IEEE802.3ah协议学习（1）——多点MAC控制简介

 本博客（http://blog.csdn.net/livelylittlefish）贴出作者（三二一、小鱼）相关研究、学习内容所做的笔记，欢迎广大朋友指正！

 

http://www.ieee802.org/21/doctree/2006_Meeting_Docs/2006-11_meeting_docs/802.3ah-2004.pdf

 

64. Multi-point MAC Control——多点MAC控制简介

 

64.1 Overview

 

This clause deals with the mechanism and control protocols required in order to reconcile the P2MP topology into the Ethernet framework. The P2MP medium is a passive optical network (PON), an optical network with no active elements in the signal’s paths from source to destination. The only interior elements used in a PON are passive optical components, such as optical fiber, splices, and splitters. When combined with the Ethernet protocol, such a network is referred to as Ethernet passive optical network (EPON).

本附录描述了将P2MP拓扑纳入以太网框架所需的机制和控制协议。

 

P2MP is an asymmetrical medium based on a tree (or tree-and-branch) topology. The DTE connected to the trunk of the tree is called optical line terminal (OLT) and the DTEs connected at the branches of the tree are called optical network units (ONU). The OLT typically resides at the service provider’s facility, while the ONUs are located at the subscriber premises.

P2MP是基于树状（或者是树及分枝）拓扑的非对称媒介。连接在树干上的数据终端设备（DTE）称为光线路终端（OLT），连接在树枝上的DTE称为光网络单元（ONU）。OLT一般置于服务提供商侧，而ONU一般放在用户侧。

 

In the downstream direction (from the OLT to an ONU), signals transmitted by the OLT pass through a 1:N passive splitter (or cascade of splitters) and reach each ONU. In the upstream direction (from the ONUs to the OLT), the signal transmitted by an ONU would only reach the OLT, but not other ONUs. To avoid data collisions and increase the efficiency of the subscriber access network, ONU’s transmissions are arbitrated. This arbitration is achieved by allocating a transmission window (grant) to each ONU. An ONU defers transmission until its grant arrives. When the grant arrives, the ONU transmits frames at wire speed during its assigned time slot.

在下行方向（从OLT到ONU），OLT发送的信号通过一个1:N的无源分路器（或几个分路器的级联）到达各个ONU。在上行方向（从各个ONU到OLT），一个ONU发送的信号只会到达OLT，而不会到达其他ONU。为避免数据冲突并提高用户接入网的利用效率，需要对ONU的传输进行仲裁。这种仲裁是通过给各ONU分配发送窗口（授权）来实现的。每个ONU只能等到其授权到达后才可以发送数据。当一个ONU的授权到达后，该ONU在所指配的时间窗口内以线速发送数据帧。

 

A simplified P2MP topology example is depicted in Figure 64–1. Clause 67 provides additional examples of P2MP topologies.

 

 

 

Topics dealt with in this clause include allocation of upstream transmission resources to different ONUs, discovery and registration of ONUs into the network, and reporting of congestion to higher layers to allow for dynamic bandwidth allocation schemes and statistical multiplexing across the PON.

本附录讨论的内容涉及在不同ONU之间分配上行发送资源，ONU加入网络的发现和注册，以及向上层报告拥塞状态以进行PON系统内的动态带宽分配和统计复用。

 

This clause does not deal with topics including bandwidth allocation strategies, authentication of end-devices, quality-of-service definition, provisioning, or management.

本附录不涉及具体的带宽分配策略、终端设备认证、服务质量定义、提供以及管理。

 

This clause specifies the multi-point control protocol (MPCP) to operate an optical multi-point network by defining a Multi-point MAC Control sublayer as an extension of the MAC Control sublayer defined in Clause 31, and supporting current and future operations as defined in Clause 31 and annexes.

本附录规定了多点控制协议(MPCP)，该协议定义了一个多点MAC控制子层以控制一个光多点网络。

 

 

 

Each PON consists of a node located at the root of the tree assuming the role of OLT, and multiple nodes located at the tree leaves assuming roles of ONUs. The network operates by allowing only a single ONU to transmit in the upstream direction at a time. The MPCP located at the OLT is responsible for timing the different transmissions. Reporting of congestion by the different ONUs may assist in optimally allocating the bandwidth across the PON.

