CN101193060A - Method for reliable E1 transmission based on forward error correction mechanism in packet network - Google Patents

Method for reliable E1 transmission based on forward error correction mechanism in packet network Download PDF

Info

Publication number
CN101193060A
CN101193060A CN 200610163606 CN200610163606A CN101193060A CN 101193060 A CN101193060 A CN 101193060A CN 200610163606 CN200610163606 CN 200610163606 CN 200610163606 A CN200610163606 A CN 200610163606A CN 101193060 A CN101193060 A CN 101193060A
Authority
CN
Grant status
Application
Patent type
Prior art keywords
packet
e1
end
packet network
fec
Prior art date
Application number
CN 200610163606
Other languages
Chinese (zh)
Other versions
CN101193060B (en )
Inventor
余少华
戴锦友
Original Assignee
武汉烽火网络有限责任公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date

Links

Abstract

A packet network reliable end-to-end E1 transmission realization method which adopts the FEC mechanism is provided. At the end where E1 enters the packet network, E1 data flow is stored in the buffer area (101) and then encapsulated into data packets (102) and then generates FEC packets. The encapsulated data packets and FEC packets are transmitted (103) on the packet network (107). At the end where E1 leaves the packet network, the receiving timer is set to control time delay and the de-dithering buffer area is set to eliminate the influence of the packet network time delay dithering. E1 data are extracted from the received data packets and are then stored in the de-dithering buffer area (104). FEC packets are stored in the error correcting buffer area. When the receiving timer overflows, the packets in the buffer area are checked, lost data packets are calculated and lost information (105) are restored. At the same time E1 clock is restored and E1 code stream (106) is restored according to the clock and the de-dithering buffer area. The invention can also guarantee remote seamless interconnection and intercommunication without FEC mechanism.

Description

在分组网上采用前向纠错机制实现可靠的El传送的方法技术领域 Reliable error correction mechanism prior to the packet network transmission method of El Field

本发明涉及分组网传输E1的通信方法,特别是涉及在分组网上实现传送El业务的FEC (Forward Error Correction,前向纠错,下同)方法。 The present invention relates to a communication method for the packet network transmission of E1, particularly to achieve transmission traffic over packet networks El FEC (Forward Error Correction, FEC, hereinafter the same) method.

背景技术 Background technique

1. TDM (Time-DivisionMultiplexing,时分复用)技术TDM是时分复用技术的缩写,是一种分时传送多个数据流的技术,它为 1. TDM (Time-DivisionMultiplexing, time-division multiplexing) technology is an abbreviation for time division multiplexing TDM techniques, is a technique for transmitting multiple data streams sharing, it is

每个数据流分配一个时隙,并在同一传输信道上重复传送一固定的时隙序列。 Each data stream is allocated one slot, and repeatedly transmitted a fixed sequence of time slots on the same transmission channel. 传统电信网采用PDH和SONET/SDH系统以TDM方式提供传输服务,主 Traditional telecommunications network using PDH and SONET / SDH systems provide transmission services to the TDM main

要向用户提供语音服务,这种方式虽然可以提供高可靠的服务,但其成本高、 To provide users with voice services, although in this way can provide highly reliable service, but its high cost,

价格昂贵、带宽有限、敷设周期长的缺点也非常突出。 Expensive, limited bandwidth, long drawback is also very prominent laying cycle.

另一方面,从当前网络业务和应用的发展趋势来看,SONET/SDH技术 On the other hand, from the current development trend of network services and applications point of view, SONET / SDH technology

经过二十几年的发展,在网上已经有大量应用,应用规4莫已达几十亿美元。 After 20 years of development, the Internet has been a large number of applications, application of regulations 4 Mo has reached billions of dollars. E1是TDM的一次群。 E1 is a group of TDM.

2. IP和以太网 2. IP and Ethernet

基于IP和基于以太网是应用最为广泛的两种分组传输技术,但IP和以太网处于网络的不同层次。 And different levels of IP-based Ethernet is the most widely used two kinds of packet transmission technology, IP and Ethernet in the network. IP是网络层技术,其在Internet的巨大成功使其在分组网络中处于绝对性的统治地位,值得注意的是,IP多在以太网上承载。 IP is the network layer technology, its huge success in the Internet it is in absolute dominance of the packet network, it is worth noting that, multi-bearer IP over Ethernet. 以太网处于网络层以下,由于其普遍存在(承载90%以上的接入数据流量),简单易用和高性价比等特性,实质是应用最广泛的分组网络技术。 Ethernet is below the network layer, because of its ubiquitous (bearer access data traffic over 90%), ease of use and high cost and other properties, the substance is the most widely used packet network technology.

现有网络向分组网演进是大势所趋,但演进的过程中基于El的网络和基于分组的网络是同时存在的。 Existing networks to packet network evolution is the trend, but the process of evolution based El networks and packet-based networks exist simultaneously.

分组网同时承载语音,视频和数据等业务是必要的也是必然的,而当前的语音业务多以El传输。 Other packet network is simultaneously carry voice, video and data services needed it is inevitable, and the current voice service to multiple transmit El. 因而以诸如IP和以太网的分组网传送E1是必要的。 Therefore it is necessary to transmit E1 packet network such as IP and Ethernet.

3. 诸如IP的分组网的QOS保证技术 3. A packet network such as IP QOS guarantees the art

诸如IP的分组网的QoS问题是解决分组网承载多业务的关键因素,从当前的研究成果看,主要有以下几种解决方案。 QoS issues, such as IP packet network is a key factor in solving the packet network carrying multiple services, from the current research, there are several solutions. 〈1SlntServ (Integrated Service,集成月良务) <1SlntServ (Integrated Service, Integrated month good service)

针对IPQoS的问题,IETF在早期提出了IntServ模型。 For questions IPQoS of, IETF proposed in the early IntServ model. IntServ模型又称为集成服务模型,其基本思想是在传送数据之前,根据业务的QoS需求进行网络资源预留,从而为该数据流提供端到端的QoS保证。 IntServ model, also known as integrated services model, the basic idea is before the transmission of data, network QoS resource reservation according to the needs of the business to provide end to end QoS guarantee for the data stream. 为此,集成服务通常采用面向流的资源预留协议(RSVP),在流传输路径上的每个节点为流预留并维护资源。 For this purpose, usually integrated service resource reservation protocol (RSVP) facing the flow, each node on the path to the streaming flow and maintains the resource reservation. 主机利用RSVP向网络为应用流提出QoS的请求;路由器利用RSVP将QoS请求信息传给流的路径中的其他路由器,并建立和保存该服务的信息;RSVP请求将会使得沿着数据路径的资源在路由器处预留。 Host to the network using RSVP QoS request of an application stream; router using the RSVP QoS route request information to the other routers in the stream, and create and save the information and services; RSVP requests will be such that the resources along the data path set aside at the router. 这种模型的优点是能提供端到端的绝对的QoS保证。 The advantage of this model is the ability to provide end to end QoS guarantee absolute.

但这种模型在实现上是非常困难的,主要体现在: However, this model is realized in the very difficult, mainly reflected in:

(1) 由于预留是基于每个流而进行的,因此使得节点中要保留每个流的状态信息,导致核心路由器负担太重,因此可扩展性很差。 (1) Since the reservation is performed on a per-stream, thereby enabling the node to maintain state information for each stream, resulting in heavy burden on the core router, so scalability is poor.

(2) 网络中每个节点都要维护各类数据库,并实现复杂的功能模块(如资源预留、路由、接纳控制等),造成了极大的复杂性。 (2) the network each node needs to maintain various databases, and implement complex functional modules (such as resource reservation, routing, admission control, etc.), resulting in a great deal of complexity.

IntServ适用于较小规模的网络。 IntServ for smaller networks.

<2>DiffServ (Differentiated Services,区分服务,或细分业务) <2> DiffServ (Differentiated Services, DiffServ, or business segments)

由于IntServ的局限性,IETF又提出了DiffServ模型,又称为区分服务模型。 Due to the limitations of IntServ, IETF DiffServ also proposed model, also known as differentiated service model. 区分服务模型的基本思想是在网络的入口处为每个数据包分类,并在数据包中标记相应的区分服务代码点(DSCP, DifflServCodePoint),用于指示数据包在网络转发路径的中间节点上被处理的方式。 The basic idea is to DiffServ ingress to a network to classify each packet, and mark the corresponding Differentiated Services Code Point (DSCP, DifflServCodePoint) in the data packet, to the intermediate node indicating the packet transfer path in the network way to be treated. 在网络内部的核心路由器中只保存简单的DSCP与PHB (每跳行为)的对应机制,根据数据包头部中的DSCP值对数据包进行相应的优先级转发,而业务流状态信息的保存与流量控制机制的实现等都在网络边界节点进行,内部节点是与状态无关的。 Save simply the DSCP and the PHB (Per Hop Behavior) within the network core routers corresponding mechanism, based on the DSCP values ​​of the packet header in the data packet corresponding priority forwarding, and preservation and service flow state information control mechanisms implemented in the network so the border node, interior node is state independent.

区分服务具有实现简单,扩展性好的特点。 Differentiated services have simple, good scalability. 目前在IP网中区分服务得到了绝大部分厂家的支持,其具体实现技术包括分类、重标记、速率限制、流量整形、拥塞避免、队列调度等。 Currently distinguish IP network services supported by the majority of manufacturers, specific implementation techniques include classification, marking, rate limiting, traffic shaping, congestion avoidance, and queue scheduling.

但区分服务的局限性也很突出,主要体现在: But the limitations of differentiated services is also very prominent, mainly reflected in:

(1) 区分服务只承诺相对的服务质量,因而不能对用户提供绝对的服务质量保证。 (1) Differentiated Services only promise relative quality of service, and therefore can not provide absolute assurance of quality service to users.

(2) 在拥塞发生时,区分服务模型只能采取丟弃报文的方式,而不能采用例如旁路的方式使部分流量通过其他路径到达终点', (2) when congestion occurs, discard only take DiffServ packets manner, using, for example but not in the bypass mode allow traffic through other paths to reach the end ',

(3 )对相同优先级的业务而言,设备在拥塞时对才艮文的丟弃是非智能化的,也就是说,设备只能随机地丢弃报文,其结杲是所有业务的服务质量都受到影响。 (3) For the same priority traffic, the device is discarded during congestion of a non-intelligent Gen only text, that is, the device can randomly discarding packets, which junction Gao quality of service for all traffic They are affected. 而此时希望的结果是只丟弃少部分业务流的报文,从而避免剩下的大多数的业务流的服务质量受到影响。 At a time when the desired result is only a small part of the business of dropped packets flow, thus avoiding most of the rest of the quality of service traffic flow is affected.

<3> IntServ与DiffServ结合 <3> IntServ and DiffServ binding

目前业界还提出了把IntServ与DiffServ结合的方式,其思路为:在用户网络仍使用RSVP,在运营商的DiffServ网络边界将IntServ的业务类型映射为DiffServ的业务类型,这样利用IntServ的架构来解决端到端的QoS( Quality of Service,服务质量),同时也利用DiffServ来提供好的扩展性。 The industry also proposed ways to IntServ and DiffServ combination, the idea is: still use RSVP in the user network, in DiffServ network boundary operators to map the IntServ service type for the DiffServ type of business, so take advantage of IntServ architecture to solve end to end QoS (quality of service, quality of service), but also use DiffServ to provide better scalability.

