to the requirements of the ONUs burst service. At present there are many methods of allocation of dynamic bandwidth. For example, a method of assigning dynamic bandwidth, Interlace Sounding with Adaptive Cycle Time (IPACT) for PON; a special DBA protocol for APON, which is presented in standard G.983.4 by Telecommunications International Telecommunication Union Standardization Sector (ITU-T, for its acronym in English). The main idea of IPACT is that it consults the next UN before the data sent by the previous ONU reaches the OLT, to determine whether or not to authorize and how much to authorize, etc., as well as to determine the start time of the transmission. data of the next ONU in accordance with the calculated arrival time of the last data bit of the current ONU. In relation to the second scheme, a DBA protocol is presented in ITU-T G.983.4, the UN service is divided into several types of Transmission Containers (T-CONT, for its acronym in English). In accordance with the priority of the service, there are four types of T-CONT; in descending priority, these are T-CONT1, T-CONT2, T-CONT3 and T-CONT4, while T-CONT5 is adapted to all services to reduce the number of transmission containers. Each type of T-CONT corresponds to a specific requirement for allocation of bandwidth. There are four types of bandwidth allocation requirements: constant bandwidth, assured bandwidth, unsecured bandwidth and best effort transmission bandwidth. In addition, there are three strategies in this DBA protocol: Report Without Status (NSR), Status Report (SR) and combined. In the first technological scheme mentioned above, the polling period varies adaptively with the amount of data, producing fluctuation of data transmission in the same service, so that the requirement of low delay and delay fluctuation for the service in real time does not It is satisfied. In the second technological scheme, a constant sequence sounding of All versus All is adopted, which is especially for the APON application oriented to connection with a fixed iongiiud of the Protocol Data Unit (PDU, for its acronym in English). Additionally, the bandwidth update mechanism is special for T-CONT and the bandwidth is assigned not according to different ports or different services. In the two previous schemes, given that different assignment strategies for different services are not taken into account, they are applied in a single service system in the practical application as in a system that provides only a search service or voice service. This type of system can not satisfy the requirement of Quality of Service (QoS) of more than one type of service, nor can it assign bandwidth in accordance with the client's service ports. In addition, when allocating bandwidth, the window allocation granting messages are generated basically in sequence of port reporting, without considering the delay requirement of the different ONU ports, and the bandwidth usage ratio is also low. Additionally, there is no dynamic procedure if the ONUs are offline.

BRIEF DESCRIPTION OF THE INVENTION One purpose of the present invention is to provide a method for the allocation of dynamic bandwidth in the PON. Service transparency is achieved in this dynamic bandwidth allocation method, which is applicable for different types of service requirements. Through the allocation of bandwidth for different service ports, the bandwidth use ratio is increased and the bandwidth allocation is made in an appropriate manner. In order to achieve the above purpose, in the present invention a method of assigning dynamic bandwidth for PON is provided, including PON an OLT and a plurality of ONUs having access to the OLT, wherein the method comprises: a ) Classify the services that must be communicated between the OLT and the ONUs in a plurality of types of service in accordance with different transmission requirements, as well as giving a different priority to each type of service; b) Authorize service ports for each type of service to transmit service data in sequence descending from said service priorities, as well as record the granting of information from the service ports obtained from the authorization; c) Read said granting information of each port of service to be granted from the same ONU; d) And program a given start time of data transmission of each service port to be granted from the current ONU, generating downlink granting messages that include both said granting information and said granted start time of data transmission. of each granted port of said ONU acted, transmitting messages of downlink granting to said current ONU. The method also comprises an aging information flow of ONU: A. Request the status of the ONUs one by one, determining if the status of the current ONU is invalid; if so, return to step A, continue with the question; Otherwise, proceed to step B.

B. Determine if there are MPCP messages that have been reported by said current ONU in the current bandwidth allocation polling period; if so, restart an Active Expiration Time Count of that current ONU and proceed to step C; otherwise, proceed to step C directly. C. Determine if the value of said Active Expiration Time Count of said current ONU exceeds the established off-line threshold, establish the status of the current ONU as invalid, release the corresponding resources of this ONU and proceed to step D; otherwise, proceed to step D directly. D. Determine if all ONUs are requested; if so, terminate the aging flow of the ONU information from the current bandwidth allocation polling period; otherwise, return to step A. The method also includes: Generating a table of UN Status Information indexed by ONUID, which must store all the status information of the ONUs that is generated in accordance with the communication of MPCP messages between the ONUs and the OLT. Generate an Active Expiration Time Count table of UN indexed by ONUID, which includes said Active Expiration Time Counter and reported indicators to indicate if MPCP messages have been reported by the corresponding ONUs.

