CN103546949A

CN103546949A - An Access Network Bandwidth Scheduling Method Based on Dormant Energy Saving Mechanism

Info

Publication number: CN103546949A
Application number: CN201310504805.5A
Authority: CN
Inventors: 张崇富; 肖能武; 邱昆
Original assignee: University of Electronic Science and Technology of China
Current assignee: University of Electronic Science and Technology of China
Priority date: 2013-10-23
Filing date: 2013-10-23
Publication date: 2014-01-29
Anticipated expiration: 2033-10-23
Also published as: CN103546949B

Abstract

The invention discloses a method for dispatching the access network bandwidth based on a sleep energy-saving mechanism. For the condition that the total service bandwidth of a system cannot completely meet the request for the total bandwidth of three types of services, firstly, the request of the EF service is totally met preferentially; secondly, the remaining total bandwidth is distributed to the AF service and the BE service. For distribution of the ONU bandwidth to the AF service and the BE service, a sleep benefit factor of each ONU is obtained through ONU bandwidth request calculation, and then the sleep benefit values, generated at different remaining bandwidths, of each ONU are obtained through calculation. When ONU bandwidth distribution is conducted, through comparison between the sleep benefit value, generated when the bandwidth is distributed to each ONU, of each ONU and the sleep benefit value, generated when the bandwidth is not distributed to the corresponding ONU, of the corresponding ONU, whether the bandwidth needs to be distributed to the ONU is determined, and therefore the optimal sleep benefit value is obtained. When remaining total bandwidth distribution is conduced, a priority-balance-factor is set so that QoS can be guaranteed, and therefore the comprehensive performance of the system is improved. According to the method for dispatching the access network bandwidth based on the sleep energy-saving mechanism, the energy-saving effect is achieved in a soft control mode, and therefore the hardware cost is not increased.

Description

An Access Network Bandwidth Scheduling Method Based on Dormant Energy Saving Mechanism

技术领域technical field

本发明属于接入网技术领域，更为具体地讲，涉及一种基于休眠节能机制的接入网带宽调度方法。The invention belongs to the field of access network technology, and more specifically relates to a bandwidth scheduling method of an access network based on a dormant energy-saving mechanism.

背景技术Background technique

光纤宽带接入网中的PON（Passive Optical Network，无源光网络）系统结构主要由中心局的光线路终端(OLT:Optical Line Terminal)、包含无源光器件的光分配网(ODN:Optical Distribution Network)、用户端的光网络单元(ONU：OpticalNetwork Unit)组成，通常采用点到多点的树型拓扑结构。在下行方向，IP数据、语音、视频等多种业务由位于中心局的OLT，采用广播方式通过ODN中的1：N无源光分配器分配到PON上的所有ONU单元。在上行方向，来自各个ONU的多种业务信息互不干扰地通过ODN中的1：N无源光合路器耦合到同一根光纤，最终传输到位于局端的OLT接收端，类似于点到点的结构。在高速率、大带宽需求的驱动下，接入网带宽的需求量呈指数增长，这导致采用PON技术的光纤宽带接入网取得了迅猛的发展，但其系统过高的能耗问题也逐渐引起了广泛的关注。目前国内外研究主要集中在物理层节能方式和数据链路层节能方式，具体如下：The PON (Passive Optical Network, Passive Optical Network) system structure in the fiber optic broadband access network is mainly composed of the optical line terminal (OLT: Optical Line Terminal) of the central office, and the optical distribution network (ODN: Optical Distribution Network) containing passive optical devices. Network) and optical network unit (ONU: Optical Network Unit) at the user end, usually using a point-to-multipoint tree topology. In the downlink direction, various services such as IP data, voice, and video are distributed to all ONU units on the PON by the OLT located in the central office through the 1:N passive optical distributor in the ODN in broadcast mode. In the upstream direction, multiple service information from each ONU is coupled to the same optical fiber through the 1:N passive optical combiner in the ODN without interfering with each other, and finally transmitted to the OLT receiving end at the central office, similar to point-to-point structure. Driven by the demand for high speed and large bandwidth, the demand for access network bandwidth is increasing exponentially, which leads to the rapid development of optical fiber broadband access network using PON technology, but the problem of excessive energy consumption of the system is gradually aroused widespread concern. At present, research at home and abroad mainly focuses on the energy-saving methods of the physical layer and the energy-saving methods of the data link layer, as follows:

a）基于物理层的节能方式a) Energy-saving methods based on the physical layer

基于物理层的节能方式无需改变上层协议，通常改进一些物理层的设备或器件结构，如Claudio BIANCO,Flavio CUCCHIETTI等在“Energy consumptiontrends in the Next Generation Access Network-a Telco perspective”一文中通过不断改进集成电路技术达到节能效果；EITETSU IGAWA,MASAMICHI NOGAMI等在“Symmetric10G-EPON ONU Burst-Mode Transceiver Employing DynamicPower Save Control Circuit”一文中采用更好的节能装置，如突发模式的激光器等。The energy-saving method based on the physical layer does not need to change the upper layer protocol, and usually improves some physical layer equipment or device structures, such as Claudio BIANCO, Flavio CUCCHIETTI, etc. in the article "Energy consumption trends in the Next Generation Access Network-a Telco perspective" through continuous improvement of integration Circuit technology achieves energy-saving effect; EITETSU IGAWA, MASAMICHI NOGAMI, etc. adopt better energy-saving devices, such as burst-mode lasers, in the article "Symmetric10G-EPON ONU Burst-Mode Transceiver Employing Dynamic Power Save Control Circuit".