但这种方法仍然存在IntServ的信令复杂、难于管理等问题,而且由于在运营商的网络采用DiffServ,因此在这一段网络也只能提供相对的QoS,从而使端到端的服务质量得不到硬性的带宽保证。 However, this method still exists IntServ signaling complex, difficult management issues, and because the use DiffServ in the operator's network, in this period of the network can only provide relative QoS, so that the end to end quality of service can not be rigid bandwidth guarantee. 该方法目前仍处于一种理论的研究阶段。 This method is still in the research stage of a theory.

<4> MPLS ( Multiprotocol Label Switching) & QoS <4> MPLS (Multiprotocol Label Switching) & QoS

利用多协议标签交换MPLS技术,可以协助解决QoS问题。 Use Multiprotocol Label Switching MPLS technology, QoS can help solve the problem. MPLS是一种结合第二层和第三层的交换技术,引入了基于标签的机制,把路由选择和数据转发分开,由标签来规定一个分组通过网络的路径。 MPLS is a switching technique in conjunction with the second and third layers, the introduction of label-based mechanism, the routing and data forwarding are separated by a predetermined tag to a path through a packet network. MPLS网络由核心部分的标签交换路由器(LSR)、边缘部分的标签边缘路由器(LER)组成。 MPLS network from the core part of the label switching router (LSR), the edge portion of the label edge router (LER) composition.

由于MPLS采用标签交换来进行MPLS转发,因此其转发效率高于传统IP通过路由器的转发,从而通过减少转发时间来提高QoS。 Because the MPLS label switching using MPLS forwarding is performed, so the more efficient than conventional IP forwarding forwarded through the router, thereby improving QoS forwarding by reducing time. 此外,MPLS的报文头中包含一个3bit的EXP字段,通过该字段可以标记该MPLS报文的优先级,从而使设备在转发该MPLS报文时能根据优先级标志进行区别对待。 In addition, the MPLS packet header contains a 3bit EXP field may mark the priority of the MPLS packet through the field, so that the device can be treated differently when forwarding the MPLS packet according to the priority flag.

这种方式的局限性在于:首先它必须基于MPLS网络实现,而当前许多网络上并没有实施MPLS;另外随着近几年芯片技术的不断发展,路由转发与交换转发之间的性能差异也越来越小;而且通过EXP进行优先级区分实际上也是DiffServ的实现方式,因而这种方式也不可避免地具有DiffServ所具有的一些局限性。 The limitation of this approach is that: First, it must be based on MPLS networks, while many current network does not implement MPLS; another in recent years with the development of chip technology, routing and forwarding performance differences between the more forward exchange smaller and smaller; and prioritization for implementation by actually DiffServ EXP, thus having in this way inevitably has some limitations DiffServ.

<5> MPLS-TE&QoS <5> MPLS-TE & QoS

流量工程(TE: Traffic Engineer)是指为业务流选择路径的处理过程, 以在网络中不同的链路、路由器和交换机之间平衡业务流负载。 Traffic Engineering (TE: Traffic Engineer) refers to the process of selecting a path of traffic flow in different links in the network, traffic balance traffic loads between routers and switches. 其目标是在给一定节点与另一节点之间计算一条路径(源路由),该路径不违反它的约束 The goal is to calculate a route (source route) between a certain node and another node, the path does not violate its constraint

(例如带宽/管理要求),并且从一些数量指标看来是最优的。 (Eg, bandwidth / management requirements), and from a number of quantitative targets appears to be optimal.

MPLS由于自身路由与转发分离的特点,适合与TE的结合,形成MPLS-TE技术。 MPLS routing and forwarding are separated due to its characteristics, for TE and combined to form a MPLS-TE technology. 应用MPLS-TE,可以提高网络的QoS,主要体现在: Application of MPLS-TE, can improve the QoS of the network, mainly reflected in:

(1 )利用MPLS-TE,可以在多条可能的转发路径中进行负栽平衡,从而避免拥塞,提高QoS。 (1) using the MPLS-TE, can be planted negative balance in the plurality of possible forwarding paths in order to avoid congestion, improve QoS.

(2 )应用MPLS-TE,通过RSVP-TE信令创建一条具有严格的QoS带宽保证的隧道,从而支持绝对的QoS。 (2) application of MPLS-TE, has created a strict QoS guaranteed bandwidth through the tunnel RSVP-TE signaling, thus supporting absolute QoS.

(3)可以通过备份LSP、 FRR (快速重路由)等方式对隧道进行额外保护,从而提高网络的QoS。 (3) through a backup LSP, FRR (Fast Reroute), etc. on the tunnel extra protection, thereby improving the QoS of the network.

但MPLS-TE的局限性依然明显,包括:首先它必须应用在MPLS网络中,因此目前部分非MPLS现网无法支持该技术的应用;其次目前对MPLS-TE 跨域的应用仍然在研究阶段,这意味着当前MPLS-TE主要的应用只能在单个域中;另外,MPLS-TE虽然可为用户创建具有带宽保证的隧道,但如果在隧道中同时传送多种业务时,如何对这些不同优先级的业务进行区别处理也是需要研究的问题。 But the limitations of MPLS-TE is still evident, including: First, it must be applied in MPLS networks, there is currently some non-existing MPLS network can not support the application of the technology; secondly the current cross-domain application of MPLS-TE is still at the research stage, this means that the current major application MPLS-TE only a single domain; in addition, MPLS-TE though users can create a tunnel with a bandwidth guarantee, but if the transmission of multiple services at the same time in the tunnel, how these different priority class business conduct issues also need to study the differentiation process.

<6>带宽代理 <6> Bandwidth Broker

为了更有效地监视和控制全网的资源,在新一代的模型中,人们又提出了带宽代理(BB , Bandwidth Broker),也就是网络资源管理器的概念。 In order to more effectively monitor and control the resources of the whole network, the new generation of the model, it also proposed a bandwidth broker (BB, Bandwidth Broker), that is, the concept of network resources manager. 带宽代理收集网络的拓朴和节点及链路状态信息,管理网络资源,并且结合策略服务器规定的策略进行接纳控制。 Collect bandwidth broker network topology and the node and link status information, network resource management, and binding the policy server predetermined admission control policy. 同时,带宽管理还负责管理域之间的通信,通过与相邻网络域的带宽代理通信来达到跨域之间QoS实现的目的。 At the same time, bandwidth management is also responsible for managing communication between domains, through the communication network bandwidth broker domain adjacent to achieve the purpose of QoS achieved between cross-domain.

带宽管理技术的优点是将QoS控制层与数据传输层分离开,核心路由器不需要进行接入控制和QoS状态信息保存,路由器之间也简化了或者说消除了QoS信令,简化了路由器的复杂性。 Advantage is bandwidth management techniques QoS control layer and separated from the data transport layer, core routers and access control need not be stored QoS state information between routers simplified QoS signaling or eliminated, simplifying the complexity of the router sex. 此外这种方式支持绝对的QoS保证, 包括支持跨域的QoS保证。 Also in this way support absolute QoS guarantees, including support for cross-domain QoS guarantee. 还有就是这种方式中网络资源被统一地控制与管理,有利于电信运营商把QoS作为一种业务来开展。 There is in this way are uniformly controlled network resources and management, help telecom operators to QoS as a business to conduct.

但目前带宽管理技术仍然处于研究阶段,BB与业务层和承载层设备之间的信令交互,以及BB之间的信令交互都还在讨论之中。 But the bandwidth management technology is still in the research stage, signaling interaction between the BB and the service layer and bearer layer devices, and signaling interaction between BB are still under discussion. 另外,这种方式对BB的要求很高,在某些情况下,如果同时申请资源的业务流个数很多,有可能会使BB成为网络中的瓶颈。 In addition, this approach requires a high of BB, and in some cases, the number of traffic flow if applying for a lot of resources, it is possible to make the BB become a bottleneck in the network.

4. PWE3 4. PWE3

PWE3原先称为马蒂尼草案(Martini Draft ),其初衷是在MPLS上仿真以太网。 PWE3 original draft called Martini (Martini Draft), its original purpose is a simulation of Ethernet on MPLS. PWE3是一种在分组交换网络(PSN)上模拟各种点到点业务的机制, 被模拟的业务可以是E1专线、ATM、 :FR或以太网等。 Is a simulation of various PWE3 services over the point to point packet switched network (PSN) mechanism, the service may be simulated E1 lines, ATM,: FR or the like Ethernet. PWE3利用PSN上的隧道机制来模拟一种业务的必要属性,这里的隧道称为虚拟线(PW)。 PWE3 using a tunnel mechanism over PSN is necessary to simulate one service attribute, herein referred to as a virtual tunnel line (PW). PWE3 可以对特定服务的协议数据单元(PDU)进行封装,PDU里面含有仿真特定服务所必需的数据和控制信息。 PWE3 protocol data unit may be performed for a particular service (PDU) package, PDU which contains the necessary service-specific simulation data and control information. 使用PWE3机制,运营商可以将所有的传送业务转移到一个融合的网络(如IP/MPLS)之中。 Use PWE3 mechanism, operators can transfer all transmission services to a converged network (such as IP / MPLS) into. 从用户的角度来看,可以认为PWE3模拟的虛拟线是一种专用的链路或电路。 From the user's point of view, it may be considered a virtual simulation PWE3 is a dedicated line or link circuit.

El方面的标准,对PSN上传输E1的框架结构有了完整的定义。 Standard El aspect, the transmission frame structure on PSN has complete definition of E1. 图2是PSN上传输E1的功能示意图。 FIG 2 is a functional schematic diagram of a transmission of E1 PSN.

CE (CustomerEdge,用户边缘设备):它是发起或终结业务的一侧,CE 不会意识到自身采用的是仿真方式而不是原有业务的方式。 CE (CustomerEdge, Customer Edge): It is initiated or terminated business side, CE is not aware of its uses is the manner in emulation mode instead of the existing business.

PE (Provider Edge ,提供商边缘设备):它为CE提供PWE3业务。 PE (Provider Edge, provider edge device): It provides for the PWE3 service CE.

PW (PseudoWire,伪线):是基于分组网络之上的两个PE设备之间的连接。 PW (PseudoWire, PW): PE is a connection between two devices over a packet-based network. PE提供CE和PW之间的适配功能。 PE provides the adaptation function between the CE and PW.

PSN(分组交换网络)隧道:可以传送多个伪线(PW),使其透明通过分组交换网络(PSN)。 PSN of (packet switched network) tunneling: a plurality of dummy wires may be transmitted (PW), via a packet switched network it transparent (PSN).

或多条伪线,使CE之间能通过PSN通信。 Or more dummy lines, so that the communication between the CE through PSN. PSN隧道用来给伪线提供一条数据通道。 PSN tunnel used to provide a pseudo wire data channel. 伪线的业务对于核心网络来说是不可见的,即核心网络对于CE来说是透明的。 Pseudo wire service to the core network is invisible, i.e., the core network is transparent to the CE.

图3则是其协议分层结构表示。 FIG 3 it is represented the protocol hierarchy.

其中封装层为在伪线上传送的指定负荷类型提供必要的结构。 Wherein the encapsulation layer to provide the necessary structure for the specified dummy load type transmission line. PW封装 PW package

层包含三个子层:负荷汇聚、定时、排序。 Layer comprises three sublayers: the convergence load, timing order.