每个PON由一个OLT节点和多个ONU节点组成，其中OLT是树的根节点，ONU是树的叶节点。PON网络在任何一个时刻都只允许一个ONU在上行方向发送数据。OLT中的MPCP协议负责对各传输进行准确定时。各ONU报告的拥塞状态信息有助于优化PON网络内的带宽资源分配。

 

Automatic discovery of end stations is performed, culminating in registration through binding of an ONU to an OLT port by allocation of a Logical Link ID (see LLID in 65.1.3.3.2), and dynamic binding to a MAC connected to the OLT.

在ONU的自动发现过程中，通过分配逻辑链路ID（LLID）将ONU绑定到OLT的一个端口上，同时将MAC动态绑定到此OLT上，最终实现ONU的注册。

 

The Multi-point MAC Control functionality shall be implemented for subscriber access devices containing point-to-multipoint physical layer devices defined in Clause 60.

实施多点MAC控制功能，要求用户接入设备包含点到多点物理层。

 

64.1.1 Goals and objectives

 

The goals and objectives of this clause are the definition of a point-to-multi-point Ethernet network utilizing an optical medium.

 

Specific objectives met include:

a) Support of Point to Point Emulation (P2PE) as specified

b) Support multiple LLIDs and MAC Clients at the OLT

c) Support a single LLID per ONU

d) Support a mechanism for single copy broadcast

e) Flexible architecture allowing dynamic allocation of bandwidth

f) Use of 32 bit timestamp for timing distribution

g) MAC Control based architecture

h) Ranging of discovered devices for improved network performance

i) Continuous ranging for compensating round trip time variation

一个采用光介质的点到多点以太网络应具有以下特征：
a) 支持规范定义的点到点仿真（P2PE）；
b) OLT 支持多个LLID 和MAC 客户端；
c) 每个ONU 至少支持一个LLID；
d) 支持单拷贝广播机制；
e) 支持动态带宽分配的灵活体系结构；
f) 使用32 比特时间戳来发布定时信息；
g) 基于MAC 控制的体系结构；
h) 对已发现设备进行测距来提高网络性能；
i) 进行连续测距以补偿往返时间的变化。

 

64.1.2 Position of Multi-point MAC Control within the IEEE 802.3 hierarchy

 

Multi-point MAC Control defines the MAC control operation for optical point-to-multi-point networks. Figure 64–2 depicts the architectural positioning of the Multi-point MAC Control sublayer with respect to the MAC and the MAC Control client. The Multi-point MAC Control sublayer takes the place of the MAC Control sublayer to extend it to support multiple clients and additional MAC control functionality.

多点MAC控制定义了点到多点光网络的MAC控制。图 2描绘了多点MAC控制子层和MAC子层、MAC控制客户端的关系。多点MAC控制子层取代并扩展了MAC控制子层以支持多个客户端及其它的MAC控制功能。

 

Multi-point MAC Control is defined using the mechanisms and precedents of the MAC Control sublayer. The MAC Control sublayer has extensive functionality designed to manage the real-time control and manipulation of MAC sublayer operation. This clause specifies the extension of the MAC Control mechanism to manipulate multiple underlying MACs simultaneously. This clause also specifies a specific protocol implementation for MAC Control.

多点MAC控制的定义延续了MAC控制子层的定义，MAC控制子层定义参见IEEE 802.3 31章。MAC控制子层设置了多种功能来处理实时控制和MAC子层操作。本附录定义了MAC控制机制的扩展功能以支持同时操作多个下层MAC。本附录同时还定义了MAC控制的一种具体协议实现。

 

The Multi-point MAC Control sublayer is specified such that it can support new functions to be implemented and added to this standard in the future. Multi-Point Control Protocol (MPCP), the management protocol for P2MP is one of these protocols. Non-real-time, or quasi-static control (e.g., configuration of MAC operational parameters) is provided by Layer Management. Operation of the Multipoint MAC Control sublayer is transparent to the MAC.

本附录定义的多点MAC控制子层能够支持未来新功能的实现并允许将这些新功能加入到本标准中，多点控制协议（MPCP）和P2MP的管理协议等均属于此类。非实时或准静态的控制（例如MAC操作参数的配置等）由层管理提供。多点MAC控制子层的操作对MAC是透明的。

 

As depicted in Figure 64–2, the layered system instantiates multiple MAC entities, using a single physical layer. The individual MAC instances offer a Point-to-point emulation service between the OLT and the ONU. An additional MAC is instantiated to communicate to all ONUs at once. This instance takes maximum advantage of the broadcast nature of the downstream channel by sending a single copy of a frame that is received by all ONUs. This MAC instance is referred to as Single Copy Broadcast (SCB).