b）基于数据链路层的节能方式b) Energy saving method based on the data link layer

这种节能方式主要是在EPON（Ethernet Passive Optical Network，以太网无源光网络）的媒体接入控制（MAC,Media Access Control）层和GPON(Gigabit-Capable PON，千兆无源光网络)的传输汇聚（TC,TransmissionConvergence）层上设计高效的节能通信协议或算法。其中，ONU休眠方式，即当ONU端的业务流量比较少时使其处于休眠状态，是当前最重要的一种节能方式，在JUN-ICHI KANI的“Power Saving Techniques for Optical Access”及WONG S,VALCARENGHI L的“Sleep Mode for Energy Saving PONs:Advantagesand Drawbacks”等文章中均有体现。This energy-saving method is mainly in the media access control (MAC, Media Access Control) layer of EPON (Ethernet Passive Optical Network, Ethernet Passive Optical Network) and GPON (Gigabit-Capable PON, Gigabit Passive Optical Network) Design efficient energy-saving communication protocols or algorithms on the transmission convergence (TC, Transmission Convergence) layer. Among them, the ONU dormancy method, that is, when the ONU-side business traffic is relatively small, it is in a dormant state, which is currently the most important energy-saving method. It is reflected in the "Sleep Mode for Energy Saving PONs: Advantages and Drawbacks" and other articles.

以上两种方式均能达到一定的节能效果，但都不可避免的带来了一些负面影响，有的在一定程度上牺牲了系统的QoS（Quality of Service，服务质量），有的是通过对物理层设备和器件的改变增加了系统的成本。因此开发出一种既能适应现有物理层设备的要求，又能保证系统的QoS的性能优秀、复杂度低、易于实现的绿色带宽分配算法是极其重要的。Both of the above two methods can achieve a certain energy saving effect, but they inevitably bring some negative impacts, some sacrifice the system QoS (Quality of Service, service quality) to a certain extent, and some pass physical layer equipment and device changes increase the cost of the system. Therefore, it is extremely important to develop a green bandwidth allocation algorithm that can not only adapt to the requirements of existing physical layer equipment, but also guarantee the system's QoS performance, low complexity, and easy implementation.

目前在QoS技术中，为了支持多业务接入的需求,在提高节能效率的同时保证系统的QOS，最常用的是设置统一的优先级分类，并且经常将业务的优先级划分为不同类型，主要将业务的优先级划分为EF（Effort Forward,加速转发型）业务，AF（Assured Forwarding,确保转发型）业务，BE（Best Effort,尽力转发型）业务。EF业务一般都事先由用户和运营商之间进行预约，此类型业务的带宽申请大小是可以预期的，因此在进行系统带宽资源配置保证此类型业务，而剩余带宽不一定足够AF、BE两种业务，在这种情况下如何实现最优节能也是需要进行研究的。At present, in QoS technology, in order to support the requirements of multi-service access and ensure the QOS of the system while improving energy saving efficiency, the most commonly used is to set a unified priority classification, and often divide the priority of services into different types, mainly The business priority is divided into EF (Effort Forward, accelerated forwarding type) business, AF (Assured Forwarding, guaranteed forwarding type) business, BE (Best Effort, best effort forwarding type) business. EF services are generally reserved in advance between the user and the operator. The bandwidth application size for this type of service is predictable. Therefore, the system bandwidth resources are configured to ensure this type of service, and the remaining bandwidth may not be sufficient for both AF and BE. In this case, how to achieve optimal energy saving also needs to be studied.

发明内容Contents of the invention

本发明的目的在于克服现有技术的不足，提供一种基于休眠节能机制的接入网带宽调度方法，通过设置优先级平衡因子及休眠效益因子优化业务的带宽分配，从而在保证接入网系统的QoS的同时，获得良好的节能效果。The purpose of the present invention is to overcome the deficiencies of the prior art, and provide a method for scheduling access network bandwidth based on a dormant energy-saving mechanism. At the same time of QoS, good energy saving effect is obtained.

为实现上述发明目的，本发明基于休眠节能机制的接入网带宽调度方法，包括以下步骤：In order to achieve the above-mentioned purpose of the invention, the access network bandwidth scheduling method based on the dormant energy-saving mechanism of the present invention includes the following steps:

S1：光线路终端OLT接收来自光网络单元ONU对各个业务的带宽请求，记为带宽请求矩阵B：S1: The optical line terminal OLT receives the bandwidth request from the optical network unit ONU for each service, which is recorded as the bandwidth request matrix B:

$B B = = \begin{matrix} [\begin{matrix} {B B}_{EF EF,, 11} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {B B}_{EF EF,, i i} & \cdot &Center Dot; \cdot &Center Dot; \cdot \cdot & {B B}_{EF EF,, N N} \\ {B B}_{AF AF,, 11} & \cdot \cdot \cdot \cdot \cdot &Center Dot; & {B B}_{AF AF,, i i} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {B B}_{AF AF,, N N} \\ {B B}_{BE BE,, 11} & \cdot \cdot \cdot &Center Dot; \cdot &Center Dot; & {B B}_{BE BE,, i i} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {B B}_{BE BE,, N N} \end{matrix}] \end{matrix}$

其中，B_m,i表示第i个ONU对m类业务的带宽请求，m∈{EF,AF,BE}，EF表示该业务类型为加速转发，AF表示该业务类型为确保转发，BE表示该业务类型为尽力转发，i＝1,2,…,N，N表示ONU的数量；系统总业务带宽记为W_t，并且 $Σ_{i = 1}^{N} B_{EF, i} < W_{t} < Σ_{i = 1}^{N} (B_{EF, i} + B_{AF, i} + B_{BE, i});$ Among them, B _m,i represents the bandwidth request of the i-th ONU for m-type services, m∈{EF,AF,BE}, EF indicates that the service type is accelerated forwarding, AF indicates that the service type is guaranteed forwarding, and BE indicates that the service type is The service type is best-effort forwarding, i=1,2,...,N, N represents the number of ONUs; the total service bandwidth of the system is denoted as W _t , and $Σ_{i = 1}^{N} B_{EF, i} < W_{t} < Σ_{i = 1}^{N} (B_{EF, i} + B_{AF, i} + B_{BE, i});$