Said step of determining step A comprises: in accordance with the ONUID index, reading the status information of the ONU from the logging of UN Status Information one by one, determining if the current ONU is invalid in accordance with said information. of UN status. Said step of determining step B comprises: reading the table element of said active ONU from the table of Active Expiration Time Count, determining if there is a reported indicator in the corresponding table element of said current ONU; if so, it can be concluded that MPCP messages have been deported in the current bandwidth allocation polling period; otherwise it can be concluded that no MPCP messages have been reported. Between step C and step D, it also includes: removing the indicator reported from that current ONU in the UN Active Expiration Time Count table. Said determination step of step D comprises: determining whether all the table elements of the UN Status Information table are read; if so, it can be concluded that all UNs have been requested; otherwise, it can be concluded that all the ONUs have not been requested. The method also includes generating a vMAC Grant information table indexed by ONUID, which includes granting information of each of the service ports of the ONUs and indicators provided to indicate whether the corresponding service ports are authorized. The step of recording grant information in step b) comprises: registering said grant information in the vMAC Grant information table, establishing the granted indicator of said service ports granted as authorized. The step of reading the grant information in step c) comprises: searching the table elements corresponding to the service ports of the same ONU in the vMAC Grant information table in accordance with the ONUID index, searching the ports of service granted in accordance with the granted indicator, read the granting information of the service ports granted. And, after step d), it comprises: establishing the indicator granted as negative of the service ports that have read the granting information. The method also includes: Generating the UN Status Information table indexed by ONUID, which must store status information of each ONU that is generated in accordance with the communication of MPCP messages between the ONUs and the OLT. Before step c) comprising: reading the status information of the ONUs one by one from the UN Status Information table in accordance with the ONUID index, determine if the current ONU status is invalid in accordance with the information of current UN status; if so, return to read the following UN status information in the UN Status Information table; otherwise, proceed to step c). Preferably, said traffic sorting by priority in descending sequence in step a) as a fast sending service, automatic detection MPCP message service, MPCP message service without automatic detection, MF service, Secured Delivery service and Sending Best Effort In relation to services except the automatic detection MPCP message service, said authorization step in step b) comprises: b11) confirming the current current service port to be granted in accordance with the activation status of the link service upward. b12) in accordance with the residual bandwidth resource in the current bandwidth allocation polling period, determine whether the current bandwidth resource is available for the constant amount of data from that current service port to be granted for the automatic detection MPCP message service, or for the reporting information of said current service port to be granted for the other types of services; if so, go to step b13); otherwise go to step b15). b13) authorize said current service port to be granted to transmit service data and record the current grant information. b14) update the residual bandwidth in the current bandwidth allocation polling period and the relevant information of that current service port to be granted. b15) and determine if there are service ports not granted service of current priority; if so, return to step b11); otherwise authorize the ports of the next priority service. This method also includes: Generate a Register of Bitmaps of Active ONU of Service for each type of service to store active information that indicates if the type of service is activated in the ONUs. Generate a Table of Bits Map of Active Port of Service for the services with port as unit, indexed by ONUID, to store active information that indicates if this type of service is activated in the ports of access service of the ONUs. Generate a vMAC Report Information table indexed by ONUID, to store report information of the access ports of the ONUs. Said step b11) comprises: polling the Bitmap record of the Active Asset of Service and the Bits Map table of the Active Port of Service of each type of service in descending sequence of priorities, discovering a service port with positive active information and confirm the service port as said current service port to be granted. Prior to step b12) also comprising: discovering current service port report information to be granted from the vMAC Report Information table; And, in step b15), determine if there are unread table elements in the Current Service Active ONU Bitmap record and the Current Service Active Port Bitmap table; if so, return to step b11); otherwise, request the registration of the Bits Map of the corresponding Active Service Asset and the Bits Map table of the Active Port of Service for the service with the following priority. Preferably, for the fast sending service, said granting information comprises the start time of data transmission and data transmission sizes. Said reporting information includes reported sizes of data waiting to be transmitted. For the fast sending service, said step of programming the start time granted in step d) comprises: taking said start time of data transmission of said granting information as the granted start time of data transmission. For the non-automatic detection MPCP message service, said grant information comprises reserved field and type of downlink MPCP messages, wherein said type of MPCP messages comprises Discovery GATE, Normal GATE and REGISTER &GATE.

Said report information comprises reserved field and field of type of MPCP messages. For the MF service, said granting information comprises reserved field and data transmission sizes. Said reporting information includes reported sizes of data waiting to be transmitted. For the Insured Shipping service, said granting information includes amount of granting deficit and data transmission sizes of the corresponding service ports. Said reporting information includes reported sizes of data waiting to be transmitted. For the Best Effort Delivery service, said grant information includes data transmission sizes. Said reporting information includes reported sizes of data waiting to be transmitted. The method also comprises: generating a Bandwidth Information table for those service ports requesting bandwidth control, for storing quantum in each bandwidth allocation polling period. The determination step for those service ports requesting bandwidth control in step b12) comprises: discover the transmission quantum of the current service port from the Bandwidth Information table, determine whether the resource width Current bandwidth is available in accordance with the reporting information from said current service port to be granted, transmit quantum of said current service port to be granted and said residual bandwidth resource in the period of bandwidth allocation. Preferably, generate a start counter for the automatic detection MPCP message service. The authorization step of the service ports in step b) includes: b21) triggering said start counter, when said counter reaches the configured threshold, start the detection procedure of the automatic detection MPCP messages; if a message is detected MPCP automatic detection, proceed to step b22); otherwise, restart the start counter and return to step b21). b22) Determine if the current bandwidth resource is available for the constant amount of data from that current service port to be granted, in accordance with the residual bandwidth resource in the bandwidth allocation period. current band of the automatic detection MPCP message service; if so, proceed to step b23); otherwise, authorize the ports of the next priority service. b23) restart the start counter and return to step b21). Preferably, for the automatic detection MPCP message service, said grant information comprises reserved field and type of downlink MPCP messages, wherein said MPCP message type comprises Discovery GATE, Normal GATE, REGISTER &GATE. Preferably, after step d), the method also comprises: determining if there is still any ONU with an unlicensed port; if so, return to step c); Otherwise, terminate the current procedure. Preferably, the type of said downlink granting messages of the method is a downlink MPCP message GATE. Preferably, said uplink grant information of the present method being transmitted by REPORT messages. Preferably, said bandwidth allocation polling period of the present invention is a virtual frame period. It is obvious from the technical scheme mentioned above that this method has the following advantages: it performs dynamically the allocation of bandwidth directed to different services with different priority levels, in order to satisfy the requirements of different types of services and implement transparency in the toilet. When generating a grant message, authorize the start time of data transmission for a group of access ports for the same ONU. In this way, all the windows granted from different access service ports for the same ONU, are consecutive with no protection band inserted, the protection bandwidth for data transmission is reduced and the ratio of use of the width is increased of band. In addition, the aging mechanism of UN information is adopted in the present invention. By monitoring the status of the ONU in real time, the offline ONU can be processed dynamically, so that the occupied resource is released. In comparison with existing technologies, the technological scheme of the present invention has visible beneficial effects. In this scheme, the allocation of dynamic bandwidth is transparent in relation to services, to satisfy the requirements of more than one service, such as the transmission of TDM / EF, AF, DF and MPCP messages, as well as the assured transmission of CPU / OAM. The beneficial effects of this scheme also include: a greater proportion of bandwidth use, an adequate allocation of bandwidth, better resistance, better performance in real time, more effective use of the bandwidth resource and the possibility of avoiding the waste of resources.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 is a schematic map of a Passive Optical Network. Figure 2 illustrates the structure of the Bitmap record of the Active Asset of Service and the corresponding table of Bit Map of the Active Port of Service, in accordance with a preferred embodiment of the present invention. Figure 3 is a structure of the vMAC Report information table in accordance with a preferred embodiment of the present invention. Figure 4 illustrates the authorization flow of different services in accordance with a preferred embodiment of the present invention. Figure 5 is a vMAC Grant Information table structure in accordance with a preferred embodiment of the present invention. Figure 6 is a structure of the UN Status Information table in accordance with a preferred embodiment of the present invention. Figure 7 is a flow chart of the method for generating grant information in accordance with a preferred embodiment of the present invention. Figure 8 is a UN Active Expiration Time Count table structure in accordance with a preferred embodiment of the present invention. Figure 9 is a flow diagram of the information aging method of UN in accordance with a preferred embodiment of the present invention.