PWE3也对PSN上传输E1对PSN性能的适应能力提出了需求: 包丢失:E1电路边缘到边缘仿真可假定分组网的端到端丢包率较低,特别 PWE3 also the ability to adapt to PSN PSN E1 transmission performance requirements proposed: packet loss: E1 circuit emulation edge to edge may be assumed that the packet network end PLR low, particularly

的,不需要重传机制。 It does not require retransmission mechanism. 为将分组网丟包率的影响降至最小,封装层应具备以下 Effect of the packet network packet loss rate is minimized, the encapsulating layer should have the following

功能: Features:

(1 )在E1的出口,应能独立解析每个数据包封装的E1数据。 (1) at the outlet E1, the E1 data should be independent parse each packet encapsulated. (2)可靠探测丢失的数据包。 (2) reliably detect lost packets.

(3) 将数据包对时钟恢复的影响降至最小。 (3) Effects of the clock recovery packet is minimized.

(4) 增加接口的弹性,可以适当的数据取代丢失的数据包。 (4) increase the flexibility of the interface, data can be appropriately substituted lost packets.

数据包顺序:封装层必须保证携带E1信息的数据包传递的顺序性: Packet sequence: encapsulating layer must ensure that data packets carry sequence information of E1 transmission:

(1) 能够检测出乱序状态。 (1) can be detected scrambled state.

(2) 能够重新顺序化。 (2) can be re-serialization.

拥塞控制:El电路具有恒定的比特率,给PSN带来的负载也是恒定的。 Congestion Control: El circuit having a constant bit rate, to the PSN of the load that is constant. 不需要诸如TCP的流控机制。 You do not need control mechanisms such as TCP streams. 5. ITU-T 8261 5. ITU-T 8261

ITU-T 8261详细的描述了分组网传送E1的同步和定时方面的要求。 ITU-T 8261 is described in detail in timing and synchronization requirements of the packet network transmission of E1. ITU-T8261指出,除了在两端提供同一的参考时钟外,在分组网传送E1 时的时钟恢复主要有两种方案: ITU-T8261 noted that, in addition to the reference clock provided at both ends of the same, when the clock recovery packet network transmission E1 There are two main options:

<1>差分时钟恢复法(Differential Methods )。 <1> differential clock recovery method (Differential Methods).

两端都能够提供参考时钟,并且可以测出本端时钟与参考时钟的偏差, 并将这个偏差传送到远端,远端根据这个差异,结合参考时钟得到E1时钟。 A reference clock can be provided at both ends, and may measure the local clock and the reference clock offset, and transmits the deviation to the distal end, the distal end according to this difference, in conjunction with a reference clock obtained E1 clock. <2>自适应时4中'恢复法(Adaptive Methods ) 两端并不能同时提供参考时钟,业务时钟依据数据包恢复。 <2> 4 when adaptive 'ends recovery method (Adaptive Methods) and at the same time does not provide a reference clock, recover the service clock according to the data packet. ITU-T 8261对E1接口的漂移的规定如下表所示: ITU-T 8261 on an E1 Interface predetermined drift in the following table:

<table>table see original document page 12</column></row> <table> <Table> table see original document page 12 </ column> </ row> <table>

ITU-T 8261指出,分组丢失引起的信息错误对分组网传送E1的影响极大。 ITU-T 8261 noted that the impact of the error packet loss due to packet network transmission E1 great. 原因之一是分组中某一位出错,将导致整个数据包丟失,从而引起E1 信息流的突发性错误。 One reason is that an error in a packet, the packet will cause the entire loss, thereby causing a burst error of the E1 stream. 因此中等程度的丟包率(参考常规分组网的情况)将导致E1电路不可用。 Thus moderate packet loss rate (refer to the case of a conventional packet network) will result in unusable E1 circuits.

6. FEC技术 6. FEC technology

FEC是一种纠错技术,在通信和存储领域有较广泛的应用。 FEC is an error correction technique, there are more widely used in the field of communications and storage.

FEC的在通信应用的基本原理是:对于要传输A比特的信息位数据,根据某种编码方法对其编码,生成比特的数据,将&比特的信息位数据和生成的比特的冗余位一起发送,接收方可以利用冗余位恢复传输中丢失的信息位数据;接收方将接收到的/t'(f^:)比特的数据,运用相应译码方法恢复A比特的信息位数据。 The basic principles of the FEC in communications applications are: for bits of information to be transmitted A-bit data, encodes it according to a certain encoding method, data bits generated, the bit information & bits of data and redundant bits generated together transmission, the receiver can recover the information bits using a redundancy bit of lost data transmission; receiving party receives the / t '(f ^ :) bits of data, using the decoding method corresponding to the information bit a bit recovery data.

FEC技术在存储领域的应用的基本原理可表述为:在数据在被存储到存储介质之前,预先按一定的算法进行编码处理,加入带有数据本身特征的冗余码,在从存储介质读取数据时,按照相应算法对读出的数据进行解码,从而找出存储介质产生的错误码并将其纠正。 The basic principles of the FEC in the application storage field can be expressed as: pre-coding algorithm according to a certain process before being stored in the data storage medium, the data added with the redundant code features themselves, read from a storage medium data, decodes the data read out in accordance with the appropriate algorithm to find the error code generated in the storage medium and corrected.

在视频处理领域也采用了基于分组的FEC技术。 In the field of video processing technology also uses FEC packet-based. 7.分组网传送E1的不足分组网传送E1已有实际的应用。 7. E1 has practical application packet network packet network transmission E1 lack of transfer.

分组网,不管是IP还是以太网,都是基于统计复用的异步的无连接的网络,对QOS (Quality Of Service,服务质量)的支持相对较弱。 Packet network, whether IP or Ethernet networks are based on statistical multiplexing asynchronous connectionless support for QOS (Quality Of Service, Quality of Service) is relatively weak.

上述的各种方案,大部分并不能完全实现QOS保证。 The above-mentioned programs, the majority does not fully implement QOS guarantees. 采用IntServ或MPLS 技术可以最大程度的保证QOS,但并非现有网络设备所能提供的,而且实现起来复杂性和成本都较高。 Using IntServ or MPLS technology can ensure the greatest degree QOS, but not existing network devices can provide, but complex to implement and costs are higher. 同时,无论哪种方案,都是需要分组网络中的所有网 At the same time, no matter what program, it is required for all network packet network

现有网络设备并不能都支持这些方案中的一种。 Existing network equipment does not support one of these solutions. 这使得在现有分组网络上传输E1的QOS并不能采用这些方案来保障。 This makes the existing packet transmission network QOS E1 and can not be used to protect these programs.

因此分组网的QOS能力与基于同步的E1对QOS的高要求的矛盾非常突出。 Therefore, the packet network QOS capabilities and E1-based synchronization of conflicting requirements of high QOS is very prominent. 在用分组网络传送E1时,网络拥塞造成的分组丢失或时延超过门限都会在接收端表现为信息的丢失,这严重地影响了El的质量。 When a packet network transmission E1, packet loss due to network congestion or latency exceeds the threshold will be at the receiving end performance for the loss of information, which seriously affected the quality of El.

现有的分组网传送E1的技术和产品,多依赖分组网来提供QOS的保证。 E1 conventional packet network transmission technology and products, more dependent on the packet network to provide a guaranteed QOS. 在分组网络不能保障其要求的性能而导致信息丢失时,多以全"1 "取代丢失的数据,或用收到的前一个帧来替代丢失的帧。 When it can not guarantee the performance requirements of the network packet loss of information, and more to all "1" substituted lost data or received with a previous frame instead of the lost frame. 这显然无法保证E1的质量。 This obviously can not guarantee the quality of E1.

虽然并未有在分组网传送E1应用FEC的先例,但由理论分析可知,采用FEC机制在现有分组网络上传送El,能够在性能获得极大的改善。 Although there is not packet network transmission E1 applications FEC precedent, but by the theoretical analysis shows that the use of FEC mechanisms of El transmitted over existing packet networks, you can get great improvement in performance. 发明内容 SUMMARY

TDM对QOS的要求非常高,而现有分组网技术的QOS体系尚不完善, 不能提供QOS保证。 TDM QOS requirements are very high, and QOS system existing packet network technology is not perfect, it does not provide QOS guarantee.

PWE3任务组提出的标准和草案也并没有规定和要求分组网提供QOS保证。 PWE3 standards and drafts proposed by the Task Force also did not provide provisions and requirements of packet network QOS guarantees.

正是由于现有的分组网传输TDM的技术和产品需依赖分组网络来提供QOS保证,而现有分组网络中的节点设备并不能提供QOS保障。 Because of the conventional TDM transmission packet network technology and products need to rely on packet network to provide guaranteed QOS, the existing packet network node apparatus does not provide QOS guarantee. 所以分组网传输E1需要结合其它技术和机制,使得其不必依赖分组网络的QOS体系, 从而拓展自身的应用领域。 Therefore, the packet network transmission of E1 in combination with other techniques and mechanisms needed, such that it does not have to rely on the system QOS packet network, thereby expand its applications.

也正是为了解决分组网传输El等TDM业务过度依赖分组网络的性能的问题,本发明提出了一种釆用FEC机制在分组网上可靠传送端到端El的方法。 In order to solve the packet network is also transmitted El TDM services over packet networks dependence of the properties, the present invention provides a method of internet packets with FEC mechanism in the El preclude reliable delivery end. 其目的是:在网络QOS较差的情况下,仍能通过FEC机制保障El在分组网上传输的端到端性能。 Its purpose is: in the case of poor network QOS, can still end performance by El FEC mechanisms to protect transmission over packet networks.

本发明的核心内容可以用下面三个步骤描述: The core of the present invention may be described by the following three steps:

a) 在源结点(E1进入分组网的一端,也可称为发送结点,区分源节点和目的节点只为叙述方便,仅相对单向传送而言,源节点和目的节点都同时具有发送/接收功能)将El按照特定的规则和协议封装成一个数据包序列,并根据这个数据包序列采用特定的编码规则和协议生成一个FEC数据包组。 a) at the source node (E1 into the end of the packet network, and may also be referred to as the sending node, the source node and the destination node to distinguish only for convenience of description, only a relatively one-way transmission, the source node and the destination node are transmitted simultaneously having / receiving function) will be packaged in accordance with specific rules El and protocol into a sequence of packets, and generates an FEC packet groups according to the sequence of packets using a specific coding rule and protocol.

b) 将这个数据包序列和FEC数据包组通过分组网发送到目的节点(E1离开分组网的一端,也可称为接收结点)。 b) that the data packet sequence and the FEC data packet group to the destination node through the packet network (E1 leaving the end points of the network, may also be referred to as a receiving node).

c) 在所述目的节点根据FEC包組和数据包序列,如存在数据包丢失,则依据特定的规则再生丟失的数据包,提取或再生El净载荷并还原El码流。 c) in the destination node according to the FEC packet group and the data packet sequence, such as the presence of packet loss, depending on the particular rules of the lost packets reproduced, reproducing the extracted El or El and reduced payload stream.

本发明能够兼容不含FEC机制的同类技术,也即,具有FEC机制的源结点可以和不具有FEC机制的目的节点互通互联。 The present invention is free of technical grade compatible FEC mechanism, i.e., the source node having a FEC mechanism can not have a destination node interconnection FEC mechanism.

本发明以自协商的方式确定目的节点是否具备相同的FEC机制。 The present invention is determined in a self-negotiation whether they have the same destination node FEC mechanism. 如目的节点不具备相同的FEC机制,则将本端设置成非FEC工作模式。 The destination node does not have the same FEC mechanism, then ends the present operation mode set to the non FEC.