S2：按照业务类型分配业务总带宽，其中EF业务的总带宽

此时剩余总带宽W_rest＝W_t-W_EF，将剩余总带宽W_rest分配给AF业务和BE业务；S2: Allocate the total bandwidth of the business according to the business type, among which the total bandwidth of the EF business

At this time, the remaining total bandwidth W _rest =W _t -W _EF , and allocate the remaining total bandwidth W _rest to the AF service and the BE service;

S3：将各业务的总带宽分配给各个ONU，其中EF业务按照各ONU的带宽请求分配，即W_EF,i＝B_EF,i，对于AF业务和BE业务，分别采用以下带宽分配方法为各ONU分配带宽：S3: Allocate the total bandwidth of each service to each ONU, wherein the EF service is allocated according to the bandwidth request of each ONU, that is, W _EF,i = B _EF,i . For the AF service and BE service, the following bandwidth allocation methods are used for each ONU respectively. ONU allocated bandwidth:

S3.1：计算每个ONU的休眠效益因子

再计算每个ONU在不同剩余带宽时的休眠效益值g(i,j)，j表示剩余带宽，取值范围为0≤j≤W_m，对于第i＝1个ONU，当j≥B_m,1时，g(1,j)＝p₁，当j＜B_m,1时，g(1,j)＝0；对于第i＝2,3…,N个ONU，当j≥B_m,i并且g(i-1,j)≤g(i-1,j-B_m,i)+p_i时，g(i,j)＝g(i-1,j-B_m,i)+p_i，否则g(i,j)＝g(i-1,j)；S3.1: Calculate the dormancy benefit factor of each ONU

Then calculate the dormancy benefit value g(i,j) of each ONU at different remaining bandwidths, j represents the remaining bandwidth, and the value range is 0≤j≤W _m , for the i=1th ONU, when j≥B _{m ,1} , g(1,j)=p ₁ , when j<B _m,1 , g(1,j)=0; for the i=2,3...,Nth ONU, when j≥B _{m ,i} and g(i-1,j)≤g(i-1,jB _m,i )+p _i , g(i,j)=g(i-1,jB _m,i )+p _i , Otherwise g(i,j)=g(i-1,j);

S3.2：按照序号从N到1依次对每个ONU进行带宽分配，当前剩余带宽记为R，剩余带宽初始值为W_m，记最佳休眠效益为Q，初始值为0；对于第i＝N,N-1,…,2个ONU，根据步骤3.2中得到的各ONU对应不同剩余带宽的休眠效益值，如果g(i,R)＞g(i-1,R)，则W_m,i＝B_m,i，更新R＝R-B_m,i，Q＝Q+p_i，否则W_m,i＝0；对于第i＝1个ONU，如果g(N,W_m)-Q＝p₁，则W_m,1＝B_m,1，Q＝Q+p_i，否则W_m,1＝0；S3.2: According to the serial number from N to 1, the bandwidth is allocated to each ONU in sequence. The current remaining bandwidth is recorded as R, and the initial value of the remaining bandwidth is W _m , and the best dormancy benefit is recorded as Q, and the initial value is 0; for the i-th =N, N-1,..., 2 ONUs, according to the dormancy benefit value corresponding to different remaining bandwidths of each ONU obtained in step 3.2, if g(i, R)>g(i-1, R), then W _{m ,i} =B _m,i , update R=RB _m,i , Q=Q+p _i , otherwise W _m,i =0; for the i=1th ONU, if g(N,W _m )-Q= p ₁ , then W _m,1 =B _m,1 , Q=Q+p _i , otherwise W _m,1 =0;

S4：根据步骤S2和步骤S3即可得到ONU的带宽分配矩阵：S4: According to step S2 and step S3, the bandwidth allocation matrix of the ONU can be obtained:

$W W = = \begin{matrix} [\begin{matrix} {W W}_{EF EF,, 11} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {W W}_{EF EF,, i i} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {W W}_{EF EF,, N N} \\ {W W}_{AF AF,, 11} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {W W}_{AF AF,, i i} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {W W}_{AF AF,, N N} \\ {W W}_{BE BE,, 11} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {W W}_{BE BE,, i i} & \cdot &Center Dot; \cdot \cdot \cdot \cdot & {W W}_{BE BE,, N N} \end{matrix}] \end{matrix}$

OLT将带宽分配矩阵W向各ONU广播，完成接入网带宽调度。The OLT broadcasts the bandwidth allocation matrix W to each ONU to complete access network bandwidth scheduling.

其中，步骤S2中将剩余总带宽W_rest分配给AF业务和BE业务的方法为：设置优先级平衡因子k，k的取值范围为

AF业务的总带宽为BE业务的总带宽W_BE＝W_rest-W_AF。Wherein, in the step S2, the method of distributing the remaining total bandwidth W _rest to the AF service and the BE service is: setting the priority balance factor k, and the value range of k is

The total bandwidth of the AF service is The total bandwidth of the BE service W _BE =W _rest -W _AF .

其中，步骤S3.1中在计算ONU的休眠效益值之前，将ONU按休眠效益因子从大到小进行排列。Wherein, in step S3.1, before calculating the dormancy benefit value of the ONU, the ONUs are arranged according to the dormancy benefit factor from large to small.