DETAILED DESCRIPTION OF THE INVENTION To clarify the purpose, the technological scheme and advantages of the present invention, the present invention will be described in detail with reference to the attached drawings. In the present invention, a method of assigning dynamic bandwidth in Passive Optical Network is provided. This method comprises service registration, allocation of bandwidth, generation of messages of granting and aging of information of UN, and so on. First, different services are classified in priority sequence and receive a different data structure, so that they perform different assignment strategies. When allocating bandwidth, different access service ports to the same ONU are distinguished in terms of different types of service and processed separately by assigning a corresponding window size. When granting messages are generated, different ports of access service to the same ONU are processed collectively, the start time of data transmission is assigned and, in this way, the granted windows of the different ports of access services to the same ONU are consecutive, without any protective band inserted, so that the bandwidth use ratio increases. In addition, the status of the ONUs is monitored in real time in the present method and the offline ONU is dynamically released. Here, the service port mentioned is a port that transmits a certain service. The service port is a logical conception, which does not really correspond to a physical UN port. For example, if a physical port transmits multiple services at the same time, this port can be divided into more than one logical service port. A preferred embodiment of the present invention is introduced in detail in the following manner. To process different service ports separately, different services in a PON must be classified into a plurality of service types. In addition, the data structures of different services, as well as different ports, must be configured for the system to process separately. In a preferred embodiment of the present invention, in a priority descending sequence, traffic is classified into the following services: The Fast Shipping service: as Temporary Division Multiplexing (TDM), Expedited Shipping (EF) , by his sigias in English), et cetera. These services can be named with the common EF service name. This type of service needs to be sent in real time, with little delay and delay fluctuation with no bandwidth limit, so it has the highest priority. The Multipoint Control Protocol message service (MPCP): The MPCP message is used for normal MPCP communication, which includes the automatic detection MPCP message service and the MPCP message service without automatic detection. The relevant parameters of the MPCP message are constant in their entirety. In the present invention, the MPCP message of a ONU is used to judge whether this ONU is online. The Mandatory Delivery Service (MF): as a message from the Central Processing Unit and an OAM message, while the messages must be transmitted correctly in this type of services that is used for Operation Management, Maintenance and Provisioning (OAM &P, for its acronym in English) and communication between the OLT and the ONUs has, therefore also a very high priority. The Insured Shipping service: as Insured Shipping (AF, for your sigias in English); The minimum bandwidth specified should be ensured in this service, but with a low real-time requirement. The Best Effort Delivery Service: as Pre-Set Delivery (DF) there is no minimum bandwidth requirement, so it has the lowest priority. In the present invention, the form of collective control of reporting information is adopted with the administration status of the OLT of each ONU. In a preferred embodiment of the present invention, the IEEE 802.3ah standard is adopted in the automatic detection of the ONU and the information interaction of the registration path. When the data transmission authorization is being carried out, in order to conduct polls in accordance with the various services and priority sequences, in a preferred embodiment of the present invention, the Bitmap record of the Active Asset Service is configured to store UN activation information of different services, such as Active MPCP Income Binary Map (IMAB) of the MPCP message service, Active Binary Map of Mandatory Income Sending (IMFAB, for its English acronym) of the MF service, Active Bits Map of Expedited Entry of Delivery (IEFAB, for its initials in English) of the EF service, Map of Active Bits of Insured Sending of Income (IAFAB, for its acronym in English) of the service AF and Map of Active Bits of Preset Shipping of Income (IDFAB, for its acronym in English) of the DF service. Among others, some services take a port as a unit, such as the services of EF, AF and DF; then, the Bits Map table of the Active Port of Service must be configured in these services to store active information of the ports, such as the Table of Active Bits of Expedited Incoming Delivery (IEFABT) of the EF service. , Table of Map of Active Bits of Insured Shipping of Income (IAFABT, for its acronym in English) of the AF service and Table of Map of Active Bits of Preset Shipping of Income (IDFABT, for its acronym in English) of the service DF. With respect to Figure 2, the left frame is the Bitmap record of the Active ONU of Service, which comprises N bits corresponding respectively to N ONUs and is used to indicate if this service is active, for example with 1 as active and 0 as inactive. The right box is the Bits Map table of the Active Port of Service, which comprises N table elements corresponding to the ONUs, with the contents of each of the tables storing information relative to each of the ports of access to the Corresponding ONU, and this information indicates whether the port is active or not, for example with 1 as active and 0 as inactive. In the system, these tables are placed in a registry and can be indexed in relation to the relevant ONU table elements in accordance with the corresponding Optical Network Unit Identifier (ONUID), convenient for reading information active service port. It is necessary to specify that during the course of the service registration, the Logical Link Identifier (LLID) assigned to the ONU has the format as ONUID, Active Bitmap K which is convenient for indexing different active information of service and bandwidth allocation information using ONUID and compensation of each port, which is also convenient to develop multicast service. The UN MPCP port is activated during registration and the registration information is stored in IMAB. The MF port is activated after the ONU is registered and the registration information is stored in IMFAB. Through this, each ONU is preset to operate with OAM and CPU service. After successful registration, for those services with a port as a unit, the ONUs request bandwidth from the OLT, while the bandwidth information is supplied by the network administrator through the interaction with the CPU. The bandwidth information includes active information of the service port EF, active information and bandwidth of the service port DF and AF. By this, since the bandwidth in any period of the virtual frame of the EF service is constant, it is not necessary then to include the port bandwidth in the bandwidth information. Active information of the EF services ports, AF and DF are stored, respectively, in IEFAB, 1AFAB, IDFAB and the corresponding table element of IEFABT, 1AFABT, 1DFABT. The corresponding bits of IEFAB, IAFAB and ÍDFAB indicate, respectively, if there is active information in the EF, DF or AF services of the ONU. The corresponding table elements of IEFABT, IAFABT and IDFABT indicate, respectively, which UN ports are active AF, DF and ED services. In relation to those ports that request constant bandwidth, such as EF, since the bandwidth in any period of the virtual frame is constant, a record with constant bandwidth information can be established to configure the constant bandwidth, or A predefined bandwidth can be predefined, after acquiring the port type through the information interaction, after which the data can be transmitted over this preset bandwidth. In relation to those ports that request dynamic bandwidth control, such as DF and AF services, the Bandwidth Information Table (BIT) is established to store bandwidth provision information. For convenience, the provisioning bandwidth is converted and stored in a transmission quantum format. Here, the transmission quantum means the amount of data that is allowed to transmit in each period of constant sounding. This transmission quantum is calculated by multiplying the provisioning bandwidth by the polling period, with Byte or Word as the unit. In the present invention, the virtual frame period of the EF service is adopted as the polling period of the bandwidth allocation operation. Similarly, by indexing ONUID and Compensation of each port, the corresponding table element content is read and the provision bandwidth information is acquired from the BIT table by the system. When interacting with the network administrator, the system OLT can dynamically operate and modify relevant bits of the Active Service ONU BIT Map record, activate the contents of relevant table elements in the Service Port Bitmap table Active and the BIT. In this way, the reconfiguration and dynamic insertion / deletion of the bandwidth information of the ONU service port are performed. In each probing period, the ONU reports the data sizes that are expected to be transmitted to the OLT through the REPORT message of the IEEE 802.3ah standard. In accordance with the IEEE 802.3ah standard, each REPORT message comprises 8 reported element dispositions of data sizes. In addition, the corresponding relationship between each disposition and the ONU port is predefined, so that the OLT can specify which port are the reported sizes. Meanwhile, the valid indicator of each disposition is established, for example each bit of 8 bits indicates, respectively, if each of the dispositions is valid or is active, with bit 0 indicating disposition 0, bit 1 indicating disposition 1 , etc. For example, bit 0 indicates that the corresponding arrangement is inactive, while 1 indicates that it is active or valid. When more than 8 ports have access to the UN, multi REPORT message attributes are adopted and indicators are transmitted in the messages to indicate the continuity of the multi REPORT messages in a period, so that OLT can determine, starting from the messages of multi REPORT, to which port corresponds the item reported. In order to separately process the reporting information of each port, the OLT stores the reported data sizes that are expected to be transmitted in the vMAC Report Information Table (RIT) sorted by MAC Virtual ( vMAC, for its acronym in English). With respect to figure 3, the structure of the RIT is shown on the left: each table element corresponds to each of the parts of the reporting information of each UN port, while the elements are indexed by ONUID. In relation to the general service data, such as the general delivery service, CPU and OAM, the table element comprises a reported indicator ("Reported Flag") and the reported data sizes waiting to be transmitted ("Report") . Here, "Reported Flag" indicates whether vMAC should report the data stored in the current virtual frame period. The numerical unit of "Report" is the same as that of the transmission quantum, in particular Byte or Word. For the MPCP message service, with respect to Figure 3 on the right, the Reported Table elements, the uplink MPCP message type ("Type") and the reserved region ("Reserved") are included. in the RIT. Here, "Type" shows the type of this uplink MPCP message, such as "REGISTER_REQ", "REGISTER-ACK", "Normal REPORT", etc., clarifying to the OLT what UN record period is there and defining what type of downlink MPCP message should be generated. For the EF service, as soon as the virtual frame period is set, the port can only transmit data of constant sizes in a virtual frame period, so there is no need to report the data sizes waiting to be transmitted. The allocation of bandwidth, in particular the authorization of the ONU by the OLT to transmit service data, is in fact a probe granting action in accordance with the priority of the service. This action is performed by the OLT in accordance with the active port information, bandwidth allocation information and ONU reporting information. To guarantee the QoS of the EF service, the virtual frame period is established as the polling granting period. The bandwidth assignments for other services are also made in this virtual frame period in priority sequence. At the beginning of the virtual frame period the allocation of EF bandwidth is made. The MPCP, MF, AF and DF services are performed in sequence in the residual time. The ascending service data transmission method in the embodiment of the present invention will be described in detail with reference to FIG. 4. In this mode, in descending priority sequence, the services are the EF service, the MPCP message service of automatic detection, the MPCP message service without automatic detection, the MF service, the AF service and the DF service and these services are consulted and granted sequentially. For the convenience of sending a grant message to each type of uplink service by the OLT, the Grant Information Table (GIT, for its acronym in English) of vMAC is established in this case implemented to store information from granting of transmission authorization. In the uplink reception procedure, the GIT data is read and the grant information is created by the OLT. And, then, the granting information is transmitted by the OLT to the corresponding ONU and the allocation of dynamic bandwidth is terminated.Figure 4 shows the authorization flow of each type of service in the present modality. Step 401: The OLT confirms the current service port to be granted in accordance with the active uplink service information or triggers the automatic detection MPCP message service start counter. For services such as EF, MF, AF, DF and MPCP messages without automatic detection, the current service port to be granted can be confirmed by consulting the Bitmap records of the Active Asset Service and the Map tables. of Active Bit of Income. The specificities are: consult the Bit Map of the Active Asset of Service of each service in sequence of descending priority. For the MFP and MPCP message service without automatic detection, the Bitmap record of the Active Asset of this service is queried to discover the ONU activated by the type of service through the active information of the table element and This UN port is taken as the one to be granted. For those services with a port as a unit, such as EF, AF and DF, the Bitmap record of the Active UN of this service is consulted to discover the ONU activated by this service. Subsequently, the Bitmap table of the Active Port of Service of this service is searched with the ONUID of the ONU found and, in addition, it is confirmed with which port of this ONU this service is activated.