本发明分组网的封装层采用了RTP协议,用作定时(特别是差分时钟恢复方式)恢复。 Encapsulating the packet network of the present invention uses the RTP protocol layer, used as a timing (in particular the differential mode clock recovery) recovery. 同时也与IETF PWE3任务组制定的标准相符合。 But also consistent with the IETF PWE3 task group to develop standards. 在不依赖RTP 实现定时恢复的应用,此协议可作为保留功能。 RTP realized without depending on the application timing recovery, this protocol can be used as retention function.

当配置成差分时钟恢复方式时,RTP协议的时戳字段携带本地时钟与参考时钟的差值信息。 When configured as a differential clock recovery mode, when the RTP timestamp field difference information carries the local clock and the reference clock. 当配置成自适应时钟恢复方式时,RTP协议的时戳字IS:携带依据本地业 When configured as an adaptive clock recovery mode, when the RTP timestamp word IS: based on local operators carrying

务时钟生成的时戳信息。 When the clock of time stamp information service.

本发明分组网的封装层的控制字的FORMID采用扩充定义,并用以标识FEC机制使用的规则。 Control word packet network encapsulation layer of the present invention uses extension FORMID defined, and the rules for identifying the FEC mechanism used. 本发明利用分组网的封装层的控制字(TDMoIP)中最左边的保留字段(共四个比特)定义以标识FEC机制采用的规则。 The present invention utilizes a control word packet network encapsulation layer (of TDMoIP) leftmost reserved field (of four bits) define rules to identify the FEC mechanism employed. 本发明将该保留字段重新定义后命名为FEC类型。 After the reserved field of the present invention redefines named FEC type.

本发明分组网的封装层的控制字的长度字段不再用于表示长度信息,而是用以标识本地测量的远端到本地单向端到端丟包率,并将次信息通过数据包反馈回远端。 Control word packet network encapsulation layer of the present invention is not the length field for indicating length information, but rather are used to identify the local measurements to the distal end of the one-way local end packet loss rate, and the feedback time information packet back to the far end. 本发明将长度字段命名为长度/性能。 The present invention will be named length field length / performance.

本发明分组网的封装层的控制字的序列号定义成一个二维结构,主序列号和次序列号两级。 Packet network of the present invention, the control word encapsulation layer sequence number is defined as a two-dimensional structure, primary and secondary sequence number SEQ ID NO levels. 主序列号与时隙组对应,用以标识一个数据包组,次序列号则用以区分数据包组内的各个数据包。 Main group sequence number corresponding to the time slot for identifying a group of data packets, the secondary sequence number is used to distinguish each packet in the packet group. 其中序列号lk为时隙组使用, 序列号kn为FEC组使用。 Wherein the sequence number is a slot group used lk, kn sequence number used for the FEC group.

本发明的分组交换网层采用IP协议和帧结构。 Packet-switched network layer of the present invention uses the IP protocol and frame structure.

本发明的分组网的数据链路层采用以太网链路层协议和帧结构。 The packet network of the present invention, data link layer and the Ethernet link layer protocol frame structure.

本发明所述a)步骤还进一步包括以下子步骤: The present invention step a) further comprises the substeps of:

本发明设计一个32位业务时钟计数器对本地业务时钟计数。 The present invention is designed to service a 32-bit clock counter for counting a clock local traffic. 同时设计一个32位参考时钟计数器对外部参考时钟计数。 At the same time a 32-bit design of the external reference clock counter counts the reference clock.

映射数据包时,将E1码流以按时隙分组。 When mapping data packet, the E1 code stream packets to time slots. 当一个完整的时隙进入緩沖区后,记录业务时钟计数器和参考时钟计数器的值以用于生成本时隙的时戳。 When a full slot into the buffer, recording the value of the service clock and the reference clock counter to timestamp the time counter used to produce the slots.

映射数据包时,时隙组映射生成的所有的数据包中RTP字段需携带确定的时戳信息。 When mapping data packets, all packets generated slot group map for an RTP timestamp field is determined that the information carried.

如果配置成自适应时钟模式,则时戳信息依据记录的业务时钟计数值生成。 If the count value of the service clock configured to adaptively clock mode, the timestamp information is generated based on record.

如果配置成差分时钟恢复方式,则时戳信息依据记录的业务时钟计数值和外部参考时钟计数值,并根据一定规则生成。 If configured as a differential clock recovery mode, the timestamp information service based on the reference clock count value and the external clock count value recorded, and generated according to a certain rule.

将一组E1时隙(称为El时隙组)集中映射成一组数据包。 A group E1 timeslot (timeslot group called El) mapped to a centralized group of data packets. 其中El时隙组中包含的E1时隙的个数可配置成8, 16或32。 Wherein E1 timeslot number of time slots contained in the group El may be configured to 8, 16 or 32. 之所以选择2Am作为时隙组的成员个数,是因为采用这样的值,设计緩冲区结构、相应处理以及生成FEC纠错包都比较方便。 2Am chosen as the slot number of the member of the group, because such a value, the buffer structure design, the FEC and generating corresponding treatment are more convenient package.

将E1时隙组集中映射成分组数据包时,可以配置使用每个时隙独立映射 When the focus mapping E1 timeslot group component group of data packets, each slot can be configured to use independent mapping

成包的独立映射方式,也可选择从E1时隙组中的每个时隙选取部分数据,再封装成数据包的合成映射方式。 Independent mapping mode to the packet portion also choose to select data from the E1 time slots each time slot in the group, and then packaged into a composite data packet mapping method.

独立映射方式处理较为简单,数据包之间相互牵连比较少,但信息安全性较差,而分组网并不能提供信息要全的完善保证。 Independent mapping relatively simple manner, each data packet between less involved, but the information security is poor, and the sub-network does not provide the full information to guarantee perfect.

合成映射方式正好相反,可以打乱E1净荷信息的排列,相当于对信息进行了加密,有助于E1信息在分组网上安全的传输(相对于E1网络来说,分组网的安全性较差)。 Synthesis opposite mapping mode, the arrangement can be disrupted E1 payload information, the information corresponding to the encrypted information helps E1 E1 networks (with respect to the safety of the packet network transmission of a packet network inferior in safety ). 但其处理较为复杂,并且那样的话, 一个E1时隙组包含的数据耦合非常紧密,要么全部错误,要么完全正确。 But the process is complicated, and in that case, a data coupling E1 timeslot group contains very close, either all error, or correct.

将一组E1时隙集中映射成分组数据包时,如果使用合成映射方式,则从每个时隙的每一字节选择对应的位拼成字节或字,再组成帧。 When the focus mapping a set of component E1 timeslot group of data packets, if the synthetic mapping mode, each slot from each byte selector corresponding to the byte or word bit makes up, and then the composition of the frame. 以时隙组包含8个时隙为例,用b(i, j, k)表示第i个时隙的第j个字节的第k位,则映射成的第i个数据包的净信息的第j个字节(用B(i,j)表示)可以表示成b( 1, j, i) b (2, j, i) b (3, j, i) b (4, j, i) b (5, j, i) b (6, j, i) b (7, j, i) b (8, j, i)。 Net information slot group contains eight time slots, for example, by B (i, j, k) represents the k-th bit of the j-th byte of the i-th time slot is mapped to the i-th data packet the j-th byte (represented by B (i, j)) can be expressed as b (1, j, i) b (2, j, i) b (3, j, i) b (4, j, i ) b (5, j, i) b (6, j, i) b (7, j, i) b (8, j, i).

将一个El时隙组集中映射成数据包后,根据生成的分组数据包序列生成FEC数据包组。 After a focus group El slot mapped into packets, packet groups according to generate FEC data packet sequence of packets generated. FEC数据包组的数据依据时隙组信息生成,数据包头的生成过程与普通数据包类似。 FEC data group of data packets based on information generation slot group, the packet header generation process is similar to the normal data packets.

生成FEC包组,采用Reed-Solomon ( n, k)编码。 Generating a set of FEC packets using Reed-Solomon (n, k) code. 该编码体系为非二进制编码体系,以数据块为单位实现检错和纠错,n是数据块的长度,k是数据块包含的净信息长度,nk是纠错码长度。 The non-binary coding system encoding scheme, data block units to achieve error detection and correction, n being the length of the block, k is the net length of the block contains information, nk is the length of the error correction code.

生成FEC包组,采用Reed-Solomon (n, k)编码时,k可配置成8, 16 或32, n的取值可有k + 4, k+8,k+16,k+32等选项。 Generating a set of FEC packets using Reed-Solomon (n, k) encoding, k may be configured to 8, 16 or 32, n can have the value k + 4, k + 8, k + 16, k + 32 and other options . 这些选项可根据远端反馈的性能信息自动配置。 These options can be automatically configured in accordance with the distal end of the performance information feedback.

生成FEC包组的步骤是: The step of generating FEC packets are set:

将k个时隙(也就是说一个时隙组包含k个时隙)中的每个时隙(包含32 字节的信息)的第i字节组合在一起得到k字节序列,根据这个k字节序列,并结合生成多项式生成nk字节的纠错码,这nk字节分别作为第1,2...nk个FEC包的第i字节。 The k time slots (i.e. a slot group contains slots k) in each slot (comprising 32 bytes of information) of the i-th byte of the byte sequence together to give k, k based on this byte sequence, in combination with a generator polynomial to generate the byte error correcting code nk, nk which each byte as the first byte of i-th ... nk 1,2 FEC packet.

本发明的b)步骤还进一步包括以下子步骤: b) step of the present invention further comprises the substeps of:

发送数据包时,普通数据包的发送时间取决于数据包(或其携带的时隙) 对应的业务时钟计数值,也就是i兌,时隙組内两个相邻凄t据包的发送时间间 When transmitting packets, the transmission time of normal data packets depends on the packet (or carrying slots) corresponding to the service clock count value, i.e. against i, the timeslot group t sad two adjacent packet data transmission time between

隔与两个相邻业务时钟计数值的差值成比例。 Two compartments and proportional to the difference clock count value of the adjacent traffic.

发送数据包时,在两个普通数据包的发送间隔内,发送FEC数据包。 When transmitting packets, the transmission interval in two ordinary data packets, and transmits the FEC data packet.

每个时间间隔发送的FEC数据包个数,根据n和k的取值确定。 FEC number of packets sent per time interval is determined according to the values ​​of n and k. 发送FEC Send FEC

数据包的时间确定原则为:尽量使相邻数据包发送的时间间隔均匀。 Determining time data packet principle: try to make the time interval of adjacent data packets transmitted uniformly. 本发明的c)步骤还进一步包括以下子步骤: 在接收端为控制时延设置接收定时器,溢出周期等于k*125us。 c of the present invention) step further comprises the substeps of: receiving end disposed to receive control delay timer overflow period is equal to k * 125us. 将接收到的数据包去掉帧头后,根据数据包携带的包序号,将得到的净 Net received data packet header is removed, based on the packet sequence number carried in the data packet, and the resulting

数据信息存入去抖緩沖区的相应位置。 Shake information into data corresponding to the buffer.

如果配置成基于时戳的自适应恢复模式,则需要将数据包携带的RTP协 If the adaptive timestamp recovery mode is configured based on the need to carry an RTP packet Co

议中的时戳信息存入缓冲区的远端时戳存储空间。 Stamp storage space when the proposed stamp information into the distal end of the buffer. 同时利用本地业务时钟生 Clock generation while taking advantage of local service

成时戳,并将这个时戳存入緩冲区的本地时戳存储空间。 When to poke, poke and poke into the local buffer memory space of this time.