本发明基于休眠节能机制的接入网带宽调度方法，针对系统总业务带宽无法完全满足三种类型业务请求的总带宽的情况，优先完全满足EF业务，再将剩余总带宽分配给AF业务和BE业务，对于AF业务和BE业务的ONU带宽分配，通过ONU带宽请求计算得到每个ONU的休眠效益因子，再计算得到每个ONU在不同剩余带宽时的休眠效益值，在进行ONU带宽分配时，通过比较给ONU分配带宽和不给ONU分配带宽的休眠效益值来判断是否给该ONU分配带宽，从而得到最佳休眠效益。在进行剩余总带宽分配时，还可通过设置优先级平衡因子来保证QoS，从而提高系统综合性能。并且，本发明完全采用软控制的方式来达到节能效果，没有增加硬件成本。The access network bandwidth scheduling method based on the dormant energy-saving mechanism of the present invention aims at the situation that the total service bandwidth of the system cannot fully satisfy the total bandwidth requested by the three types of services, giving priority to fully satisfying the EF service, and then allocating the remaining total bandwidth to the AF service and BE For the ONU bandwidth allocation of AF service and BE service, the dormancy benefit factor of each ONU is calculated through the ONU bandwidth request, and then the dormancy benefit value of each ONU at different remaining bandwidths is calculated. When performing ONU bandwidth allocation, It is judged whether to allocate bandwidth to the ONU by comparing the dormancy benefit values of allocating bandwidth to the ONU and not allocating bandwidth to the ONU, so as to obtain the best dormancy benefit. When allocating the remaining total bandwidth, QoS can also be guaranteed by setting the priority balance factor, thereby improving the overall performance of the system. Moreover, the present invention fully adopts the soft control method to achieve the energy-saving effect without increasing the hardware cost.

附图说明Description of drawings

图1是本发明基于休眠节能机制的接入网带宽调度方法一种具体实施方式流程图；Fig. 1 is a flow chart of a specific embodiment of an access network bandwidth scheduling method based on a dormant energy-saving mechanism in the present invention;

图2是计算第i＝1个ONU的休眠效益值的流程图；Fig. 2 is the flowchart of calculating the dormancy benefit value of the i=1 ONU;

图3是计算第i＝2,3…,N个ONU的休眠效益值的流程图；Fig. 3 is the flow chart of calculating the dormancy benefit value of the i=2, 3..., N ONUs;

图4是ONU带宽分配的流程图。Fig. 4 is a flowchart of ONU bandwidth allocation.

具体实施方式Detailed ways

下面结合附图对本发明的具体实施方式进行描述，以便本领域的技术人员更好地理解本发明。需要特别提醒注意的是，在以下的描述中，当已知功能和设计的详细描述也许会淡化本发明的主要内容时，这些描述在这里将被忽略。Specific embodiments of the present invention will be described below in conjunction with the accompanying drawings, so that those skilled in the art can better understand the present invention. It should be noted that in the following description, when detailed descriptions of known functions and designs may dilute the main content of the present invention, these descriptions will be omitted here.

实施例Example

图1是本发明基于休眠节能机制的接入网带宽调度方法一种具体实施方式流程图。如图1所示，本发明基于休眠节能机制的接入网带宽调度方法包括以下步骤：Fig. 1 is a flow chart of a specific embodiment of the access network bandwidth scheduling method based on the sleep energy saving mechanism of the present invention. As shown in Figure 1, the access network bandwidth scheduling method based on the dormant energy-saving mechanism of the present invention includes the following steps:

S101：中心局的光线路终端OLT收集来自各个ONU对各类型业务的带宽请求，记为带宽请求矩阵B：S101: The optical line terminal OLT of the central office collects bandwidth requests from each ONU for various types of services, which is recorded as a bandwidth request matrix B:

$B B = = \begin{matrix} [\begin{matrix} {B B}_{EF EF,, 11} & \cdot \cdot \cdot &Center Dot; \cdot &Center Dot; & {B B}_{EF EF,, i i} & \cdot \cdot \cdot &Center Dot; \cdot &Center Dot; & {B B}_{EF EF,, N N} \\ {B B}_{AF AF,, 11} & \cdot &Center Dot; \cdot \cdot \cdot &Center Dot; & {B B}_{AF AF,, i i} & \cdot &Center Dot; \cdot \cdot \cdot &Center Dot; & {B B}_{AF AF,, N N} \\ {B B}_{BE BE,, 11} & \cdot &Center Dot; \cdot &Center Dot; \cdot \cdot & {B B}_{BE BE,, i i} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {B B}_{BE BE,, N N} \end{matrix}] \end{matrix}$

B_m,i表示第i个ONU对m类业务的带宽请求，m∈{EF,AF,BE}，EF表示该业务类型为加速转发型业务，AF表示该业务类型为确保转发型业务，BE表示该业务类型为尽力转发型业务，i＝1,2,…,N，N表示ONU的数量。本发明针对的是系统总业务带宽无法完全满足三种类型业务请求的总带宽的情况，将系统总业务带宽记为W_t，并且 $Σ_{i = 1}^{N} B_{EF, i} < W_{t} < Σ_{i = 1}^{N} (B_{EF, i} + B_{AF, i} + B_{BE, i}) .$ B _m,i represents the bandwidth request of the i-th ONU for m-type services, m∈{EF,AF,BE}, EF indicates that the service type is accelerated forwarding service, AF indicates that the service type is guaranteed forwarding service, BE Indicates that the service type is a best-effort forwarding service, i=1, 2, ..., N, and N represents the number of ONUs. The present invention is aimed at the situation that the total service bandwidth of the system cannot fully satisfy the total bandwidth of the three types of service requests, and the total service bandwidth of the system is recorded as W _t , and $Σ_{i = 1}^{N} B_{EF, i} < W_{t} < Σ_{i = 1}^{N} (B_{EF, i} + B_{AF, i} + B_{BE, i}) .$

S102：基于优先级平衡因子的各业务总带宽分配：S102: Allocation of the total bandwidth of each service based on the priority balance factor:

本发明中，将完全满足EF业务的请求带宽，即分配给EF业务的总带宽

此时剩余总带宽W_rest＝W_t-W_EF，可见此时剩余总带宽W_rest无法完全满足AF业务和BE业务的带宽请求，因此在向AF业务和BE业务分配带宽的时候，只能部分满足，即

W_{AF} < Σ_{i = 1}^{N} B_{AF, i}, W_{BE} < Σ_{i = 1}^{N} B_{BE, i} .

