Bandwidth allocation

Abstract

An optical line terminal (OLT) is disclosed. The OLT includes at least one processor and at least one memory including machine-readable instructions. The at least one memory and the machine-readable instructions are configured, together with the at least one processor, to cause the OLT to determine, based on one or more variables, a relationship between the bandwidth efficiency and latency of communication between the contents of a queue buffer of an optical network unit and the OLT via an optical distribution network, and to determine burst scheduling for the queue buffer based on the determined relationship.

Description

Technical Field

[0001] This disclosure relates to bandwidth allocation in communication networks. Background Technology

[0002] Passive optical networks (PONs) are used to deliver broadband access services. A PON system has a point-to-multipoint topology, where an optical line terminal (OLT) on the network side is used to connect to multiple user modems (called optical network units (ONUs) on the user side via an optical distribution network (ODN). The ODN includes passive optical power dividers. Because the communication links in the ODN are shared, the OLT schedules the ONUs to transmit upstream via time division multiplexing (TDM), where transmission time slots (also known as bursts) are allocated to service bearer entities (TCONTs) within the ONU. Summary of the Invention

[0003] A first aspect of the present invention provides an optical line terminal, the optical line terminal comprising: at least one processor; and at least one memory including machine-readable instructions; wherein the at least one memory and the machine-readable instructions are configured, together with the at least one processor, such that the optical line terminal: determines, based on one or more variables, a relationship between the bandwidth efficiency and latency of communication between the contents of a queue buffer of an optical network unit and the optical line terminal via an optical distribution network, and determines burst scheduling of the queue buffer based on the determined relationship.

[0004] In one implementation, at least one memory and machine-readable instructions are configured, together with at least one processor, to cause the optical line terminal to determine burst scheduling for a queue buffer by: determining the contents of the queue buffer of the optical network unit and the bandwidth of communication between the optical line terminal; determining burst parameter allocation for the contents of the queue buffer of the optical network unit and communication between the optical line terminal based on the determined bandwidth; and determining burst scheduling based on the determined burst parameter allocation.

[0005] In one implementation, at least one memory and machine-readable instructions are configured, together with at least one processor, to cause the optical line terminal to determine bandwidth based on a determined relationship.

[0006] In one implementation, at least one memory and machine-readable instructions are configured, together with at least one processor, to cause the optical line terminal to determine burst scheduling for a queue buffer by: determining the contents of the queue buffer of the optical network unit and the burst parameter allocation of the communication of the optical line terminal based on the determined relationship, and determining the burst scheduling based on the determined burst parameter allocation.

[0007] In one implementation, at least one memory and machine-readable instructions are configured, together with at least one processor, to cause the optical line terminal to determine bandwidth by selecting a function from a plurality of functions stored in at least one memory or modifying parameters of a function stored in at least one memory based on a determined relationship, and determining bandwidth based on the selected or modified function.

[0008] In one implementation, at least one memory and machine-readable instructions are configured, together with at least one processor, to cause the optical line terminal to determine burst parameter allocation by: selecting a function from a plurality of functions stored in at least one memory or modifying the parameters of a function stored in at least one memory based on the determined relationship, and determining burst parameter allocation based on the selected or modified function.

[0009] In one implementation, at least one memory and machine-readable instructions are configured, together with at least one processor, to cause the optical line terminal to determine the relationship by retrieving variables characterizing the relationship between bandwidth efficiency and latency of the communication from at least one memory.

[0010] In one implementation, at least one memory and machine-readable instructions are configured, together with at least one processor, to cause the optical line terminal to determine the relationship by receiving, as input to the optical line terminal, a variable defining the contents of a queue buffer for an optical network unit and communication preferences for the optical line terminal, and determining the relationship based on the communication preferences.

[0011] In one implementation, at least one memory and machine-readable instructions are configured, together with at least one processor, to cause the optical line terminal to determine the relationship by: evaluating a variable representing the utilization of the optical distribution network for previous communication between the optical network unit and the optical line terminal, and determining the relationship based on the utilization.

[0012] In one implementation, at least one memory and machine-readable instructions are configured, together with at least one processor, to cause the optical line terminal to determine the relationship by: characterizing variables representing the contents of a queue buffer of an optical network unit and the type of communication of the optical line terminal, and determining the relationship based on the type of communication.

[0013] In one implementation, at least one memory and machine-readable instructions are configured, together with at least one processor, to cause the optical line terminal to determine the relationship by identifying the time of communication between the contents of the queue buffer for the optical network unit and the optical line terminal based on a time representation variable, and determining the relationship based on the identified time.