如果配置成差分时钟恢复模式,则以这个时戳结合外部时钟得到本地业 If configured as a differential clock recovery mode, places the external clock timestamp combined to give local operators

务时钟。 Service clock.

如果配置成基于时戳的自适应恢复模式,且本时隙组的时戳信息尚未存储,则需要将RTP协议中的时戳信息存入时戳緩沖区。 If the adaptive timestamp recovery mode, and the present set of time slots has not been stored timestamp information configured based on the needs of the RTP protocol time stamp information into the time stamp buffers.

将接收到的FEC包去掉包头后,根据数据包携带的包序号,将得到的净纠错码信息存入纠错緩冲区的相应位置。 After the received FEC packet header is removed, based on the packet sequence number carried in the data packet, and the resulting net position error correction code corresponding to the error correction information into the buffer.

当接收定时器溢出时,检查去抖緩沖区的数据,如存在信息丢失,则利用纠错緩冲区的信息,结合去抖緩冲区的数据,再生丢失的数据包携带的净信息。 Upon receiving the timer overflow check debounce buffer data, information is lost, the error correction information using the buffer, the buffer debounce binding data, reproducing lost packets carrying information such as the presence of a net.

当接收定时器溢出时,并且丢失信息恢复完成后,需要恢复E1的业务时钟。 When receiving the timer overflow, and information is lost after the recovery is completed, the need to restore the E1 service clock. 可以根据是否具有参考时钟,配置E1时钟的恢复方法为差异方法和自适应方法。 Depending on whether the reference clock, the configuration of the clock recovery method as a difference E1 and adaptive methods.

差异方法是利用RTP协议携带的时戳,结合外部参考时钟生成本地业务时钟。 Difference method is carried using the RTP protocol time stamp, in conjunction with an external reference clock generating a local clock business.

采用自适应方法,是根据去抖緩冲区的使用情况或时戳信息,采用反馈调节的方式恢复时钟。 Adaptive method, according to the embodiment is a clock recovery debounce time stamp information of buffer usage or using feedback regulation.

采用自适应方法,可以根据需要选用基于去抖緩冲区填充级和基于RTP 时戳两种模式。 Adaptive method, based on the buffer fill-level debounce and RTP timestamp based modes according to the welding.

采用自适应方法,可以配置恢复的业务时钟的精度范围。 Adaptive method may be configured to restore the accuracy range of the service clock.

如果配置成自适应方式恢复时钟时,且配置成使用緩冲填充级时,则根 If the configuration when an adaptive clock recovery mode, and configured to use the buffer filling level, the root

据緩沖区的使用情况恢复本地业务时钟。 According to buffer usage, the local service clock recovery.

如果配置成自适应方式恢复时钟时,且配置成使用时戳时,则根据緩冲区存储的时戳信息恢复本地业务时钟。 If the recovered clock is configured to adaptively and configured time stamp, the time stamp information according to the recovery clock when the local service time to use the buffer memory.

然后,使用恢复的业务时钟,并结合缓沖区的数据恢复E1码流。 Then, using the service clock recovery and data recovery E1 binding buffer stream.

靠的端到端E1传送的设备。 By the end E1 transmission equipment. 附图说明 BRIEF DESCRIPTION

图1表示本发明的主要功能流程(描述单方向的El业务传送)。 Figure 1 shows the main functions of the process of the present invention (described in El traffic transmission in one direction).

图2是正TFPWE3任务组关于边缘到边缘的电路仿真的功能示意。 FIG 2 is a positive TFPWE3 Task Force on edge to the edge of the circuit simulation of a schematic function. 图3是正TF PWE3任务组关于边缘到边缘的电路仿真的协议分层结构。 3 is a front task group TF PWE3 circuit emulation protocol hierarchy on the edge to edge. 图4是IETF PWE3任务组关于边缘到边缘的电路仿真PSN层IP协议之上的协议结构,也是本发明采用的协议结构。 FIG 4 is a task group IETF PWE3 edge to edge on top of the circuit emulation protocol structure PSN layer IP protocol, the protocol structure of the present invention is employed.

图5是正TF PWE3有关TDMoIP的标准和草案定义的控制字的图示。 FIG 5 is a diagram TF PWE3 positive draft standards and definitions of TDMoIP control word.

图6是本发明定义的控制字的图示。 FIG 6 is an illustration of the control word defined in the present invention.

图7是本发明使用的协议的分层结构图示。 FIG 7 is a layered protocol structure of the present invention illustrating use.

图8是本发明使用的协议的帧结构图示。 FIG 8 is a frame structure illustrating the protocol of the present invention.

图9是发明的设备的功能结构图。 FIG 9 is a functional block diagram of the apparatus of the present invention.

图IO本发明的采用FEC机制传送E1的示意。 FIG IO E1 transmission using FEC mechanism according to the invention schematically.

图ll本发明的功能模块详解。 FIG ll Detailed functional module of the invention.

图12本发明的FEC包生成的图示。 Generating FEC packets invention 12 illustrated in FIG.

图13表示利用FEC机制纠正错误对物理信道性能的改善的情况。 13 shows a case of using the error correcting FEC mechanism to improve the physical properties of the channel. 图14本发明的编码模块的流程图。 Flowchart of an encoding module 14 of the present invention FIG. 图15本发明的封包模块的流程图。 Packet module of the present invention in a flowchart in FIG. 15. 图16本发明的解包模块的流程图。 Decapsulating module flowchart of Figure 16 of the present invention. 图17本发明的解码模块的流程图。 Flowchart decoding module 17 of the present invention of FIG.

图18本发明的基于緩沖区填充级模式的自适应时钟恢复模块的流程图。 FIG flowchart block buffer fill level mode based on an adaptive clock recovery 18 invention.

图19本发明的去抖緩沖区的示意图.。 FIG invention dejitter buffer 19 is a schematic view ..

图20本发明的基于RTP时戳模式的自适应时钟恢复模块的流程图。 Based on the flow diagram block mode of the RTP timestamp of the present invention in FIG. 20 adaptive clock recovery. 图21发明的设备的正面图。 FIG 21 is a front view of the apparatus of the present invention.

图22发明的设备的背面图。 Rear view of the device of the invention in FIG 22.

图23本发明的性能评估和测试网络拓朴。 FIG performance evaluation test and network topology 23 invention. 具体实施方式 detailed description

本发明的核心内容用FPGA实现,并在发明的设备中应用。 The core of the present invention is implemented in an FPGA, and use the device in the invention.

图l表示本发明的主要功能流程。 Figure l showing the main functions of the process of the present invention.

101表示发送端将E1映射到緩冲区的过程。 101 denotes a transmission process to end E1 mapped buffer.

102表示发送端将数据封装成数据包的过程。 102 shows a data packet encapsulated into the data transmission process ends.

103表示发送端FEC数据包生成的过程。 Process 103 represents FEC packet generation sender.

104表示接收端收到的数据存入緩沖区的过程。 Process 104 represents receiving terminal stores the received data buffers.

105表示接收端利用FEC机制再生丟失的数据包的过程。 105 represents a receiving side FEC mechanism reproducing process using the lost data packets.

106表示接收端FEC机制处理后,利用数据还原El码流的过程。 106 denotes the receiving end processing FEC mechanism, using data reduction process stream El.

107表示分组网。 107 denotes a packet network.

实施的步骤分别为: Step were:

I) 确定采用的帧格式。 I) determining the frame format used.

2 )确定FEC机制使用的编码规则。 2) determining an encoding rule used for FEC mechanism.

3) 时戳緩冲区设计 3) buffers stamp design

4) 发送緩冲区设计。 4) transmit buffer design.

5) 去抖緩冲区设计。 5) to shake buffer design.

6) 纠错缓冲区设计。 6) Correction buffer design.

7) 接收定时器设计。 7) receiving timer design.

8) 封包模块设计。 8) Design packet module.

9) 解包模块设计。 9) decapsulating module design.

10) 编码模块设计。 10) encoding module design.

II) 解码模块设计。 II) decoding module design.

12) 时钟恢复模块设计等。 12) a clock recovery module design.

13) 对端类型确定(是否具备FEC机制)。 13) determines the type of the peer (if FEC mechanism is provided). 以下是各个步骤的详细内容 The following are the details of the individual steps

1)确定采用的帧格式。 1) determining a frame format used.

图3描述了正TFPWE3任务组制定的标准说明的帧结构,本发明以此帧 3 depicts a frame structure of the positive TFPWE3 task force developed standards described, the present invention this frame

结构为模板。 Structure as a template. 但由于要采用FEC机制,并且发送端应能够通过帧头信息通告 However, since the FEC mechanism to be used, and the sender shall be able to advertise through the header of the information

接收端FEC机制的配置,以便接收端正确的解码,所以帧结构并不能和标准说明的完全一样。 FEC mechanism receiving end disposed to receive the correct decoder, the frame structure and the standard instructions are not exactly the same.

同时为互联互通性考虑,本发明的帧结构与IETF PWE3任务组制定的标准和草案兼容。 At the same time as interoperability considerations, the frame structure of the invention with the IETF PWE3 task group developed a draft standard and compatible.

为符合标准和兼容性考虑,本发明并不改变前述帧格式的主要框架,主要有以下几个方面的改动: Is a standard and compatibility reasons, the present invention does not change the frame format of the main frame, the main changes in the following aspects:

(1) 用正TFPWE3关于TDMoIP控制字中的最左边四比特的保留字段重新定义FEC类型字段,FEC类型表示FEC机制的配置信息。 (1) to redefine the FEC type field with n TFPWE3 TDMoIP reserved field in the control word on the leftmost four bits, FEC type indicates configuration information FEC mechanism.

(2) 用IETF PWE3关于TDMoIP控制字中的第8、 9两个比特的保留字段重新定义FEC使能字段,FEC使能表示是否禁用本发明的FEC机制。 (2) with respect to IETF PWE3 8, two 9-bit reserved field in the control word TDMoIP redefine FEC enable field, FEC FEC mechanism can enable the present invention indicates whether to disable.

(3) 将16位序列号分解为一个二维的结构:高10位作为主序列号,主序列号标识数据包组,低6位作为次序列号,表示数据包组内的每个数据包顺序。 (3) A 16-bit sequence number is decomposed into a two-dimensional structure: 10-bits of the sequence number as the master, the master serial number identification packet group, the lower 6 bits as the secondary sequence number, represent each data packet in the packet group order.

(4) 长度域重新定义为长度/性能级字段,用来表示网络的单向端到端性能。 (4) the length of the length field is redefined / performance level field used to indicate the one-way end-to-network performance.

原协议结构如图5所示,各字段分别为: Original protocol structure shown in Figure 5, the fields are:

RES(最左边四位),保留字段。 RES (leftmost four), reserved field. 当指示是MPLS网络时,该四位应该设置成全0。 When the indication is an MPLS network, the setting should fulfill four 0.

L字段,l位,指示本地失效。 L field, l bits, indicating local failure.

R字段,l位,指示远端失效。 R field, l bits, indicating the end failure.

M字段,2位,指示L字段的含义,当前未定义。 M field, 2, L indicates the meaning of fields is currently undefined.

RES (第8位和第9位),保留字段,应设置为全0。 RES (No. 8 and 9), reserved field, should be set to all zeros.

长度域,6位,指示包含控制字在内的净负荷的长度。 Length, 6, indicating the length of the payload contains control words including the.

序列号,用于凝:据包的定序。 Serial number for coagulation: sequencing data packets.