In the present invention, the requested bandwidth of the EF service will be fully satisfied, that is, the total bandwidth allocated to the EF service

At this time, the remaining total bandwidth W _rest = W _t - W _EF , it can be seen that the remaining total bandwidth W _rest cannot fully meet the bandwidth requests of the AF service and the BE service at this time, so when allocating bandwidth to the AF service and BE service, only a part satisfied, that is

W_{AF} < Σ_{i = 1}^{N} B_{AF, i}, W_{BE} < Σ_{i = 1}^{N} B_{BE, i} .

将剩余总带宽W_rest分配给AF业务和BE业务时，可以按照设置的比例进行分配。为了保证两种业务之间的相对公平性，防止由于某业务的业务量过大而占据了另一业务的带宽，而使得另一业务的大部分业务无法正常运行的情况发生，因此在本实施方式中，可以采用优先级平衡因子对剩余总带宽W_rest进行分配，具体方法为：When allocating the remaining total bandwidth W _rest to the AF service and the BE service, it may be allocated according to a set ratio. In order to ensure the relative fairness between the two services and prevent the situation that most of the services of the other service cannot run normally due to the excessive traffic of one service occupying the bandwidth of the other service, in this implementation In the method, the priority balance factor can be used to allocate the remaining total bandwidth W _rest , and the specific method is as follows:

设置优先级平衡因子k，k的取值范围为

AF业务的总带宽为BE业务的总带宽W_BE＝W_rest-W_AF。采用优先级平衡因子k可以调节AF业务和BE业务的QoS性能。Set the priority balance factor k, the value range of k is

The total bandwidth of the AF service is The total bandwidth of the BE service W _BE =W _rest -W _AF . The QoS performance of the AF service and the BE service can be adjusted by using the priority balance factor k.

S103：对EF业务进行ONU带宽分配，由于

因此EF业务的总带宽W_EF按照各ONU的带宽请求分配，即W_EF,i＝B_EF,i。S103: Perform ONU bandwidth allocation for EF services, because

Therefore, the total bandwidth W _EF of the EF service is allocated according to the bandwidth request of each ONU, that is, W _EF,i =B _EF,i .

S104：根据休眠效益因子对AF、BE业务进行ONU带宽分配：S104: Perform ONU bandwidth allocation for AF and BE services according to the dormancy benefit factor:

从步骤S102可知，此时分配给AF业务和BE业务的总带宽W_AF和W_BE不能满足所有ONU，只能提供给部分ONU，则那些未分配到带宽的ONU将在下个轮询周期内不进行上行数据的发送，进入半休眠状态，从而实现节能。为了使休眠的节能效益最大化，本发明引入了休眠效益因子p_i来分配AF业务和BE业务的ONU带宽，其具体步骤包括：It can be seen from step S102 that the total bandwidth W _AF and W _BE allocated to the AF service and BE service cannot satisfy all ONUs at this time, and can only be provided to some ONUs, and those ONUs that have not been allocated bandwidth will not be available in the next polling cycle. Send uplink data and enter a semi-sleep state, thereby realizing energy saving. In order to maximize the energy-saving benefit of dormancy, the present invention introduces the dormancy benefit factor _pi to distribute the ONU bandwidth of AF service and BE service, and its specific steps include:

S4.1：计算每个ONU的休眠效益因子

此时可以将ONU按照休眠效益因子p_i的大小，即申请带宽的大小进行降序排列，使休眠效益值大的ONU优先满足其申请的带宽，从而保证系统整体获取最大的休眠效益值。S4.1: Calculate the dormancy benefit factor of each ONU

At this time, the ONUs can be arranged in descending order according to the size of the dormancy benefit factor _pi , that is, the size of the application bandwidth, so that the ONU with a large dormancy benefit value can first satisfy the bandwidth applied for, thereby ensuring that the system as a whole obtains the maximum dormancy benefit value.

S4.2：根据步骤S4.1得到的休眠效益因子计算每个ONU在不同剩余带宽时的休眠效益值g(i,j)，j表示剩余带宽，取值范围为0≤j≤W_n。可见当j＜B_m,i时，第i个ONU无法得到请求的带宽，即不给第i个ONU分配带宽，此时休眠效益值g(i,j)＝g(i-1,j)；而当j≥B_m,i时，有两种情况：S4.2: Calculate the dormancy benefit value g(i, j) of each ONU at different remaining bandwidths according to the dormancy benefit factor obtained in step S4.1, where j represents the remaining bandwidth, and the value range is 0≤j≤W _n . It can be seen that when j<B _m,i , the i-th ONU cannot obtain the requested bandwidth, that is, no bandwidth is allocated to the i-th ONU. At this time, the sleep benefit value g(i,j)=g(i-1,j) ; and when j≥B _m,i , there are two cases:

情况1：不给第i个ONU分配带宽，同样此时休眠效益值g(i,j)＝g(i-1,j)；Case 1: Do not allocate bandwidth to the i-th ONU, and at this time the dormancy benefit value g(i,j)=g(i-1,j);

情况2：给第i个ONU分配带宽，此时产生休眠效益p_i，剩余带宽为j-B_m,i，由于本发明采用倒序的方式进行ONU带宽分配，因此此时的剩余带宽j-B_m,i可分配给B_m,1,B_m,2,…B_m,i-1，此时的休眠效益值g(i,j)＝g(i-1,j-B_m,i)+p_i。Case 2: Allocate bandwidth to the i-th ONU. At this time, dormancy benefit p _i is generated, and the remaining bandwidth is jB _m,i . Since the present invention uses a reverse order method to allocate ONU bandwidth, the remaining bandwidth jB _m,i at this time can be Assigned to B _m,1 , B _m,2 ,...B _m,i-1 , the dormancy benefit value at this time g(i,j)=g(i-1,jB _m,i )+p _i .