For the automatic detection MPCP message service, to make the new online ONU register dynamically, an automatic detection MPCP message service start counter is established. MPCP automatic detection starts as soon as the counter reaches the configured threshold. If the service is detected MPCP automatic detection, go to step 402; otherwise, we return to step 401. In this way, the periodic MPCP automatic detection procedure is implemented and not much bandwidth is occupied. Step 402: the OLT looks for the report information of the current service port to be granted in the RIT table and judges whether the current bandwidth is available for the grant; if so, it authorizes the bandwidth and records the granting information in the GIT; otherwise, for step 404. Through this, for the MPCP, EF and MF message services, the availability of bandwidth resource is judged in accordance with the provisioning bandwidth of the port, in particular with the allowed amount of data (for the EF and MF service) or the constant amount of MPCP data (for the MPCP service) in a virtual frame period, as well as the residual bandwidth in the current virtual frame period, when the first is greater than the second, the bandwidth resource is available and granting is allowed. For the AF service, only when the following conditions are met, the bandwidth resource can be considered available: there are pri- mary data reported on the port and the amount of pri- mary data reported is less both the amount of residual data permitted and the sum of the amount of transmission data of provision of gauge width and deficit quantum. Here, the amount of bandwidth provisioning transmission data can be replaced by the requested transmission quantum in the BIT table through ONUID and Compensation index. For the DF service, only when the following conditions are met, the bandwidth resource can be considered available: there are pri- mary data reported in the port and the residual reported pri- mary data are not less than the maximum of the amount of data reported and the amount of transmission data of bandwidth provision. Here, the amount of provision transmission data can be replaced by the requested transmission quantum in the BIT table through ONUID and Compensation index. If the bandwidth is available, the OLT authorizes the current port to transmit and record the grant information in the GIT. For the EF service, the OLT confirms the Start Tempo and the data transmission sizes in accordance with the Reserved Data Arrival Seal and protective band and, subsequently, records said Start Time and said data transmission sizes in the corresponding table element of the port in the GIT of this service. The identifier of said Reserved Data Arrival Seal is the identifier of the arrival time of predefined data in the next uplink virtual frame period. The reference standard of the identifier is the oscillatory clock in the OLT. For the MPCP message service, the OLT confirms the type of downlink MPCP message in accordance with the uplink message type of the current MPCP port in the RIT. Here, for the automatic detection MPCP message service, the type of downlink message is constant. Then, the OLT records the confirmed MPCP type and the Indicator Awarded in the MPCP port corresponding to the GIT table element. If the service is the automatic detection MPCP message service, the OLT can register only the Indicator Awarded in the GIT. For the MF and DF service, the OLT authorizes a window size granted for the current port in accordance with the information of the data sizes reported in the corresponding table elements of the RIT, as well as the window size granted in the RIT. corresponding GIT table element. For the AF service, the OLT authorizes the size of the granted window in accordance with the information of data sizes reported in the corresponding table elements of the RIT and the existing grant deficit. Meanwhile, the OLT calculates a new grant deficit. Subsequently, the OLT records the granted window size and the new grant deficit in the corresponding GIT table element.