[0014] In one implementation, the optical network unit includes multiple discrete queue buffers, and at least one memory and machine-readable instructions are configured, together with at least one processor, to cause the optical line terminal to determine a respective burst schedule for each of the multiple queue buffers.

[0015] A second aspect of this disclosure provides a computer-implemented method for determining burst scheduling for a queue buffer of an optical network unit, the method comprising: determining a relationship between bandwidth efficiency and latency of communication between the contents of the queue buffer and an optical line terminal via an optical distribution network based on one or more variables, and determining burst scheduling for the queue buffer based on the determined relationship.

[0016] A third aspect of this disclosure provides a computer program including instructions that, when executed by a computer, cause the computer to perform the methods described above.

[0017] These and other aspects of the invention will become clear from the embodiments described below. Attached Figure Description

[0018] To facilitate a clearer understanding of the invention, embodiments of the invention will now be described by way of example with reference to the accompanying drawings, in which:

[0019] Figure 1 An example of a communication network embodying one aspect of this disclosure is illustrated schematically. The communication network includes an optical line terminal and multiple optical network units.

[0020] Figure 2 A graph illustrating delay as a function of bandwidth estimation for an example communication system is shown schematically.

[0021] Figure 3 The previous reference is shown schematically. Figure 1 The optical line terminal and multiple optical network units are identified;

[0022] Figure 4 The components of the bandwidth allocation module of the optical line terminal are schematically shown;

[0023] Figure 5 The sub-modules of the bandwidth allocation module are illustrated schematically;

[0024] Figure 6 The diagram schematically illustrates the burst parameter allocation pattern for communication performed by an optical network unit;

[0025] Figure 7 A table illustrating the communication parameters used in an optical network unit is shown schematically; and

[0026] Figure 8The process involved in bandwidth allocation performed by the bandwidth allocation module is illustrated. Detailed Implementation

[0027] First refer to Figure 1 The communication network 101, embodying one aspect of this disclosure, has a point-to-multipoint (P2MP) topology and includes an optical line terminal (OLT) 102 on the network side, multiple user modems (ONUs) 1 to ONU M (referred to as optical network units (ONUs)) on the user side, and an optical distribution network (ODN) (generally indicated as 103) for communication between the OLT 102 and the multiple ONUs. The ODN 103 includes a passive optical power divider 104 for splitting and combining communication channels between the OLT and the ONUs. In this example, the communication network 101 is a passive optical network (PON), for example, a 10-Gigabit symmetric PON (XGS-PON) conforming to ITU-T G.9807.1, or a 50-Gigabit PON (G.hsp) conforming to ITU-T G.9804.1 / 2 / 3. In this example, the communication network 101 can be deployed for delivering broadband access services.

[0028] Since the communication link (e.g., fiber optic channel) of ODN 103 is shared by multiple ONUs, the OLT schedules the ONUs to transmit upstream using Time Division Multiplexing (TDM) technology (e.g., via TDM or Time Wavelength Division Multiplexing (TDWM) methods). This is also known as burst-mode upstream transmission. In this way, the bandwidth of the ODN is time-divided, making it possible to allocate bandwidth resources to the ONUs. Such bandwidth allocation can be performed statically, meaning it can occur during the provision of the communication network, or dynamically, i.e., through Dynamic Bandwidth Allocation (DBA), whereby bandwidth can be dynamically allocated by the OLT based on simultaneous monitoring of service conditions during the operation of the communication network. Using DBA, the OLT allocates upstream transmission opportunities or upstream bandwidth allocations by allocating burst transmission time slots to the ONUs at specific times based on dynamically estimated ONU activity indicators and their corresponding configured service protocols (i.e., service layer protocols). In the example, each of ONUs 1 to M has multiple service bearer entities, also referred to as transport containers (T-CONTs). A T-CONT is an ONU object that represents a set of logical connections presented as a single entity for upstream transmission bandwidth allocation by the OLT, i.e., as a queue buffer. Therefore, each service bearer entity or T-CONT represents a queue buffer. In the example, bandwidth allocation can be performed on a per-T-CONT basis. Each ONU can support multiple such T-CONTs, and each T-CONT can represent a queue buffer for a specific type of service (e.g., video, voice, or data) or even a specific combination of these. The queue buffer represented by each T-CONT can, for example, consist of one or more physical buffers.

[0029] In many applications, such as XGS-PON, DBA is the preferred mode for bandwidth allocation because it may be desirable to allow for more efficient bandwidth allocation. For example, using DBA, transmission opportunities / slots can be allocated based on the real-time service needs of the ONUs, thereby maximizing the utilization of available ODN bandwidth resources. Therefore, DBA can advantageously facilitate the deployment of more ONUs for an ODN with a given bandwidth capacity, and / or can allow for the provision of enhanced services to ONUs, such as services that require variable rates and peak bandwidth levels exceeding those that can be statically allocated to all ONUs.