本发明所期望得到的最佳休眠效益即为情况1和情况2两种休眠效益值的最大值，于是有递推关系式为：The best dormancy benefit that the present invention expects to obtain is the maximum value of two kinds of dormancy benefit values of situation 1 and situation 2, so there is a recursive relational expression as:

$g g ((i i,, j j)) = = \{\begin{matrix} g g ((i i - - 11,, j j)),, & j j < < {B B}_{m m,, i i} \\ max max ((g g ((i i - - 11,, j j)),, g g ((i i - - 11,, j j - - {B B}_{m m,, i i})) + + {p p}_{i i})),, & j j &GreaterEqual; &Greater Equal; {B B}_{m m,, i i} \end{matrix}$

此算法的边界条件为公式：The boundary condition for this algorithm is the formula:

$g g ((11,, j j)) = = \{\begin{matrix} {p p}_{11},, & j j < < {B B}_{m m,, 11} \\ 00,, & j j < < {B B}_{m m,, 11} \end{matrix}$

因此在计算休眠效益值时，需要先得到第i＝1个ONU的休眠效益值，然后依次计算第i＝2,3…,N个ONU的休眠效益值。图2是计算第i＝1个ONU的休眠效益值的流程图。如图2所示，计算第i＝1个ONU的休眠效益值包括以下步骤：Therefore, when calculating the dormancy benefit value, it is necessary to first obtain the dormancy benefit value of the i=1 ONU, and then calculate the dormancy benefit value of the i=2, 3...,N ONUs in sequence. FIG. 2 is a flow chart of calculating the dormancy benefit value of the i=1th ONU. As shown in Figure 2, calculating the dormancy benefit value of the i=1 ONU includes the following steps:

S201：初始化剩余带宽j＝0。S201: Initialize the remaining bandwidth j=0.

S202：判断是否剩余带宽j≥B_m,1，如果是，进入步骤S203，如果不是，进入步骤S204。S202: Determine whether the remaining bandwidth j≥B _m,1 , if yes, go to step S203, if not, go to step S204.

S203：g(1,j)＝p₁，进入步骤S205。S203: g(1,j)=p ₁ , go to step S205.

S204：g(1,j)＝0，进入步骤S205。S204: g(1,j)=0, go to step S205.

S205：判断是否j＝W_m，如果是，计算结束，如果不是，进入步骤S206。S205: Determine whether j=W _m , if yes, the calculation ends, if not, go to step S206.

S206：令j＝j+ΔW，其中ΔW表示设置的步长，返回步骤S202。S206: set j=j+ΔW, where ΔW represents the set step size, and return to step S202.

从以上计算过程可以看出，将会求出多个不同剩余带宽j对应的休眠效益值，这些值将在之后的达到最佳休眠效益的ONU带宽分配中使用。为了使这些剩余带宽j能够覆盖在ONU带宽分配中可能得到的当前剩余带宽，在步长ΔW的设置方面，应该以带宽请求中最小值的数量级来确定，例如，当带宽请求值为1M、3M、6M、400k，此时设置的步长可以为100k。一种比较简便的方法是求所有带宽请求值的公约数。It can be seen from the above calculation process that multiple dormancy benefit values corresponding to different remaining bandwidth j will be calculated, and these values will be used in subsequent ONU bandwidth allocation to achieve the best dormancy benefit. In order to make these remaining bandwidth j cover the current remaining bandwidth that may be obtained in the ONU bandwidth allocation, the setting of the step size ΔW should be determined by the order of the minimum value in the bandwidth request, for example, when the bandwidth request value is 1M, 3M , 6M, 400k, the step size can be set to 100k at this time. A simpler method is to find the common divisor of all bandwidth request values.

图3是计算第i＝2,3…,N个ONU的休眠效益值的流程图。如图3所示，计算第i＝2,3…,N个ONU的休眠效益值包括以下步骤：Fig. 3 is a flow chart of calculating dormancy benefit values of i=2, 3..., N ONUs. As shown in Figure 3, calculating the dormancy benefit value of the i-th=2, 3..., N ONUs includes the following steps:

S301：初始化i＝2。S301: Initialize i=2.

S302：初始化j＝0。S302: Initialize j=0.

S303：判断是否剩余带宽j≥B_m,i，如果是，进入步骤S304，如果不是，进入步骤S306。S303: Determine whether the remaining bandwidth j≥B _m,i , if yes, go to step S304, if not, go to step S306.

S304：判断是否g(i-1,j)≤g(i-1,j-B_m,i)+p_i，如果是，进入步骤S305，如果不是，进入步骤S306。S304: Determine whether g(i-1,j)≤g(i-1,jB _m,i )+p _i , if yes, go to step S305, if not, go to step S306.

S305：令g(i,j)＝g(i-1,j-B_m,i)+p_i，进入步骤S307。S305: Let g(i,j)=g(i-1,jB _m,i )+p _i , go to step S307.