本发明定义的协议结构如图6所示。 Protocol structure defined in the present invention is shown in Fig. 各字段分别为: The fields are as follows:

FEC类型,4位。 FEC type, four. 具体定义参见下面的描述。 Specific definitions - see below.

L字段,l位,与原定义相同。 L field, l bits, defined the same as the original.

R字段,l位,与原定义相同。 R field, l bits, defined the same as the original.

M字段,2位,指示L字段的含义,当前未定义。 M field, 2, L indicates the meaning of fields is currently undefined.

FEC机制使能(第8位和第9位),设置为11表示本发明的FEC机制 FEC mechanism is enabled (bit 8 and 9), is set to 11 showing the FEC mechanism according to the present invention

使能,设置为00时表示,为与远端非FEC机制的网络设备互通,禁用本发明的FEC机制。 Enabled, is set to 00, the distal end of the device communicate with a network of non-FEC mechanism, FEC mechanism according to the present invention is disabled.

性能级,6位,表示网络的单向端到端性能。 Level, 6, showing end-way network performance. 共有64个值,具体定义如 A total of 64 values, defined as the specific

下: under:

0…单向无丟包。 0 ... no one-way packet loss.

I— 单向丢包率小于或等于1%。 I- unidirectional packet loss rate less than or equal to 1%.

2…单向丢包率大于1%。 2 ... one-way packet loss rate is greater than 1%. ,但小于或等于5%。 But less than or equal to 5%. 3…单向丢包率大于5%。 3 ... one-way packet loss rate is greater than 5%. ,但小于或等于10%。 But less than or equal to 10%.

4— 单向丢包率大于10%。 4- one-way packet loss greater than 10%. ,但小于或等于15%。 But less than or equal to 15%.

5— 单向丢包率大于15%。 5- one-way packet loss greater than 15%. ,但小于或等于20%。 But less than or equal to 20%.

6— 单向丢包率大于20。 6--way loss greater than 20. /。 /. . ,但小于或等于25%。 But less than or equal to 25%.

7— 单向丢包率大于25。 7- one-way packet loss rate is greater than 25. /。 /. . ,但小于或等于30%。 But less than or equal to 30%.

8— —单向丟包率大于30%。 8--- way loss greater than 30%. ,但小于或等于35%。 But less than or equal to 35%. 9…单向丢包率大于35%。 9 ... 35% greater than the one-way packet loss. ,但小于或等于40%。 But less than or equal to 40%. 10…单向丢包率大于40%。 10 ... one-way loss greater than 40%. ,但小于或等于45%。 But less than or equal to 45%.

II— 单向丢包率大于45%。 II- way packet loss rate is greater than 45%. ,但小于或等于50%。 But less than or equal to 50%.

12— 单向丟包率大于50%。 12- way loss greater than 50%. ,但小于或等于55%。 But less than or equal to 55%.

13— 单向丢包率大于55%。 13- 55% greater than the one-way packet loss. ,但小于或等于60%。 But less than or equal to 60%.

14— 单向丢包率大于60%。 14- 60% greater than the one-way packet loss. ,但小于或等于65%。 But less than or equal to 65%. 15…单向丢包率大于65%。 15 ... one-way loss greater than 65%. ,但小于或等于70%。 But less than or equal to 70%.

16— 单向丟包率大于70%。 16- way loss greater than 70%. ,但小于或等于75%。 But less than or equal to 75%.

17— 单向丢包率大于75%。 17- 75% greater than the one-way packet loss. ,但小于或等于80%。 But less than or equal to 80%. 18…单向丢包率大于80%。 18 ... one-way loss greater than 80%. ,但小于或等于85%。 But less than or equal to 85%. 19—单向丢包率大于85%。 19- 85% greater than the one-way packet loss. ,但小于或等于90%。 But less than or equal to 90%. 20…单向丟包率大于90%。 20 ... one-way loss greater than 90%. ,但小于或等于95%。 But less than or equal to 95%.

21— 单向丟包率大于95%。 21 one-way loss greater than 95%. ,但小于或等于100%。 But less than or equal to 100%.

22- 63—保留。 22-63- reserved.

主序列号,10位,表示数据包组的组号。 Main sequence number, 10, represents the group number of the group of data packets.

次序列号,6位,表示每个数据包在数据包组内的顺序。 The secondary sequence number, 6, indicates the order of each packet in the group of data packets.

图7是本发明采用完整的帧结构。 FIG 7 is complete the present invention employs a frame structure.

701是以太网帧头,共14字节。 701 is an Ethernet frame header of 14 bytes.

702是IP头,共20字节。 702 is the IP header of 20 bytes.

703是UDP头,为8字节。 703 is a UDP header is 8 bytes. 704是RTP协议,共12字节。 RTP is a protocol 704, a total of 12 bytes. 705是携带FEC机制的控制字,为4字节。 705 is a control word carries FEC mechanism, it is 4 bytes. 706是TDM净负荷,共32字节。 706 is a TDM payload of 32 bytes.

2)确定FEC机制使用的编码规则。 2) determining an encoding rule used for FEC mechanism.

FEC常常用于纠正物理通信和存储设备中的误码,应用非常广泛。 FEC physical communication and are often used to correct errors in the storage device, it is widely used. 传输网的主流技术SDH系统也采用了FEC机制。 Mainstream technology SDH transmission network system also uses FEC mechanism.

但查新并未发现分组网传送E1采用FEC机制的先例。 But the new investigation found no precedent for the use of FEC packet network transmission E1 mechanism. 这里采用RS (n, k)编码体系。 Used here RS (n, k) code system.

理论分析表明,采用RS (n, k)编码,海明(Hamming)距离为 Theoretical analysis shows that, using the RS (n, k) coding, Hamming (the Hamming) distance

<formula>formula see original document page 22</formula> <Formula> formula see original document page 22 </ formula>

则可以纠正的错误数至多为: Errors can be corrected up to:

<formula>formula see original document page 22</formula> <Formula> formula see original document page 22 </ formula>

而删除纠错能力为: The error correction capability to delete:

本发明使用的生成多项式为: <formula>formula see original document page 22</formula> A generator polynomial used in the present invention is: <formula> formula see original document page 22 </ formula>

恢复分组网传送E1的丟包,也就是通过FEC再生丢失的数据包的每一个字节。 Recovery E1 transmission packet network packet loss, i.e. each byte FEC regenerated by lost data packets. 通过序列号可以判断丢失的分组的位置,也就可以确定每个数据块出错的位置,所以它实质是一种删除纠错能力。 Can be judged by the sequence number of packets lost position, it can determine the location of the error of each data block, so that it is essentially a deleted error correction capability. 因此,采用RS(n, k)编码,当每一组数据包丢失的个数小于(nk)时,丢失的数据包都可用解码器完全正确的生成。 Therefore, RS (n, k) encoding, each set of data when a packet loss is less than the number (NK), lost packets are generated can be used exactly the correct decoder.

本发明的k的取值有8, 16, 32三个选项。 k has the value of the present invention, 8, 16, 32 three options.

当k4时,n可以选用12, 16, 24, 32等四个值。 When k4, n is 12 can be selected, 16, the four values ​​24, 32 and the like.

当k46时,n可以选用20, 24, 28, 32等四个值。 When k46, n is selected can be 20, 24, four values ​​28, 32 and the like.

当1^=32时,n可以选用40, 48 , 56, 64等四个值。 When 1 ^ = 32, n can use 40, 48, 56, 64 four values.

k的取值可由软件配置,n的取值则有反馈协商机制确定,也就是说,当远端测出的丢包率较低时,n的取值可设置成较小的数,反之则可设置成较大的数。 software configurable value of k, n for determining the value there is a feedback mechanism negotiation, that is, when the distal end is low measured packet loss rate, n for the value may be set to a smaller number, and vice versa a larger number may be provided.

例如,当1^8时,远端测出的丢包率大于等于20%时,n的值选取32, 而当远端测出的丢包率等于10%时,n的值选取16。 For example, when 1 ^ 8, the distal end of the measured packet loss rate is equal to greater than 20%, 32 to select the value of n, when the distal end of the measured packet loss rate equal to 10%, 16 to select the value of n.

结合前述的帧结构,如杲用b3b2blb0表示所述的FEC类型,则可以用b3b2blb0的值来表示带有FEC机制的数据包,具体定义如下: The binding frame structure, as represented by the Gao with b3b2blb0 FEC type, can be represented by the value b3b2blb0 packets with FEC mechanism, defined as follows:

b3b2: 00--表示不采用FEC机制,当丢包率处于较低的水平时采用。 b3b2: 00-- means that no FEC mechanism, when using the packet loss rate at a low level. 01—表示采用FEC机制,并且k二8, 01- represents using FEC mechanism, and k = 8,

10— 表示采用FEC机制,并且k-16, 10- represents using FEC mechanism, and k-16,

11— -表示采用FEC机制,并且k二32。 11--- expressed using FEC mechanism, and k = 32. Mb0: 00—当k=8时,n= 12, Mb0: 00- When k = 8, n = 12,

当k=16时,n = 20, 当k=32时,n = 40 MbO: 01—-当k=8时,n= 16, 当k=16时,n = 24, 当k=32时,n = 48 blbO: 10—当k=8时,n = 24, 当k=16时,n = 28, 当k =32时,n = 56 blbO: 11—当k=8时,n = 32, 当k=16时,n = 32, 当k=32时,n = 64 当FEC使能字段为00时,指示禁用FEC字段,这时b3b2blb0的定义失效,应设置为全O。 When k = 16, n = 20, when k = 32, n = 40 MbO: 01-- when 8 k =, n = 16, when k = 16, n = 24, when k = 32, n = 48 blbO: 10- while when k = 8, n = 24, when k = 16, n = 28, when k = 32, n = 56 blbO: 11- when 8 k =, n = 32, when the time k = 16, n = 32, when k = 32, n = 64 when the FEC enable field is 00, indicating disabling FEC field defining b3b2blb0 case of failure should be set to all O.

下面以k-8为例,分析本发明的FEC机制在现有网络上的性能表现。 K-8 below to an example, analysis of the mechanism of the present invention, FEC performance on existing networks. 在网络上传输的各个数据包实际上是相对独立的,相互之间并无联系。 Each packet is actually transmitted over the network is relatively independent, there is no contact with each other. 现假设在分组网络上由于拥塞造成的实际丟包率为p,根据传输网络性能分析可知, 传输n个数据包丟失m (m<n)个以上的数据包的概率P(m)遵从二项分布, 可以用下式表示: It is assumed that the packet network due to congestion of the actual packet loss rate p, according to the transmission network performance analysis, transmission of n data packet loss m (m <n) the probability of more than one packet P (m) to comply with two distribution can be represented by the following formula:

乂'u Qe 'u

那么,当m〈^-k时,丢失的数据包可以利用已收到的凄t据包和FEC数据包 Then, when the m <^ - when k, lost data packets may be utilized desolate t already received packet and FEC packet.

完全再生。 Complete regeneration. 反之,则不能恢复。 On the contrary, it can not be restored.

下表是P (k=8)与p的关系,P(8)是在k-8时,当端到端丢包率为p时,采用FEC机制可以达到的最小丢包率。 The table below is the relationship P (k = 8) and the p, P (8) is k 8-time, packet loss rate p when the end, with a minimum loss rate FEC mechanism can be achieved. <table>table see original document page 24</column></row> <table>由此可见 <Table> table see original document page 24 </ column> </ row> <table> shows

当采用RS(32,8)编码并实现FEC机制,即使网络拥塞高达50%时,实际数据包丢失的概率也仅在0.1 %,这样的丢包率可以满足应用的要求。 When using the RS (32,8) encoding and FEC mechanism implemented, even when the network is congested up to 50%, the actual probability of packet loss is only 0.1% packet loss rate required to meet such applications.