[0030] In this example, communication network 101 is depicted as including three ONUs 1 to M. However, in practice, communication networks embodying this disclosure may include more or fewer ONUs, for example, 512 ONUs.

[0031] Next reference Figure 2In situations where ODN bandwidth resources are relatively scarce compared to the service demands of subscribed ONUs, efficient bandwidth allocation among ONUs (or their T-CONTs) is desirable to maximize bandwidth utilization. In other words, the goal is to minimize the over-allocation of bandwidth to ONUs, thus avoiding unwanted unused "waste" of bandwidth. Therefore, bandwidth efficiency is related to the efficiency of ODN bandwidth allocation to T-CONTs. Conversely, insufficient bandwidth allocation to ONUs can adversely impair communication between ONUs and the OLT, and may especially lead to increased communication (packet scheduling) latency, which can be unacceptable in some applications. Latency refers to the time a data object (such as a packet) must wait in the ONU's buffer queue (T-CONT) before being sent to the OLT in an upstream burst format. Since a stream typically consists of several packets, the metric used can be average latency, or another latency-related statistic such as maximum, minimum, 25th percentile, etc. Therefore, it can be understood that bandwidth efficiency and latency (or average queue buffer fill) are strongly coupled, such that changing bandwidth allocation to improve bandwidth efficiency may affect communication latency, and vice versa.

[0032] Figure 2 The diagram schematically illustrates this Pareto optimality between bandwidth efficiency and latency. For the example communication system, it can be seen that the bandwidth estimate plotted on the X-axis (which is generally inversely proportional to bandwidth efficiency) is negatively correlated with the average packet scheduling latency plotted on the Y-axis. Therefore, it can be seen that communication latency decreases as the bandwidth estimate increases (and thus as bandwidth efficiency decreases). Consider, for example, two operating modes A and B. Operation of the example communication system in mode A results in a low bandwidth estimate (and therefore high bandwidth efficiency) but high latency. Operation of the communication system in mode B, where the bandwidth estimate increases (and therefore bandwidth efficiency decreases), desirously reduces latency.

[0033] Therefore, one goal of bandwidth allocation is to allocate bandwidth efficiently to each ONU, that is, to ensure that the bandwidth allocated to each ONU closely corresponds to the actual service needs of the ONU, and to minimize the over-allocation and under-allocation of bandwidth resources to the ONU.

[0034] Bandwidth allocation can be performed by the OLT based on variables, such as service control parameters called service descriptors supplied to the ONU (or its T-CONT, as described above). Such service descriptors may include bandwidth-related parameters defining, for example, fixed bandwidth, guaranteed bandwidth, and / or maximum bandwidth parameters, and latency-related parameters such as jitter tolerance and / or maximum latency, as noted in, for example, section C.7.1.1 of ITU-T Standard G.9807. In this regard, new (additional) parameters related to the relationship between bandwidth efficiency and latency can be added to the service descriptor. These parameters indicate the priority of bandwidth efficiency versus latency-sensitive operations for active service flows, within the range indicated by other parameters. For example, a low-value parameter configuration indicates a trade-off between relatively high bandwidth efficiency (e.g., smaller bandwidth allocation) resulting in relatively higher latency (e.g., larger average queue fill), while a high-value parameter configuration indicates a trade-off between relatively low bandwidth efficiency (e.g., higher bandwidth allocation) resulting in relatively lower latency (e.g., smaller average queue fill).

[0035] Defects have been found in PONs using such service descriptors for bandwidth allocation because such descriptors can lead to inefficient bandwidth allocation. For example, the provided service descriptor may define a maximum latency time for the ONU (or its T-CONT) that is not achievable in a particular application, for example, due to the nature of the service. This often results in excessive bandwidth allocation, even though the specified latency cannot be met. Typically, service descriptors provide limits. In the example above, the limits may be incorrect, or they may be unmet. However, even if the limits are correct, some flexibility may be gained. For example, the DBA can allocate bandwidth from a guaranteed bandwidth (e.g., 10 Mbps) to a maximum bandwidth (100 Mbps), where the guaranteed bandwidth and maximum bandwidth can be provided as parameters.