如图 2所示，分层系统包括多个MAC实体和一个物理层。每个MAC实体在OLT和ONU间提供点到点仿真业务。对于MAC实体与所有的ONU同时通信的特殊情况，可通过发送所有ONU都能接收到的单个帧拷贝来实现，这最大限度地利用了下行信道的广播特性，称为单拷贝广播(SCB)。

 

The ONU only requires one MAC instance since frame filtering operations are done at the RS layer before reaching the MAC. Therefore, MAC and layers above are emulation-agnostic at the ONU (see 65.1.3.3).

由于帧在到达MAC前由RS子层进行过滤，所以每个ONU只需要一个MAC实体。这样使得ONU MAC下层的仿真功能对MAC及其上层是透明的。

 

Although Figure 64–2 and supporting text describe multiple MACs within the OLT, a single unicast MAC address may be used by the OLT. Within the EPON Network, MACs are uniquely identified by their LLID which is dynamically assigned by the registration process.

虽然图 2中OLT内有多个MAC，但实际上OLT也可以只使用一个单播MAC地址。在EPON网络内，MAC是通过注册过程中动态指配的LLID来唯一标识的。

64.1.3 Functional block diagram

 

Figure 64–3 provides a functional block diagram of the Multi-point MAC Control architecture.

 

64.1.4 Service interfaces

 

The MAC Client communicates with the Control Multiplexer using the standard service interface specified in Clause 2.3. Multi-point MAC Control communicates with the underlying MAC sublayer using the standard service interface specified in 4A.3.2. Similarly, Multi-point MAC Control communicates internally using primitives and interfaces consistent with definitions in Clause 31.

MAC客户端通过标准的服务接口和控制复用器通信。多点MAC控制子层通过标准的服务接口和下层的MAC子层通信。类似地，多点MAC控制子层内部通过原语和接口进行通信。

 

64.1.5 State diagram conventions

 

The body of this standard comprises state diagrams, including the associated definitions of variables, constants, and functions. Should there be a discrepancy between a state diagram and descriptive text, the state diagram prevails. The notation used in the state diagrams follows the conventions of 21.5. State diagram timers follow the conventions of 14.2.3.2 augmented as follows:

状态图包括相关的变量、常量、函数的定义。状态图中的标注和定时器约定如下：

 

a) [start x_timer, y] sets expiration of y to timer x_timer.

b) [stop x_timer] aborts the timer operation for x_timer asserting x_timer_not_done indefinitely.

[start x_timer, y]设置定时器x_timer的失效时间y。

[stop x_timer]通过设置x_timer_not_done为不确定，放弃对定时器x_timer的操作。

The state diagrams use an abbreviation MACR as a shorthand form for MA_CONTROL.request and MACI as a shorthand form for MA_CONTROL.indication.

状态图使用如下缩写， 例如MA_CONTROL.request 缩写为MACR，MA_CONTROL.indication缩写为MACI。

The vector notations used in the state diagrams for bit vector use 0 to mark the first received bit and so on (for example data[0:15]), following the conventions of 3.1 for bit ordering. When referring to an octet vector, 0 is used to mark the first received octet and so on (for example m_sdu[0..1]).

状态图中的比特向量的标注使用0表示第一个接收到的比特，以此类推（例如数据[0:15]）。当使用字节向量时，0用来标识第一个接收到的字节，以此类推（例如m_sdu[0..1]）。

a < b: A function that is used to compare two (cyclic) time values. Returned value is true when b is larger than a allowing for wrap around of a and b. The comparison is made by subtracting b from a and testing the MSB. When MSB(a-b) = 1 the value true is returned, else false is returned. In addition, the following functions are defined in terms of a < b:

a>b is equivalent to !(a<b or a=b)

a≥b is equivalent to !(a<b)

a≤b is equivalent to !(a>b)

a<b：用于比较两个（周期）时间值的函数，如b大于a那么返回值为真。这个比较是将a减去b并检查MSB得到。如果MSB(a-b)=1，那么返回值为真，否则返回值为假。同时，还定义了如下函数：
a>b等价于!(a<b或a=b)；
a≥b等价于!(a<b)；
a≤b等价于!(a>b)。

 

http://www.ieee802.org/21/doctree/2006_Meeting_Docs/2006-11_meeting_docs/802.3ah-2004.pdf