3)时戳緩沖区设计 3) buffers stamp design

时戳緩冲区应能存储三种时戳,分别如下: (1 ) TDM时隙进入緩冲区的时戳(业务时钟计数器的值)。 When the time stamp buffer should be able to stamp stores three, are as follows: (1) TDM time slot into the time stamp buffer (the service clock counter value). (2 )收到的数据包携带的时戳(来自于RTP时戳域)。 (2) Upon receipt of the data package stamp (the time stamp field from RTP). (3 )收到数据包时本地时戳(业务时钟计数器的值)。 (3) time stamp (value of the clock counter service) when receiving the local packet.

本发明时戳緩冲区设计为1024字节,共可存储256个时戳信息。 The present invention is designed stamp buffer is 1024 bytes, a total of 256 store timestamp information.

4)发送緩沖区设计 4) Transmit Buffer Design

这里的发送緩冲区是指在El信息进入分组网时,緩存El信息以便封装形成数据包的緩冲区。 Here transmit buffer means when El incoming message packet network, the packet buffer cache information for a package-forming El. 需要緩存的信息主要有如下三部分: Need cached information mainly has the following three parts:

(1) 时隙组的缓存:由于这里是以时隙组为单位来封装的,所以,緩冲区大小不能小于一个时隙组的大小。 (1) Cache set of time slots: slot since this group is encapsulated as a unit, so the buffer size is not smaller than the size of a set of time slots. 这里一个时隙组最大可包含32个时隙,而每个E1时隙包含的净数据的长度为32字节,因此,发送緩沖区的大小应大于32*32=1024字节=8192比特。 Here a timeslot group contains a maximum 32 slots, and the length of each slot contains E1 net data is 32 bytes, and therefore, the transmission buffer size should be greater than 32 * 32 = 1024 bytes = 8192 bits.

(2) FEC信息的緩存:同时,在緩存一个时隙组后,需要生成FEC信息, 并将时隙组和FEC信息都封装成数据包,因此还需有緩存FEC信息和封装好的数据包的空间。 (2) FEC information cache: while a set of time slots in the cache, needs to generate FEC information, and slot group and the FEC information is encapsulated into packets, and therefore should also have cached information and FEC packets encapsulated Space.

由前述内容可知,FEC信息不大于最大时隙组(32个时隙)包含的信息长度。 Seen from the foregoing, FEC information is not greater than the maximum slot group (32 slots) contains information length.

(3)数据包的緩存:同时緩存的数据包的数量S可以比较小,即使S为1的情况下也可正常工作,但S为l时,效率上的损失很大,基本退化为先封装后发送的串行工作模式。 Buffer (3) data packets: data packets buffered at the same time the number of S may be relatively small, even when S is 1 can also work, but S is L, a large loss in efficiency, as the first substantially degraded Package after sending the serial mode of operation.

这里选取4作为緩存的数据包的数量,既可做到封包过程和发送数据包过程为并行操作,占用的緩存数量也较少,总的緩存数据包的大小<512字节。 Here the number of selected data packets as a buffer 4, the packet can be done during the process and send packets to operate in parallel, the number of buffer occupancy is also less, the total size of the packet buffer <512 bytes.

发送緩冲区的总量设计为3K字节,既满足要求,又有一定的备用空间存储时戳信息,或供功能拓展之需。 Total transmit buffer is designed to 3K bytes, both to meet the requirements, there are certain timestamp information stored spare space, or the need for expansion of the function.

在发送数据包时,普通数据包的发送时间取决于数据包(或其携带的时隙)对应的业务时钟计数值,也就是说,时隙组内两个相邻数据包的发送时间间隔与两个相邻业务时钟计数值的差值成比例。 When sending a packet, the transmission time of normal data packets depends on the packet (or carrying slots) corresponding to the service clock count value, that is, transmission time slots within the groups of two adjacent intervals with packets two proportional to the difference clock count value of the adjacent traffic. FEC数据包在两个普通数据包的发送间隔内发送,其中每个时间间隔发送的FEC数据包个数,根据n 和k的取值确定。 FEC data packets transmitted during the transmission interval of two ordinary data packets, wherein each time interval FEC number of packets transmitted, is determined according to the values ​​of n and k. 发送FEC数据包的时间确定原则为:尽量使相邻数据包发送的时间间隔均匀。 FEC packet transmission time determination principle: try to make the time interval of adjacent data packets transmitted uniformly.

5)去抖緩冲区的设计 5) designed to shake buffer

本发明的去抖緩冲区有两个职能:在不具有定时传送机制的情况下,用于自适应方式的定时信息恢复和用于丢失信息的再生。 Debounce buffer of the present invention has two functions: in the case of having no timer transmission mechanism for adaptively timing information for reproducing and restoring loss of information.

用作丢失信息的再生,则去抖緩沖区的大小至少能存放一个E1时隙组对应的信息量,由于k《32,而一路El的一个时隙为32字节,则去抖緩沖区的最低大小为<formula>formula see original document page 25</formula>字节=8192比特。 At least the size of the loss is used as reproduction information, the debounce buffer can store information corresponding to a set of time slots E1, since k "32, and one slot is 32 bytes El way, the de-jitter buffer minimum size of <formula> formula see original document page 25 </ formula> byte = 8192 bits. 实际去抖緩冲区的大小设置成最低大小的L倍,L=2, 4, 8...,可根据使用的可编程器件(这里是FPGA)的容量来决定。 The actual debounce buffer sized to L times the minimum size, L = 2, 4, 8 ..., can be used in a programmable device (here FPGA) capacitance is determined.

用作自适应时钟的恢复,则理论上去抖緩冲区的大小越大越好,而且不能太小。 As adaptive clock recovery, the theoretical jitter buffer size up better, and not too small. 因为去抖緩冲区太小,在恢复的时钟尚未锁定时,去抖緩沖区会出现上溢和下溢,造成数据丢失或同步问题。 Because the buffer is too small to shake, when the clock recovery has not been locked, to shake the buffer will overflow and underflow, resulting in loss of data or synchronization problems.

另一方面,去抖緩冲区过大,则导致的时延会有较大的增加。 On the other hand, to shake buffer is too large, resulting in delay there will be a greater increase. 图19表示的去抖缓冲区的结构,是一个环形緩沖池结构。 Debounce buffer configuration shown in FIG. 19, the annular structure is a buffer pool. 1901表示该緩沖区对应的时隙组数据已完全到达或部分到达但通过FEC 机制完全恢复,但尚没有被用作恢复成E1码流。 1901 denotes a slot corresponding to the set of data of the buffer has been completely or partly reaches the arrival but totally recovered by FEC mechanism, but not yet been restored to as E1 stream. 1902表示该緩沖区区对应的时隙组无数据包到达或该緩沖区的数据已被 The buffer region 1902 was no corresponding slot or a data packet arrives to the buffer has been

用作恢复成E1码流。 E1 as a recovery to codestream.

1903表示该緩沖区对应的时隙组包含的数据部分到达(没有全部到达), 1903 represents a data set comprising a slot corresponding to the buffer portion reaches (not all arrive),

并且没有使用FEC恢复丢失数据的情况。 And without the use of FEC to recover lost data.

1901表示该緩冲区对应的时隙组数据已完全到达或部分到达但通过FEC Slot group 1901 indicates that the data corresponding to the buffer have been completely or partially reaches through the FEC reaches but

机制完全恢复,且正在被恢复E1码流的情况。 Mechanism fully recovered, and the recovery is being E1 stream.

l卯5表示1903对应的时隙组中已经到达的数据包(阴影部分)。 5 l d denotes slot group 1903 corresponding to the data packet has arrived (hatched portion). 1906表示1903对应的时隙组中未到达的数据包(空白部分)。 1906 denotes slot group 1903 corresponding to the non-arrival packets (blank portion). l卯7表示环形緩沖池的写指针,指向当前时隙组中已到达的次序列号最 l d 7 represent the ring buffer pool write pointer, which points to the sequence number of the current slot group has reached the most

大的数据包包含的信息的最后一个字节。 Large data packets containing the last bytes of information.

1908表示环形緩冲池的读指针,指向恢复El码流时将要读取的数据。 1908 represents the ring buffer pool read pointer El recovered data symbol stream to be read when. 一般的,采用自适应方法恢复时钟时,初始填充级约为1/2,所以去抖緩 In general, adaptive clock recovery method, the initial filling of the order of about 1/2, so slow debounce

沖区引起的时延约为处理一半去抖緩沖区的时间。 Delay caused by the buffer zone to about half the processing time jitter buffer. 如果去抖緩沖区能存储L If you go to shake the buffer can store L

个时隙组,则去抖緩沖区会带来172*]<:* 125us的时延。 A slot group, the de-jitter buffer will bring * 172] <: * 125us of delay.

El对时延有比较高的要求。 El relatively high requirements for delay. El多携带语音信息,端到端语音信息的时 When El carry more than voice information, end-to-speech information

延达到25ms以上时,就会产生比较明显的回声,使语音质量下降。 When the delay reaches more than 25ms, it will produce more obvious echo, the voice quality. 使用专门 Use special

的回声消除装置虽然可以抑制回声,但成本较高。 Although the echo cancellation apparatus can suppress an echo, but the cost. 因此,设计去抖緩冲区不能导致时延过大。 Therefore, the buffer can not shake designed to cause too much delay.

分组网传送E1的端到端时延D是发送端处理延迟Dht,传送时延Dt, 转发时延Df已及接收端处理时延Dhr的累积和。 The packet network transmission of E1-end delay is the transmitting side processing delay D Dht, transmission delay Dt, Df is the forwarding delay and processing delay of the receiving side and Dhr accumulation.

现有网络长距离传输一般使用光纤作为传输媒体,100公里的光纤带来的传输时延约在300〜500US,因此,大多数情况下,Dt的影响可以忽略。 Conventional long-distance transmission network is generally used as the optical transmission medium, 100 km fiber transmission delay brought about 300~500US, therefore, in most cases, the influence can be ignored Dt.

现有的网络节点设备, 一般采用存储转发机制,因此会带来一定转发时延,整个分组网的端到端转发时延应是经过的路径所有网络节点设备的转发时延之和。 Existing network node apparatus, the general store and forward mechanism, and therefore will bring some forwarding delay, the entire end of the packet network forwarding delay of the delay should be forwarded through all the paths of the network nodes and devices. 但是,现有网络节点设备的转发机制多采用高速ASIC或NP实现, 在没有网络拥塞的情况下,转发时延相对较小。 However, existing network node forwarding mechanism device to use more or NP-speed ASIC implementation, in the absence of network congestion, the forwarding delay is relatively small.

如果在出现拥塞的情况下,则实际导致的时延是一个随机变化的量,有可能很大。 If the delay in the event of congestion, the actual amount is due to a random change, there may be significant. 换句话说,时延抖动有可能处于一个很高的水平。 In other words, it is possible to delay jitter at a very high level. 但结合超时定时器使用FEC机制,可以使较大的时延抖动得到抑制。 However, the timeout timer used in conjunction with FEC mechanism, a large delay jitter can be suppressed. 因此分析时延时,可以将拥塞的情况当作没有网络拥塞的情况处理。 Therefore, the analysis of delay, congestion may be the case as no network congestion processing.