Step 403: The OLT updates the residual bandwidth in the current virtual frame period and the information relevant to the current service port to be granted. Specifically comprising: updating the residual bandwidth in the current virtual frame period; update the arrival time stamp of the current reserved data and the residual bandwidth in the virtual frame period in the EF service; with respect to the automatic detection MPCP, restart the start counter and return directly to step 401; eliminate the Indicator Reported in the corresponding RIT of MPCP, MF, AF and DF. Here, eliminating the Reported Indicator means avoiding repeating the authorization in the next virtual framework period. In a preferred embodiment of the present invention, a "funnel algorithm" is adopted as the administration method of granting in the AF service. The person skilled in the art will understand that the weighting algorithm can also be adopted as a grant administration method in the AF service, without affecting the essence and scope of the present invention. Step 404: The OLT judges if there are ports not granted; if so, go back to step 401; otherwise, the port authorization of this priority level terminates and passes to another port authorization action of the next priority level.

Specifically, for MF and MPCP without automatic detection, the aforementioned determination action consists in that: the OLT judges whether there are unread table elements in the currently viewed Active Serviceable ONU Bitmap record; if so, there are ports not granted; otherwise, the authorization of this type of service is terminated. Meanwhile, for services with a port as a unit, such as EF, AF, and DF, the OLT judges first if there is any unread table element in the currently viewed Active Port Port Bits Map table; if so, continue requesting the next table item; otherwise, it continues to judge whether there are unread table elements in the currently viewed Queued Bit Server Map of Service Asset; if so, there are ports not granted; otherwise, the authorization of this type of service is terminated. Here, for automatic detection MPCP, since there is at most one authorization in a virtual frame period, the OLT ignores step 404 and directly takes the authorization action for the service with the next priority level. Figure 5 shows the structure of the GIT, with each table element corresponding to the granting information of each UN port, and the elements are indexed by ONUID. For the uplink EF service, the table element comprises the granted indicator, the start time of the data transmission ("Start Time") and the size of the data transmission ("Length"). Here, the Granted Indicator indicates whether this service port is granted, the Start Time clarifies to this granted port when to initiate the transmission of data; Length equals the granted window size and indicates how much data is granted to be transmitted. For the uplink MF service, the table element comprises the Granted, Reserved and Length Indicator. Here, Reserved means the undefined part reserved for future use. For the MPCP service, the table element comprises the Indicator Awarded, Type and Reserved. Here, Type comprises "Discovery GATE", "Normal GATE" and "REGISTER &GATE". Here, downlink MPCP with GATE type is used especially to transmit the bestowal message. For other services, such as DF and AF, the table element includes "Granted Flag", "Deficit Count" and "Length". Here, Deficit Count is used to record the authorization deficit of this port. The granting deficit reflects the degree to which the cumulative allocation bandwidth exceeds the bandwidth reported in the previous uplink service period. In addition, the system makes an equal QoS guarantee of the AF or DF service through the authorization deficit; for the DF service, since the authorization deficit is ignored, the value of the Deficit can be specified in zero. During the authorization stage, as indicated above, the Start Time is only authorized for the EF service. For other services, such as the automatic detection MPCP message service, its log window size is constant for the PON system, with a constant ONU number and a constant maximum propagation space. For MPCP without automatic detection, the granted length is constant, since the length of the MPCP messages is constant, while the granted length is no longer written in the GIT. For the AF and DF services, the granted Length must be written in the GIT, given that the granted Length of the service port is not constant. For convenient handling of the ONU, the Status Information Table (SIT) of the ONU is established in the present invention to display UN status information. Figure 6 illustrates the structure of the Status Information Table. Each ONU corresponds to a table element that contains the ONU status and the elements are indexed by ONUID. In a preferred embodiment of the present invention, two bits are used to display the status of ONU of Invalid Introduction, Registered ONU and Registration ONU. Here Invalid Introduction indicates that there is no ONU assigned to this ONUID. Registered UN means that this UN has completed registration and that service data can be sent. Registered ONU means that the UN is registering, in this status, the current interacted messages that can be confirmed in accordance with the RIT table element of the MPCP port and the present GIT table element of the ONU. The UN status information table is generated in the interactive procedure of the MPCP messages.