[0036] Therefore, it has been determined that in PON (e.g., Figure 1 In some applications, bandwidth allocation efficiency in the XGS-PON type (illustrated in the diagram) can ideally be improved by allocating bandwidth to ONUs (or their T-CONTs) based on control parameters (i.e., service descriptors) that define the desired relationship between bandwidth efficiency and latency in communication. For example, such allocation methods can ideally avoid over-allocating bandwidth to ONUs (or their T-CONTSs), thereby improving bandwidth efficiency while maintaining a given latency. Therefore, aspects of this disclosure relate to bandwidth allocation in P2MP shared-channel communication networks (such as PON) based at least in part on desired bandwidth efficiency and latency characteristics.

[0037] Next reference Figure 3In this example, OLT 102 includes a bandwidth allocation module 301 for allocating bandwidth resources of ODN 103 to multiple ONUs 1 to M. In this example, bandwidth allocation module 301 is a dynamic bandwidth allocation module for dynamically allocating bandwidth to the ONUs.

[0038] Each ONU in an ONU includes one or more T-CONTs. Each T-CONT in the network is identified by a unique allocation identifier (Alloc-ID) known to the OLT. The Alloc-ID is an identifier (e.g., a number) assigned by the OLT to the ONU to identify the T-CONT as the receiver of upstream bandwidth allocation within the corresponding ONU. Therefore, in this example, ONU 1 supports T-CONTs identified by Alloc-IDs A through C, ONU 2 supports T-CONTs identified by Alloc-IDs D through F, and ONU M supports T-CONT Z identified by Alloc-IDs X through F.

[0039] In the example, each T-CONT or Alloc-ID is provided with service control parameters, i.e., service descriptors, which define the service level of the corresponding T-CONT. These service descriptors or service parameters are stored in the memory of the bandwidth allocation module 301. For example, the service descriptor may include a service descriptor that defines the expected relationship between bandwidth efficiency and latency for communication between the corresponding T-CONT and OLT 102.

[0040] Using DBA, the OLT 102 allocates upstream transmission opportunities or upstream bursts to T-CONTs within a subscribed ONU on a per-T-CONT basis. The OLT can utilize dynamic estimations of ONU and T-CONT activity defined by service control parameters (called service descriptors) and their corresponding service protocols (i.e., service layer protocols). Activity status indications can be explicit via buffer status reports, implicit via transmissions of idle XGEM frames during upstream transmission opportunities, or both. After estimation, the OLT performs bandwidth allocation, which is then used for burst parameter assignment (or allocation). Based on these, burst scheduling is calculated, defining the exact timing of bursts for each T-CONT. The OLT authorizes these bursts to ONUs via bandwidth mapping on a per-T-CONT basis, such as... Figure 3 As shown by the dashed lines, this includes control signals sent in the downstream direction. The ONU then controls the transmission of its queue buffer / T-CONT according to the bandwidth grant defined in the bandwidth map.

[0041] Next reference Figure 4In the example, the bandwidth allocation module 301 of the OLT 102 includes a processor 401, a memory 402, an input / output interface 403, and a system bus 404. The bandwidth allocation module 301 is configured to run a computer program for dynamically allocating the bandwidth of the ODN to multiple ONUs, for example, on a per-T-CONT basis, the bandwidth of the ODN is converted into upstream transmission opportunities (bursts).

[0042] Processor 401 is configured to execute instructions for a computer program that dynamically allocates bandwidth of the ODN to multiple ONUs. Memory 302 is configured as a non-volatile storage for the computer program defining machine-readable instructions for execution by processor 401, and serves as a read / write memory for storing operational data associated with the computer program executed by processor 301. Input / output interface 403 is provided for connecting bandwidth allocation module 301 to other components of OLT 102, and for connecting bandwidth allocation module 301 to ODN 103 to facilitate the exchange of information (e.g., upstream status reports and downstream bandwidth mappings) between bandwidth allocation module 301 and ONUs 1 to M. Components 401 to 403 of bandwidth allocation module 301 communicate via system bus 404.

[0043] Next reference Figure 5 In the example, components 401 to 403 of bandwidth allocation module 301 support multiple functional submodules 501 to 506 for dynamically allocating bandwidth to the T-CONT of ONUs 1 to M. In this diagram, the logical connections between submodules 501 to 506 are schematically depicted by arrows, where the direction of the arrows indicates the typical direction of data flow between the connected submodules.

[0044] The bandwidth efficiency and delay determination submodule 501 is configured to determine the expected relationship between bandwidth efficiency and delay for communication between each ONU and OLT based on one or more variables. In the example described in detail herein, the dynamic bandwidth allocation module 301 is configured to allocate bandwidth to the ONU on a per-T-CONT basis, i.e., such that each T-CONT (Alloc-ID A to Alloc-ID Z) is granted an individual bandwidth allocation. In these examples, the bandwidth efficiency and delay determination submodule 501 is configured to determine the corresponding expected relationship between bandwidth efficiency and delay for communication between each T-CONT and OLT.