S306：令g(i,j)＝g(i-1,j)，进入步骤S307。S306: Let g(i,j)=g(i-1,j), go to step S307.

S307：判断是否j＝W_m，如果是，进入步骤S309，如果不是，进入步骤S308。S307: Determine whether j=W _m , if yes, go to step S309, if not, go to step S308.

S308：令j＝j+ΔW，返回步骤S303。S308: Let j=j+ΔW, return to step S303.

S309：判断是否i＝N，如果是，计算结束，如果不是，进入步骤S310。S309: Determine whether i=N, if yes, end the calculation, if not, go to step S310.

S310：令i＝i+1，返回步骤S302。S310: set i=i+1, return to step S302.

S4.3：按照序号从N到1依次对每个进行ONU带宽分配。本发明中，对于某个ONU的AF业务和BE业务的带宽请求，要么给其分配满足要求的带宽，记x_i＝1；要么不给其分配带宽，记x_i＝0；某个分配结果可产生的休眠效益即为

本发明的最终目的即获取最大的休眠效益

S4.3: Perform bandwidth allocation to each ONU sequentially according to the serial number from N to 1. In the present invention, for the bandwidth request of the AF service and BE service of a certain ONU, either allocate bandwidth that meets the requirements to it, record _xi = 1; or do not allocate bandwidth to it, record _xi = 0; a certain allocation result The resulting dormancy benefit is

The final purpose of the present invention promptly obtains maximum dormancy benefit

图4是ONU带宽分配的流程图。如图4所示，ONU带宽分配包括：Fig. 4 is a flowchart of ONU bandwidth allocation. As shown in Figure 4, ONU bandwidth allocation includes:

S401：初始化ONU序号i＝N，当前剩余带宽R＝W_m，最佳休眠效益为Q＝0。S401: Initialize the ONU serial number i=N, the current remaining bandwidth R=W _m , and the best dormancy benefit is Q=0.

S402：根据步骤S4.2中得到的各ONU对应不同剩余带宽的休眠效益值，判断是否g(i,R)＞g(i-1,R)，如果是，进入步骤S403，如果不是，进入步骤S404。可见，本步骤是通过比较给ONU分配带宽和不给ONU分配带宽的休眠效益值大小来判断是否给该ONU分配带宽。S402: According to the dormancy benefit value of each ONU corresponding to different remaining bandwidth obtained in step S4.2, judge whether g(i, R)>g(i-1, R), if yes, enter step S403, if not, enter Step S404. It can be seen that in this step, it is judged whether to allocate bandwidth to the ONU by comparing the dormancy benefit value of allocating bandwidth to the ONU and not allocating bandwidth to the ONU.

S403：为第i个ONU分配带宽，即W_m,i＝B_m,i，更新当前剩余带宽R＝R-B_m,i，最佳休眠效益Q＝Q+p_i，进入步骤S405。S403: Allocate bandwidth for the i-th ONU, that is, W _m,i =B _m,i , update the current remaining bandwidth R=RB _m,i , best dormancy benefit Q=Q+p _i , and enter step S405.

S404：不为第i个ONU分配带宽，即W_m,i＝0，当前剩余带宽R和最佳休眠效益Q均不变，进入步骤S405。S404: Do not allocate bandwidth to the i-th ONU, that is, W _m,i =0, the current remaining bandwidth R and the best dormancy benefit Q remain unchanged, and enter step S405.

S405：令i＝i-1。S405: Set i=i-1.

S406：判断是否i＝1，如果是，进入步骤S407，如果不是，返回步骤S402。S406: Determine whether i=1, if yes, go to step S407, if not, go back to step S402.

S407：判断是否g(N,W_m)-Q＝p₁，如果是，进入步骤S408，如果不是，进入步骤S409。S407: Determine whether g(N,W _m )-Q=p ₁ , if yes, go to step S408, if not, go to step S409.

S408：给第i＝1个ONU分配带宽，即W_m,1＝B_m,1，Q＝Q+p_i，ONU带宽分配结束。S408: Allocate bandwidth to the i=1th ONU, that is, W _m,1 =B _m,1 , Q=Q+p _i , and ONU bandwidth allocation ends.

S409：不给第i＝1个ONU分配带宽，即W_m,1＝0，ONU带宽分配结束。S409: Do not allocate bandwidth to the i=1th ONU, that is, W _m,1 =0, ONU bandwidth allocation ends.

S105：根据步骤S103和步骤S104即可得到ONU的带宽分配矩阵：S105: According to step S103 and step S104, the bandwidth allocation matrix of the ONU can be obtained:

$W W = = \begin{matrix} [\begin{matrix} {W W}_{EF EF,, 11} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {W W}_{EF EF,, i i} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {W W}_{EF EF,, N N} \\ {W W}_{AF AF,, 11} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {W W}_{AF AF,, i i} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {W W}_{AF AF,, N N} \\ {W W}_{BE BE,, 11} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {W W}_{BE BE,, i i} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {W W}_{BE BE,, N N} \end{matrix}] \end{matrix}$

光线路终端OLT将带宽分配矩阵W向各ONU广播，完成接入网带宽调度，分配带宽为0的ONU将在下一轮询周期中进入半休眠状态，不进行上行数据的发送，从而达到节能的效果。The optical line terminal OLT broadcasts the bandwidth allocation matrix W to each ONU to complete the bandwidth scheduling of the access network. The ONU with the allocated bandwidth of 0 will enter a semi-sleeping state in the next polling cycle and will not send uplink data, thereby achieving energy saving. Effect.