The current UN status depends both on the previous status and on the MPCP message interacted between the OLT and the UN in the previous status. For example, in the Invalid Introduction status if the MPCP message from Discovery GATE is received by the ONU and the MPCP message from REGISTER_REQ is sent to the OLT, the ONU goes to the UN registry status. In the Registration status, if a REGISTER-ACK message is received by the OLT, the ONU goes to the Registered status. After completing the transmission authorization operation, the OLT will generate grant messages in the virtual frame period and send the grant messages to the corresponding ONU at a certain time. The principle of the generation of bestowal messages is: in order to reduce the granting messages of each UN service port and in order to have fewer protective bands and synchronization time sections, the transmission windows of all the Service ports are successive and adjacent, except in the EF service. Thus, except in the EF service, the start time of data transmission granted from all service ports is confirmed and written in the grant message during the generation period of granting messages. This is because a low delay and low delay jitter are required in the ED service, while the timing transmission is always adopted in terms of these ports. Therefore, for the EF service, the granted start time of transmission of data and sizes granted of data transmission, must be confirmed already during the period of creation of granting messages. Meanwhile, for other service ports, only the sizes granted for data transmission are confirmed during the period of creation of granting messages and the granted start time of the data transmission is confirmed during the generation period of granting messages. . The purpose is to make the granted windows of the ports of the same ONU successive and adjacent in the time field, thus reducing the number of protection bands and increasing the bandwidth usage ratio. Figure 7 shows a flow of method of generating bestowal messages based on a preferred embodiment of the present invention. A grant message is generated with ONU as a unit in the system, so that the grant message is transmitted in the downlink MPCP message with GATE type.

In addition, service ports that belong to the same ONU can be granted universally. As shown in Figure 7, in step 701, the UN status information table elements are read one by one in accordance with the ONUID index. Step 702: determine if the current ONU is invalid; in particular, determine whether the corresponding SIT table element of the ONUID of the current ONU is 00; if so, there is no granting and the next SIT table element continues reading, going to step 701; otherwise, in particular the corresponding SIT table element of the ONUID of the current ONU is 01 or 10, continuing with step 703. Step 703: reading the corresponding GIT table elements of each service port to be granted of the current UN and generate grant information, schedule the start time granted for data transmission for each port of service to be granted. This procedure specifically includes: requesting the corresponding table elements of the service ports belonging to the same ONU in the GIT through the ONUID index, acquiring the Granting Status of the service ports in accordance with the Grant Indicator of the element of corresponding table, determine if the current service port is granted, if the service port is granted, read the relevant information, such as the data transmission size of the corresponding table element and program the start time granted for transmission of data for the service port. In this way, for the EF service, requesting the corresponding elements of the GIT, acquire the data transmission size and start time of the data transmission, take the start time as the start time granted of data transmission, write it directly in the downlink MPCP message with GATE type.