[0045] The bandwidth efficiency and latency determination submodule 501 receives one or more variables related to the determination as input. In this example, the input variables include a provided service descriptor defining the desired bandwidth efficiency and latency relationship, and also include one or more use case-related triggers. The provided service descriptor can be defined by the operator of OLT 102 and stored in memory 402. For example, the use case-related triggers can be manually entered by the operator of OLT 102 or by a user of a corresponding ONU among ONUs 1 to M, or they can be determined by the bandwidth efficiency and latency determination submodule 501. Examples of use case-related trigger variables that can be input to the bandwidth efficiency and latency determination submodule 501 include:

[0046] (1) End-user requirements - The requirements of users from the corresponding ONU can be received by the bandwidth efficiency and latency determination submodule 501, for example, for communication with particularly low latency, or for bandwidth efficiency with particularly high efficiency.

[0047] (2) PON-level utilization – The utilization rate of ODN bandwidth resources can be identified by the bandwidth efficiency and delay determination submodule 501, for example, by receiving utilization status reports from the corresponding ONU 1 to M, OLT 102, or ODN 103. For example, if the PON-level utilization rate is identified as relatively low, resulting in unused bandwidth resources in principle, the relationship between minimizing delay at the expense of reduced bandwidth efficiency can be determined.

[0048] (3) T-CONT behavior - Based on the analysis of the T-CONT service trajectory by the bandwidth efficiency and delay determination submodule 501, the service type output by the corresponding T-CONT can be classified / identified. Based on this, the bandwidth efficiency and delay determination submodule 501 can determine the expected bandwidth efficiency and delay characteristics for which service categories of communication.

[0049] (4) Carrier Queue Configuration – Carriers can configure certain parameters / functions (e.g., VLAN priority, IP CoS, etc.) for T-CONTS of different applications. These configurations can be stored in the memory of the bandwidth allocation module 301; and

[0050] (5) Time / Day – Bandwidth Efficiency and Latency Determination Submodule 501 can determine the time / day of the upcoming time interval, for example, by referencing the clock of OLT 102. This parameter can be related to the determination of the bandwidth efficiency and latency relationship, as time / day can be considered as a proxy indication of the expected service type. For example, during working hours, the expected service mainly consists of document transmissions, and relatively high latency is acceptable. Conversely, outside of working hours, the expected service may include a higher proportion of video game data, for which relatively low latency transmission may be required.

[0051] (6) IP / TCP Header Information Inspection – Based on the analysis of the T-CONT service trace payload data by the bandwidth efficiency and delay determination submodule 501, the IP or TCP header is exported and its fields (such as Explicit Congestion Notification (ECN) related flags) are inspected. Based on this, the bandwidth efficiency and delay determination submodule can determine the expected bandwidth efficiency and delay characteristics for communication used in this T-CONT service. The ECN flag is related to enabling communication links for L4S (Low Latency, Low Loss, Scalable Throughput).

[0052] Therefore, based on one or more input variables, namely, input service descriptors and / or one or more use case-related triggers, the bandwidth efficiency and delay determination submodule 501 determines the desired relationship between bandwidth efficiency and delay for communication between the corresponding T-CONT and OLT, i.e., the trade-off between bandwidth efficiency and delay. The output of the bandwidth efficiency and delay determination submodule 501 (i.e., the determined relationship between bandwidth efficiency and delay) is provided to the control means determination submodule 502.

[0053] The control means determination submodule 502 is used to determine a method for employing the determined relationship between bandwidth efficiency and latency. In the example, referring to the figure, the control means determination submodule 502 outputs a control signal to one or both of the bandwidth estimation submodule 503 and the burst parameter allocation submodule 505. For example, the control means determination submodule 502 can select an algorithm from a corresponding group of algorithms or algorithm configurations for output to the bandwidth estimation submodule 503 and / or the burst parameter allocation submodule 505, thereby influencing the operation of one or both of the bandwidth estimation submodule 503 and / or the burst parameter allocation submodule 505 based on the determined relationship between bandwidth efficiency and latency. For example, as described further below, when the determined relationship between bandwidth efficiency and latency indicates a need for low-latency and low-bandwidth-efficiency communication, the control means determination submodule 502 can select an algorithm or algorithm configuration to output to the bandwidth estimation submodule 503 and the burst parameter allocation submodule 505, which results in bandwidth estimation and burst parameter allocation that favors the desired low-latency and low-bandwidth-efficiency communication.