本发明完全采用软控制的方式来达到节能效果，硬件成本没有增加。它既区别于经典的DBA（Dynamic Bandwidth Allocation，动态带宽分配）算法（只能保证系统的QoS），又不同于现今的各种数据链路层的节能方式（以系统的性能来换取节能的效果），而是可以通过休眠效益因子p_i和优先级平衡因子k的设置同时实现了QoS和节能效果。The present invention fully adopts the soft control mode to achieve the energy-saving effect, and the hardware cost does not increase. It is not only different from the classic DBA (Dynamic Bandwidth Allocation, dynamic bandwidth allocation) algorithm (which can only guarantee the QoS of the system), but also different from the energy-saving methods of various data link layers today (the performance of the system is exchanged for the effect of energy saving). ), but the QoS and energy-saving effects can be realized at the same time by setting the dormancy benefit factor p _i and the priority balance factor k.

尽管上面对本发明说明性的具体实施方式进行了描述，以便于本技术领域的技术人员理解本发明，但应该清楚，本发明不限于具体实施方式的范围，对本技术领域的普通技术人员来讲，只要各种变化在所附的权利要求限定和确定的本发明的精神和范围内，这些变化是显而易见的，一切利用本发明构思的发明创造均在保护之列。Although the illustrative specific embodiments of the present invention have been described above, so that those skilled in the art can understand the present invention, it should be clear that the present invention is not limited to the scope of the specific embodiments. For those of ordinary skill in the art, As long as various changes are within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions and creations using the concept of the present invention are included in the protection list.

Claims

1. the Access Network bandwidth scheduling method based on dormancy power-saving mechanism, is characterized in that, comprises the following steps:

S1: optical line terminal OLT receives from optical network unit ONU the bandwidth request to each business, is designated as bandwidth request matrix B:

B = \begin{matrix} [\begin{matrix} B_{EF, 1} & \cdot \cdot \cdot & B_{EF, i} & \cdot \cdot \cdot & B_{EF, N} \\ B_{AF, 1} & \cdot \cdot \cdot & B_{AF, i} & \cdot \cdot \cdot & B_{AF, N} \\ B_{BE, 1} & \cdot \cdot \cdot & B_{BE, i} & \cdot \cdot \cdot & B_{BE, N} \end{matrix}] \end{matrix}

Wherein, B _m,irepresent the bandwidth request of i ONU to m class business, m ∈ { EF, AF, BE}, EF represents that this type of service is expedited forwarding type business, and AF represents that this type of service is for guaranteeing forward type business, and BE represents that this type of service is for the forward type business of doing the best, i=1,2 ..., N, N represents the quantity of ONU; The total service bandwidth of system is designated as W _t, and

Σ_{i = 1}^{N} B_{EF, i} < W_{t} < Σ_{i = 1}^{N} (B_{EF, i} + B_{AF, i} + B_{BE, i});

S2: according to type of service distribution service total bandwidth, the wherein total bandwidth of EF business

now remain total bandwidth W _rest=W _t-W _eF, will remain total bandwidth W _restdistribute to AF business and BE business;

S3: the total bandwidth of each business is distributed to each ONU, and wherein EF business is according to the bandwidth request allocation of each ONU, i.e. W _{eF, i}=B _{eF, i}, for AF business and BE business, adopt respectively following bandwidth allocation methods for each ONU distribution bandwidth:

S3.1: the dormancy benefit factor of calculating each ONU calculate the dormancy benefit value g (i, j) of each ONU when different remaining bandwidth, j represents remaining bandwidth again, and span is 0≤j≤W _m, for i=1 ONU, as j>=B _{m, 1}time, g (1, j)=p ₁, as j < B _{m, 1}time, g (1, j)=0; For i=2,3 ..., N ONU, as j>=B _m,iand g (i-1, j)≤g (i-1, j-B _m,i)+p _itime, g (i, j)=g (i-1, j-B _m,i)+p _i, otherwise g (i, j)=g (i-1, j);

S3.2: successively each ONU is carried out to allocated bandwidth according to sequence number from N to 1, current residual bandwidth is designated as R, remaining bandwidth initial value is W _m, remember that best dormancy benefit is Q, initial value is 0; For i=N, N-1 ..., 2 ONU, according to the dormancy benefit value of the corresponding different remaining bandwidths of each ONU that obtain in step 3.2, if g (i, R) > g (i-1, R), W _m,i=B _m,i, upgrade R=R-B _m,i, Q=Q+p _i, otherwise W _m,i=0; For i=1 ONU, if g is (N, W _m)-Q=p ₁, W _{m, 1}=B _{m, 1}, Q=Q+p _i, otherwise W _{m, 1}=0;

S4: the allocated bandwidth matrix that can obtain ONU according to step S2 and step S3:

W = \begin{matrix} [\begin{matrix} W_{EF, 1} & \cdot \cdot \cdot & W_{EF, i} & \cdot \cdot \cdot & W_{EF, N} \\ W_{AF, 1} & \cdot \cdot \cdot & W_{AF, i} & \cdot \cdot \cdot & W_{AF, N} \\ W_{BE, 1} & \cdot \cdot \cdot & W_{BE, i} & \cdot \cdot \cdot & W_{BE, N} \end{matrix}] \end{matrix}

OLT to each ONU broadcast, completes Access Network bandwidth scheduling by allocated bandwidth matrix W.

2. the Access Network bandwidth scheduling method based on dormancy power-saving mechanism according to claim 1, is characterized in that, in described step S2, will remain total bandwidth W _restthe method of distributing to AF business and BE business is: priority balance factor k is set, and the span of k is

the total bandwidth of AF business is

the total bandwidth W of BE business _bE=W _rest-W _aF.

3. the Access Network bandwidth scheduling method based on dormancy power-saving mechanism according to claim 1, is characterized in that, in described step S3.1, before calculating the dormancy benefit value of ONU, ONU is arranged from big to small by the dormancy benefit factor.