For the uplink MPCP service, the MPCP message type is acquired by requesting the GIT. In addition, since the granted size of the data transmission is constant, only the data transmission granting start time of the corresponding service port must be programmed in accordance with the granted data transmission time stamp that is inherited from the period of granting of data transmission. For other services, such as the MF, AF and DF services, the granted size of data transmission is acquired by requesting the GIT and programming the granted start time of data transmission from the corresponding service port in accordance with the transmission time stamp. granted data that is inherited from the transmission grant period. Step 704: grant the messages that are being generated and released in accordance with the granting information and the granted start time of the data transmission. Here, the granting information and the start time granted for the transmission of data from the service ports, belonging to the current ONU, are transmitted by the MPCP message with GATE type. First, the type of downlink MPCP message to the current ONU is confirmed in accordance with the MPCP message types granted, which includes: determining if there is a GCP MPCP message, if there is an MPCP message granted GATE type; subsequently, the granted MPCP message is taken as the downlink MPCP message, the grant information and the granted start time of the data transmission of each granted port confirmed in step 703, are written in the MPCP message downlink with GATE type and then sent. In general, the granting information and the granted start time of four-port data transmission can be transmitted in a downlink MPCP message with GATE type. Therefore, if there are more than 4 ports requesting a grant, two or more downlink MPCP messages with GATE type are generated to transmit the grant information and the granted start time of data transmission of all ports. Step 705: Establish the Grant Indicator of the GIT table element of the current granted port as negative, in order to avoid generating a granting message repeatedly. Step 706: determine if all the elements of the SIT are read; If so, it can be concluded that all messages of granting UN are created and the procedure of generating the message of bestowal must be terminated; otherwise, return to step 701 and continue reading the SIT. Finally, the present invention provides a method for processing an out-of-line ONU or a defective ONU, in particular a method of information aging procedure of ONU. The UN Active Expiration Time Count table is established and each non-response time is counted. If the value exceeds the specified off-line threshold, this ONU will be considered invalid and the invalid ONU resource will be released. Figure 8 shows the structure of the table of Count of Active UN Expiration Time. Each ONU corresponds to a table element and the elements are indexed by ONUID. The table element includes Reported Indicator and Active Expiration Time Count (ATC). Figure 9 illustrates the flow of the ONU information aging method for an implemented case of the present case. With respect to Figure 9, in step 901, the elements of the SIT are read one by one in accordance with the ONUID index. Step 902: Determine if the ONU status information in the current element is invalid; if so, return to step 901 without processing this invalid ONU and continue reading the next element of the SIT; otherwise, go to step 903. Step 903: read the Active Expiration Time Count table item corresponding to the current ONU and go to step 904. Step 904: Determine if any MPCP message is reported by this ONU in the virtual frame period present in accordance with a reported indicator of the present element, in particular to determine whether there is an indicator reported in the table element; if so, restart the Active Expiration Time Count of the present element; Otherwise, continue with the count. In addition, whatever the result of the determination, the Reported Indicator must be eliminated. Step 905: Determine if the value of the Active Expiration Time Count exceeds the established off-line threshold; if so, it can be concluded that this UN is offline and go to step 906; otherwise, it can be concluded that this ONU is online and go to step 907. Step 906: establish the SIT element of the ONU present as invalid, release the relevant resources of this ONU and proceed to step 907. In this way, for the ONU offline, the corresponding element of the SIT is established as invalid to indicate the lack of existence of this UN Thus, the relevant ONU resources released include: the corresponding bits of the Active Bits Map of Services, such as IMPAB, IMAFAB, IEFAB, IAFAB and IDFAB; the corresponding elements of the Table of Active Bits of Service Map, such as IEFAT, IAFAT and IDFAT, the corresponding elements of the Bandwidth Information Table, as well as the ONUID originally assigned to this ONU. Step 907: determine if all elements of the SIT are read; if so, it can be concluded that all the ONUs have been requested, thus terminating the aging procedure of the UN information in the present virtual frame period; otherwise, go back to step 901 and continue reading the SIT. The aforementioned aging process can be carried out repeatedly. In a preferred embodiment of the present invention, the aging procedure the ONU information of the SIT may be initiated by a certain event. Once the aging procedure is started, all the table elements are traversed to update all the information of the ONUs. The procedure ends when all the elements are traversed, waiting for the next event to initiate this procedure. The present invention is applicable in the master-slave system and access to the ONUs must be controlled by the OLT in this system. All access and upload of uplink service data of all ONUs, such as MPCP messages or other MAC messages, is controlled by the OLT. For the downlink situation, the diffusion mode is adopted, the data being filtered and received by the ONU. The administration of the downlink transmission may be based on the priority, in particular the downlink administration sending performed in the sequence of EF, MPCP, CPU / OAM, AF, DF. The present invention is illustrated with graphs and a description with reference to some excellent implemented cases of the present invention. In spite of this, the person skilled in the art will understand that various types of changes in form and detail could be made to the present invention.

These changes are within the spirit and scope of the appended claims.

Claims (16)

NOVELTY OF THE INVENTION CLAIMS

1. - A method for the allocation of dynamic bandwidth in a Passive Optical Network (PON), said PON includes an OLT and a plurality of ONUs that have access to the OLT, comprising: a) classifying the services that must be communicated between the OLT and an ONU in a plurality of types of services in accordance with different transmission requirements, as well as granting a different priority to each type of service; b) authorizing a service port of each type of service to transmit service data in descending sequence of said service priorities, as well as recording the granting information of the service ports obtained from the authorization; c) read said granting information of each port of service to be granted from the same UN; and d) programming the start time of the data transmission granted from each service port to be granted from the current ONU, generating a downlink granting message that includes both the granting information and said data transmission start time granted from each port granted from said current ONU, as well as transmitting said downlink granting messages to said current ONU.

2. The method according to claim 1, further characterized in that it comprises: generating an Active Expiration Time Counter to count the non-response periods of each ONU; and which also includes an aging flow of UN information: A. Request the status of the ONUs one by one, determine if the current ONU status is invalid; if so, return to step A; otherwise, proceed to step B; B. Determine if the MPCP messages have been reported by said current ONU in the current bandwidth allocation polling period: if so, reset the corresponding Active Expiration Time Counter of said current ONU and proceed to step C; otherwise, proceed to step C directly; Or determine if the value of said Active Expiration Time Counter of said current ONU exceeds the established off-line threshold; if so, establish the status of the current UN as invalid, release the corresponding resources from this UN, proceed to step D; otherwise, proceed to step D directly; and D. Determine if all ONUs are requested, terminate the aging information flow of the ONU of the current bandwidth allocation polling period; otherwise, go back to step A, continue with the request for a next UN.

3. The method according to claim 2, further characterized in that it comprises: generating a table of UN Status Information indexed by ONUID, which must store every UN status information that is generated in accordance with the message communication of the ONU. MPCP between the ONUs and the OLT; generate an UN Active Expiration Time Count indexed by ONUID, which includes said Active Expiration Time Counter and reported indicators to indicate whether the MPCP messages have been reported by the corresponding ONUs; said step of determination in step A comprising: in accordance with the ONUID index, reading the status information of ONU from the table of UN Status Information one by one, determining whether the current ONU is invalid in accordance with said UN status information; said step of determination in step B comprising: reading the table item of said current ONU from the table of Active Expiration Time Count, determining if there is a reported indicator in the corresponding table element of said current ONU; if so, it can be concluded that MPCP messages have been reported in the current bandwidth allocation polling period; otherwise, it can be concluded that no MPCP messages have been reported; between step C and step D, also comprising: removing the indicator reported from said current ONU in the UN Active Expiration Time Count table; said step of determination in step D comprising: determining whether all the table elements of the UN Status Information table are read; if so, it can be concluded that all the UNs have been requested; otherwise, it can be concluded that some of the UNs have not been requested.