[0054] The bandwidth estimation submodule 503 is configured to estimate bandwidth requirements for an upcoming time interval (DBA cycle), for example, on a per-T-CONT basis. The bandwidth estimation submodule 503 operates on the fact that completely accurate information about the actual requirements for each T-CONT is not actually available to the bandwidth allocation module 301. Instead, the bandwidth allocation submodule 503 estimates the desired bandwidth requirements using a bandwidth estimation algorithm based on input dynamic service information and relying on control mechanisms to determine the output of submodule 502. The input dynamic service information may include dynamic status reports output by T-CONTS and / or service information metrics monitored by the bandwidth estimation submodule 503. The bandwidth estimation submodule 503 may, for example, access a set of two or more bandwidth estimation algorithms stored in memory 402, wherein a first algorithm tends to produce a bandwidth estimate that may result in relatively low-latency communication, e.g., where the resulting bandwidth estimate may be relatively high for a given dynamic service information, while a second algorithm tends to produce a bandwidth estimate that may result in high latency, e.g., where the resulting bandwidth estimate may be relatively low for a given dynamic service information. In other words, the bandwidth estimation submodule 503 is configured to estimate the upcoming bandwidth demand, for example, on a per-T-CONT basis, thereby estimating the impact of the output of the control means determination module 502, and thus the impact of the bandwidth efficiency and delay relationship determined by the bandwidth efficiency and delay determination submodule 501.

[0055] The bandwidth allocation submodule 504 is configured to grant bandwidth allocation to the ONU based on the bandwidth estimate generated by the bandwidth estimation submodule 503 and the input service descriptor (control parameters), for example, on a per-T-CONT basis. For example, the input service descriptor may be provided by the operator of the OLT 102 and stored in memory 402, and service control parameters such as the maximum bandwidth allocation and / or maximum delay for communication between the ONU and the OLT may be defined, for example, on a per-T-CONT basis. The output of the bandwidth allocation module 504 is therefore a bandwidth allocation that will be converted into time slot / burst opportunities for upcoming time intervals for communication between the ONU and the OLT 102, for example, on a per-T-CONT basis.

[0056] The burst parameter allocation submodule 505 is configured to determine the output of submodule 502 based on the bandwidth allocation, input service descriptors, and control mechanisms determined by the bandwidth allocation submodule 504. For example, it allocates parameters for (burst mode) communication between the ONU and OLT102 on a per T-CONT basis. In short, the burst parameter allocation submodule 505 is configured to define the desired number (frequency) and size of bursts, or other similar parameters, to support the bandwidth allocation generated by the bandwidth allocation submodule 504.

[0057] The burst parameter allocation submodule 505 is adopted based on the understanding that the size and frequency of bursts can affect the bandwidth efficiency and latency of communication between the ONU and the OLT. For example, it can be understood that several short bursts may advantageously lead to a reduction in communication latency compared to fewer longer bursts, but conversely, may undesirably lead to a reduction in bandwidth efficiency, the reasons for which will be referred to... Figure 6 Further detailed description. The burst parameter allocation submodule 505 is configured to allocate burst parameters based on the output of the control means determination submodule 502. The burst parameter allocation submodule 505 may, for example, access a set of two or more burst parameter assignment / allocation algorithms stored in memory 402, wherein a first algorithm tends to allocate burst parameters that may result in relatively low-latency communication, for example, where, given dynamic service information and bandwidth allocation, the resulting burst allocation consists of multiple relatively short bursts and results in low latency, while a second algorithm tends to allocate burst parameters that may result in high bandwidth efficiency, for example, where, given dynamic service information and bandwidth allocation, the resulting burst allocation has fewer longer bursts and results in higher latency. In other words, the burst allocation submodule 505 is configured, for example, to allocate burst parameters on a per-T-CONT basis to support bandwidth allocation, whereby the allocation is influenced by the output of the control means determination submodule 502, and therefore by the bandwidth efficiency versus latency relationship determined by the bandwidth efficiency versus latency determination submodule 501. The output of the burst parameter allocation submodule 505 is provided to the bandwidth mapping generation submodule 506.

[0058] The bandwidth mapping generation submodule 506 is configured to generate bandwidth maps defining TDM output scheduling (i.e., burst scheduling) for ONUs on a per-T-CONT basis. The generated bandwidth maps define the timing and size of upstream transmission opportunities (bursts) for each ONU on a per-T-CONT basis, i.e., providing size and timing specifications for each Alloc-ID. The bandwidth mapping generation submodule 506 is also configured to transmit the generated bandwidth maps to the ONUs via the ODN. An example of the bandwidth maps generated by the bandwidth mapping generation submodule 506 is shown in... Figure 7 Described in the text.

[0059] Next reference Figure 6 As previously referenced Figure 5As mentioned above, burst parameter assignment / allocation can affect the bandwidth efficiency and latency of communication between the ONU and OLT. This is at least partly because each burst of the data payload incurs overhead to ensure proper reception of the burst at the OLT, such as preambles for training the equalizer and delimiters for identifying the end of overhead. Therefore, it is understandable that allocating many small bursts to satisfy bandwidth allocation can reduce latency because smaller bursts are more easily scheduled into busy transmission schedules, but it can also reduce bandwidth efficiency due to the additional overhead of each burst.

[0060] Next reference Figure 7 The bandwidth mapping generated by the bandwidth mapping generation submodule 506 defines TDM output scheduling, i.e., burst scheduling, for each ONU based on each T-CONT. The generated bandwidth mapping defines the size and timing of upstream transmission opportunities for each ONU based on each T-CONT, i.e., it can provide size and timing specifications for each Alloc-ID. Figure 6 The example table depicted includes mapping data for Alloc-IDs A through C, i.e., the T-CONT for ONU 1. In other applications, the generated bandwidth mapping may include mapping data for the T-CONTs of all ONUs 1 through M.

[0061] Final Reference Figure 8 In the example, the computer program used to determine the burst scheduling of the T-CONT (i.e., queue buffer) of the ONU stored in memory 302 includes two phases.

[0062] In stage 801, the computer program causes processor 401 to determine the corresponding relationship between bandwidth efficiency and latency for communication from the contents of each queue buffer of the ONU (i.e., T-CONTS) to the OLT. (See previous reference...) Figure 4 In the example, phase 801 may be performed by the bandwidth efficiency and latency determination submodule 501 based on one or both of the bandwidth efficiency and latency descriptors (e.g., the descriptor for each T-CONT) and / or (multiple) use case-related triggers.

[0063] In stage 802, the computer program causes the processor 301 to determine the burst scheduling of the queue buffer for communication of T-CONTS content to the OLT, i.e., T-CONTS, based on the relationship between bandwidth efficiency and latency determined in stage 801. (See previous reference...) Figure 5In the example, stage 802 can be executed by the control means determination submodule 502, bandwidth estimation submodule 503, bandwidth allocation submodule 504, and burst parameter allocation submodule 505. In the example, stage 802 may also include generating a mapping of the burst schedule determined by the record, for example, by converting burst parameters into specific burst sizes and timings, and transmitting the mapping to the ONU; this can be executed by the bandwidth generation submodule 506.

[0064] Although the invention and its advantages have been described in detail, it should be understood that various changes, substitutions, and modifications may be made herein without departing from the spirit and scope of the invention as defined by the appended claims. In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite articles "a" or "an" do not exclude a plurality.

Claims

1. An optical line terminal, comprising: At least one processor; as well as At least one memory including machine-readable instructions; The at least one memory and the machine-readable instructions are configured, together with the at least one processor, to cause the optical line termination to: Based on one or more variables, determine the relationship between bandwidth efficiency and latency of communication between the contents of the queue buffer of an optical network unit (ONU) and the optical line terminal (OLT) via an optical distribution network, wherein the one or more variables include at least one of the following: (1) a variable defining communication preferences for communication between the ONU and the OLT, and the contents of the queue buffer; (2) a variable representing the utilization of the optical distribution network for previous communication between the ONU and the OLT; (3) a variable representing the type of communication between the ONU and the OLT, and the contents of the queue buffer; (4) a variable representing the timing of communication between the ONU and the OLT, and the contents of the queue buffer; or (5) a variable representing IP or TCP header fields for communication between the ONU and the OLT, and the contents of the queue buffer. Based on the established relationship, burst scheduling for the queue buffer is determined.

2. The optical line terminal of claim 1, wherein the at least one memory and the machine-readable instructions are configured, together with the at least one processor, such that the optical line terminal determines the burst scheduling for the queue buffer in such a way as follows: Determine the contents of the queue buffer of the optical network unit and the communication bandwidth of the optical line terminal. Based on the determined bandwidth, the contents of the queue buffer of the optical network unit and the burst parameter allocation for communication of the optical line terminal are determined, and The burst scheduling is determined based on the allocated burst parameters.

3. The optical line terminal of claim 2, wherein the at least one memory and the machine-readable instructions are configured, together with the at least one processor, to cause the optical line terminal to determine the bandwidth based on a determined relationship.