4. The method according to claim 1, further characterized in that it comprises: generating a table of vMAC Grant information indexed by ONUID, which includes grant information of each of the service ports of the ONUs, as well as Indicators granted to indicate if the corresponding service ports are authorized; the step of registering grant information in step b) comprising: registering said grant information in the vMAC Grant information table, establishing the granted indicator of said service ports granted as authorized; the step of reading the Granting information in step c) comprising: searching the table elements corresponding to the service ports of the same ONU in the vMAC Grant information table in accordance with the ONUID index, searching the ports of service granted in accordance with the granted indicator, read the granting information of the service ports granted; and, after step d), also comprising: establishing the indicator granted as negative of the service ports that have read the granting information.

5. The method according to claim 1, further characterized by comprising: generating the table of UN Status Information indexed by ONUID, which must store the status information of each ONU that is generated in accordance with the communication of messages of MPCP between the ONUs and the OLT; before step c), further comprising: reading the status information of the ONUs one by one from the UN Status Information table in accordance with the ONUID index, determining whether the current ONU status is invalid in accordance with the current UN status information; if so, go back and read the status information of the next ONU in the UN Status Information table; otherwise, proceed step c).

6. The method according to claim 1, further characterized in that said services are classified according to their priority in descending sequence in step a) as fast sending service, automatic detection MPCP message service, MPCP message service No automatic detection, MF service, Insured Shipping service and Best Effort Delivery service.

7. The method according to claim 6, further characterized in that, in relation to the services except for the automatic detection MPCP message service, said authorization step in step b) comprises: b11) confirming the service port current to be granted in accordance with the uplink service activation status; b12) in accordance with the residual bandwidth resource in the current bandwidth allocation polling period, determine whether the current bandwidth resource is available for the constant amount of data from that current service port to be granted for the automatic detection MPCP message service, or for the reporting information of said current service port to be granted for the other types of services; if so, proceed to step b13); otherwise, proceed to step b15); b13) authorize said current service port to be granted to transmit service data, as well as record the current granting information; b14) update the residual bandwidth in the current bandwidth allocation polling period and the relevant information of said current service port to be granted; and determine if there are service ports not granted from the current priority service; if so, return to step b11); otherwise, authorize the ports of the next priority service.

8. The method according to claim 7, further characterized in that it comprises: generating a Bitmap record of a Service Active ONU for each type of service, to store active information indicating whether the service type is activated in the ONUs; generate a Table of Bits of Active Port of Service with port as a unit, which is indexed by ONUID, to store the active information that indicates if this type of service is activated in the ports of access service of the ONUs; generate a vMAC Report Information table indexed by ONUID, to store the reporting information of the ONUs access ports; step b11) comprising: consulting the Register of Bit Bits of the Active Asset of Service and the Bits Map table of the Active Port of Service of each type of service in descending sequence of priorities, discovering a service port with positive active information and confirm the service port as said current service port to be granted; before step b12) further comprising: discovering the report information of the current service port to be granted from the vMAC Report Information table; and, in step b15), determine if there are unread table elements in the current Active Active ONU Bitmap record and in the Current Service Active Port Bitmap table; if so, return to step b11); otherwise, request the corresponding record of the Bitmap of the Active Asset of Service and the table of the Bitmap of the Active Port of Service of the service with the following priority.

9. The method according to claim 7 or 8, further characterized in that, for the fast sending service, said granting information comprises a start time of data transmission and data transmission sizes; said report information comprising reported sizes of data waiting to be transmitted; for the fast sending service, said step of scheduling a start time granted in step d) comprising: taking said start time of data transmission of said granting information as the granted start time of data transmission; for the MPCP message service without automatic detection, said grant information comprising reserved field and type of downlink MPCP messages, wherein said type of MPCP messages comprises Discovery GATE, Normal GATE and REGISTER &GATE; said report information comprising reserved field and MPCP message type field; for the MF service, said granting information comprising reserved field and data transmission sizes; said report information comprising reported sizes of data waiting to be transmitted; for the Insured Shipment service, said granting information comprising the amount of granting deficit and the data transmission sizes of the corresponding service ports; said report information comprising the reported sizes of the data waiting to be transmitted; for the Best Effort Delivery Service, said granting information comprising the data transmission sizes; said report information comprising the sizes of the reported data waiting to be transmitted.

10. The method according to claim 7, further characterized in that it comprises: generating a bandwidth information table for those service ports that request bandwidth control, to store transmission quantum in each polling period. bandwidth allocation; the determination step for those service ports requesting bandwidth control in step b12) comprising: discovering the transmission quantum of the current service port from the Bandwidth Information table, determine if the current bandwidth resource is available in accordance with the reporting information of said current service port to be granted, the transmission quantum of said current service port to be granted and said resource bandwidth residual in the current bandwidth allocation polling period.

11. The method according to claim 6, further characterized in that it comprises generating a start counter for the automatic detection MPCP message service.; the step of authorizing service ports in step b) comprising: b21) triggering said start counter, when said counter reaches the configured threshold, initiating the detection procedure of the MPCP automatic detection messages; if an automatic detection MPCP message is detected, proceed to step b22); otherwise, restart the start counter and return to step b21); b22) determine if the current bandwidth resource is available for the constant amount of data from that current service port to be generated, in accordance with the residual bandwidth resource in the bandwidth allocation period current band of the automatic detection MPCP message service; if so, go to step b23); otherwise the ports of the next priority service are authorized; b23) restart the start counter and return to step b21).

12. The method according to claim 11, further characterized in that, for the automatic detection MPCP message service, said grant information comprises reserved field and downlink MPCP message type, wherein said type of message MPCP includes Discovery GATE, Normal GATE, REGISTER &GATE.

13. The method according to claim 1, further characterized in that, after step d), it further comprises: determining whether there is still any ONU with an undelivered port; if so, return to step c); otherwise, finish.

14. The method according to claim 1, further characterized in that the type of said downlink granting messages of the method is a GATE downlink MPCP message.

15. The method according to claim 7, 8 or 10 further characterized in that said uplink grant information of the present method is carried by REPORT messages.

16. The method according to any of claims 2, 7 and 10, further characterized in that said bandwidth allocation polling period of the present method is the virtual frame period.