4. The optical line terminal of claim 1, wherein the at least one memory and the machine-readable instructions are configured, together with the at least one processor, such that the optical line terminal determines the burst scheduling of the queue buffer in the following manner: Based on the established relationship, the contents of the queue buffer of the optical network unit and the burst parameter allocation of the communication of the optical line terminal are determined, and The burst scheduling is determined based on the allocated burst parameters.

5. The optical line terminal of claim 2, wherein the at least one memory and the machine-readable instructions are configured, together with the at least one processor, such that the optical line terminal determines the bandwidth in the following manner: Based on the determined relationship, select a function from a plurality of functions stored in the at least one memory, or modify the parameters of a function stored in the at least one memory, and The bandwidth is determined based on the selected or modified function.

6. The optical line terminal of claim 3, wherein the at least one memory and the machine-readable instructions are configured, together with the at least one processor, such that the optical line terminal determines the burst parameter allocation in the following manner: Based on the determined relationship, select a function from a plurality of functions stored in the at least one memory, or modify the parameters of a function stored in the at least one memory, and The burst parameter allocation is determined based on the selected or modified function.

7. The optical line terminal of claim 4, wherein the at least one memory and the machine-readable instructions are configured, together with the at least one processor, such that the optical line terminal determines the burst parameter allocation in the following manner: Based on the determined relationship, select a function from a plurality of functions stored in the at least one memory, or modify the parameters of a function stored in the at least one memory, and The burst parameter allocation is determined based on the selected or modified function.

8. The optical line terminal of claim 1, wherein the at least one memory and the machine-readable instructions are configured, together with the at least one processor, such that the optical line terminal determines the relationship in the following manner: Retrieve from the at least one memory a variable characterizing the relationship between bandwidth efficiency and latency for the communication.

9. The optical line terminal of claim 1, wherein the at least one memory and the machine-readable instructions are configured, together with the at least one processor, such that the optical line terminal determines the relationship in the following manner: The variable defining the communication preference is received as input to the optical line terminal, the communication preference being for communication between the optical network unit and the optical line terminal, and the contents of the queue buffer, and the relationship is determined based on the communication preference.

10. The optical line terminal of claim 1, wherein the at least one memory and the machine-readable instructions are configured, together with the at least one processor, such that the optical line terminal determines the relationship in the following manner: The evaluation represents the variable indicating the utilization of the optical distribution network relative to previous communication between the optical network unit and the optical line terminal, and the relationship is determined based on the utilization.

11. The optical line terminal of claim 1, wherein the at least one memory and the machine-readable instructions are configured, together with the at least one processor, such that the optical line terminal determines the relationship in the following manner: The variable characterizes the type of communication between the optical network unit and the optical line terminal, and the relationship is determined based on the type of communication.

12. The optical line terminal of claim 1, wherein the at least one memory and the machine-readable instructions are configured, together with the at least one processor, such that the optical line terminal determines the relationship in the following manner: Based on a time-representation variable, the time of communication between the optical network unit and the optical line terminal, and the content of the queue buffer, is identified, and the relationship is determined based on the identified time.

13. The optical line terminal of claim 1, wherein the at least one memory and the machine-readable instructions are configured, together with the at least one processor, such that the optical line terminal determines the relationship in the following manner: The variable represents the IP or TCP header related field, which relates to the communication between the optical network unit and the optical line terminal, and the contents of the queue buffer, and the relationship is determined based on the IP or TCP header related field.

14. The optical line terminal according to any one of claims 1 to 13, wherein the optical network unit comprises a plurality of discrete queue buffers, and the at least one memory and the machine-readable instructions are configured, together with the at least one processor, such that the optical line terminal determines a respective burst schedule for each of the plurality of queue buffers.

15. A computer-implemented method for determining burst scheduling for a queue buffer of an optical network element, the method comprising: Based on one or more variables, determine the relationship between bandwidth efficiency and latency of communication between the contents of the queue buffer and the optical line terminal via the optical distribution network, wherein the one or more variables include at least one of the following: (1) a variable defining communication preferences for communication between the optical network unit and the optical line terminal, and the contents of the queue buffer; (2) a variable representing the utilization of the optical distribution network for previous communication between the optical network unit and the optical line terminal; (3) a variable representing the type of communication between the optical network unit and the optical line terminal, and the contents of the queue buffer; (4) a variable representing the timing of communication between the optical network unit and the optical line terminal, and the contents of the queue buffer; or (5) a variable representing IP or TCP header related fields for communication between the optical network unit and the optical line terminal, and the contents of the queue buffer. Based on the established relationship, burst scheduling for the queue buffer is determined.

16. A computer program product comprising instructions that, when executed by a computer, cause the computer to perform the method according to claim 15.

Images