下表表示的是在无拥塞的情况下,某种网络节点设备实测的不同长度的数据包通过该网络设备产生的时延。 The following table shows that in the absence of congestion situation, some devices of different lengths measured latency through the network nodes the data packets generated by the network device.

<table>table see original document page 27</column></row> <table> <Table> table see original document page 27 </ column> </ row> <table>

由于本发明的数据包都在128字节以下,因此,需经过120台以上的这种网络设备才会导致lms的时延,因此,大多数情况下也可以忽略Df的影响。 Since the data packet in the present invention are less than 128 bytes, and therefore, subject to the above network device 120 such units would result in delay of lms, therefore, can be ignored in most cases also influence the Df.

因此,在端到端距离较短,且经过路径上的网络设备较少,则D - Dht + Thus, in the end a short distance, and after less network devices on the path, the D - Dht +

Dhr。 Dhr.

由于在发送端需要将完整的时隙組同时映射成数据包,而每个时隙的周期为125us,因此发送端会带来固有的k"25us的緩冲延迟,而去抖緩沖区则也会带来L/2承k举125us的时延,加上封包时延,编码时延,解码时延以及解包时延的总和不能超过门限,因此L在满足要求的情况下,不能设置过大。 Since the transmitting side needs to map the full set of time slots simultaneously into packets, each slot period is 125us, thus transmitting end will bring inherent k "25us buffering delay, jitter buffer is also away delay will bring L / 2 k supporting the lifting 125us, plus the sum of packet delay, delay encoding, decoding delay and delay unpacking not exceed the threshold, so in the case of L meet the requirements, you can not be set too Big.

封包时延+编码时延与时隙组的大小相关,如果k为32,则封包时延+编码时延的和不能大于32*125us=4ms。 + Size of packet delay associated with coding delay slot group, and if k is 32, the encoded packet delay + delay can not exceed 32 * 125us = 4ms.

同样,解码时延和解包时延也与时隙组的大小相关,,如果k为32,则解码时延+解包时延的和不能也大于32*125us=4ms。 Similarly, the decoding delay unpacking delay is also associated with the size of the set of time slots, that if k is 32, the decoder unpacking delay + delay and also not greater than 32 * 125us = 4ms.

如果为满足端到端的总时延小于25ms,贝'J To meet the end is less than the total delay of 25ms, Tony 'J

L/2 * k * 0.125+解码时延+解包时延+封包时延+编码时延<25, L / 2 * k * 0.125+ decoder unpacking delay + delay + delay + encoder packet delay <25,

则l^k〈272,这就是需要满足的条件。 The l ^ k <272, this is the condition to be met.

本发明设计L值最大为32。 The present invention is designed maximum value of L 32.

在k=8时,L的值一般设计为32。 When k = 8, the value of L is typically designed to be 32.

在k-16时,L的值一般设计为16。 When k-16, the value of L is typically designed to be 16.

在k=32时,L的值一般设计为8。 When k = 32, the value of L is typically designed to 8.

6)接收定时器设计。 6) receiving timer design.

接收定时器的职能是为数据包恢复提供定时信息。 Receiving timer function is to provide a timing information packet recovery.

图19表示去抖緩冲区,它实际是一个环形队列,使用读指针Pll(如1907 所示)和写指针P12 (如1908所示)两个指针指示。 19 shows a debounce buffer, it is actually a circular queue, the read pointer using Pll (shown as 1907) and write pointer P12 (shown as 1908) indicates two pointers. 其存放的L个时隙组逻 Which store logical group of L time slots

辑上可以看作是有序的,主序列号相对小(由于是环形结构,并不存在绝对顺序)的在前面。 Can be seen as the ordered series, the main sequence number is relatively small (because it is an annular structure, there is no absolute order) at the front.

令写指针指向第S个緩冲区,读指针指向第T个緩冲区。 So that the S-th write pointer points to a buffer, the read pointer points to a buffer of the T. 同时令纠错缓 At the same time so slow correction

冲区存储的FEC包组的个数为R。 FEC packet group number stored in the buffer zone is R.

当接收定时器溢出时,系统从第(T+l )。 Upon receiving the timer overflows from the system (T + l). /。 /. L个緩冲区开始检查,到第S个緩沖区结束。 L start checking buffers, buffers to the second end of the S.

首先检查去抖緩冲区的信息,分以下两种情况讨论: First check the information to shake buffer, divided into the following two cases discussed:

〈1〉如果(S + L- (T+l ))。 <1> if (S + L- (T + l)). /oL大于R,表示该缓冲区不在本次FEC 处理的范围,如果该緩冲区对应的时隙组存在某一个或多个时隙丟失,也不能够使用FEC机制恢复,则有以下三个可能: / OL greater than R, indicating that the buffer is not in the scope of this FEC processing, if the presence of the buffer corresponding to slot group lost one or more time slots, not using the FEC mechanism can be restored, there are the following three may:

若丟失的时隙的上一个时隙存在,则以上一个时隙填充本时隙,否则, 若丢失的时隙的下一个时隙存在,则以下一个时隙填充本时隙,否则, 若丟失的时隙的上一个时隙和下一个时隙都不存在,则以全1填充本时隙。 If missing slots on a time slot exists, more than one slot of this time slot is filled, otherwise, if the lost time slot one slot is present, then the following the present slot fill a slot, or, if lost on a slot by slot and the next does not exist, the whole of this time slot 1 is filled.

<2>如果(S + L- (T+1 ))。 <2> if (S + L- (T + 1)). /。 /. L小于或等于R,表示该緩冲区在本次FEC处理的范围,如果该緩冲区对应的时隙组存在某一个或多个时隙丟失,则结合纠错緩冲区的信息判定是否能够纠错,如能够纠错,则将相应数据送交解码模块,通过特定算法恢复丟失的信息。 L is less than or equal to R, indicating that the FEC buffer scope of this process, if the presence of the buffer corresponding to slot group one or more time slots is missing, then the combined information determines whether an error correction buffer error correction is possible, as can the error correction, then the corresponding data is sent to the decoding module, recovering a lost through a specific algorithm.

接收定时器的周期的设置非常重要。 Set to receive timer period is very important.

过长则会导致信息丢失。 Too long it will lead to loss of information. 如果溢出周期过长,当接收定时器溢出时,可能读指针已经越过应该处理的时隙组对应的緩冲区。 If the overflow is too long period, when receiving the timer overflow slot group may read pointer corresponding to a buffer to be processed has been crossed. 造成信息丢失。 Loss of information.

周期过短可能不能保证在一个周期内处理完所有的从第(T+l) 。 Cycle is too short may not be processed in a guaranteed period from all the (T + l). /。 /. L个緩冲区开始的,到第S个缓冲区结束的所有緩沖区。 L buffers from the beginning to the end of the first S buffers all buffers.

这里设置周期约等于一个时隙组的时长。 Here a set period of time approximately equal to the length of slot group.

7)纠错緩冲区设计 7) Correction buffer design

纠错缓冲区用于存放FEC数据包,由前述内容可知, 一个FEC数据包组包含的数据包数最大可达到32,因此,纠错緩沖区必须保证能存放32个时隙对应的数据。 FEC correction buffer for storing data packets, can be seen from the foregoing, the maximum FEC data packet group comprises a number of packets up to 32, therefore, must ensure that the error correction buffer 32 can store data corresponding to the slot.

在网络时延变化量较大时,可能导致去抖緩沖区很多缓冲区都处于部分填充状态(即緩冲区对应的时隙组只有部分时隙的数据接收到),最极端的情况,可能有L/2个缓冲区处于这样一来的情况,那么与之对应的纠错緩沖区 In a large network delay variation, jitter buffer may lead to a lot of buffers are in a partially filled state (i.e., only a portion of the received data slot buffer corresponding to slot group), the most extreme case, it may there are L / 2 buffers in this way in the case, then the corresponding error correction buffer

最大也应能存放L/2个FEC包组。 It should also be able to store the maximum L / 2 th FEC packet groups.

但经分析确定绝大多数情况下不需要设置L/2个FEC包组,且太大对资源的利用率也较低。 But not necessary to provide L / 2 th FEC packet group under analysis is determined by the majority of cases, and the utilization of resources is too low.

本发明设计的纠错緩沖区的大小为可存放8个FEC包组。 Size of the present invention is designed for error correction buffer can store 8 FEC packet groups.

8) 编码模块设计。 8) Design encoding module.

图14是编码模块的流程图。 FIG 14 is a flowchart of an encoding module.

本发明的FEC机制就是使用一些冗余数据包来帮助再生丢失的数据包。 FEC mechanism according to the present invention is the use of a number of redundant packets to help regeneration of lost packets. 采用RS (n, k)编码来生成冗余数据包。 Using RS (n, k) code to generate the redundant data packets.

RS(n, k)码是一种非二进制纠错码,有很强的纠正突发性错误的能力。 RS (n, k) code is a non-binary error correction code, has a strong ability to correct burst errors. 它利用一个基本数据块经过运算,生成冗余块,将冗余块与基本数据块合并一起得到一具有纠错功能的数据块。 It uses a basic block after operation, generates redundant block, the redundant block is combined with the basic data block to obtain a block of data having an error correction function.

由于每个时隙组对应多个普通数据包,而每个普通数据包并不与FEC包 Since each group corresponds to plurality of slots ordinary data packets, and each packet is not the normal data packet and FEC

——对应,因此并不能使用编码器直接得到FEC包组。 - correspondence, and therefore can not be used directly to obtain FEC encoder packet group.

将一个时隙组集中映射成分组数据包时,可以配置使用每个时隙独立映射成包的独立映射方式,也可选择从一组E1时隙中的每个时隙选取部分数据,再封装成数据包的合成映射方式 When a set of time slots as the map component group of data packets, each slot can be configured using separate mapping mode mapped into separate packets, or choose from a set of selected portions of the data in the E1 time slots each time slot, then the package synthesis mapping method into data packets

本发明的主要思想是利用时隙组中每个时隙的对应字节(以i表示)组合成一个k字节的基本数据块,该数据块经过编码逻辑(例如余数生成逻辑) 得到(nk)长度的冗余块,冗余块的每一字节分别作为FEC包组中每一数据包的i字节。 The main idea of ​​the invention is the use of slot group of bytes corresponding to each slot (represented by i) into a k byte blocks of basic data, the logical block of data encoded (e.g. remainder generation logic) to give (nk ) redundant block, the redundant block length of each byte of each packet group as i FEC bytes of each packet.

如果由反馈信息获取的端到端丢包率满足应用要求,则不需要形成FEC包。 If meet the application requirements by the end of the packet loss rate feedback information acquisition, the FEC packet is formed is not required.

9) 封包模块设计 9) packet module design

图4是本发明采用的协议结构。 FIG 4 is a protocol structure used in the present invention.

图5是IETF PWE3任务组标准或草案中定义的TDMoIP的控制字的图示。 FIG 5 is an illustration of a control word TDMoIP Task Force IETF PWE3 draft standard or defined.

图6是本发明定义的控制字的图示。 FIG 6 is an illustration of the control word defined in the present invention.

图7是本发明完整的协议分层结构图示。 FIG 7 is a layered structure illustrating the complete protocol of the present invention. 200610163606.2 200,610,163,606.2

说明书第24/32页 Instructions Page 24/32

图8是本发明完整的帧结构图示。 FIG 8 is a complete frame structure of the present invention is illustrated.

封包模块的功能就是将发送缓沖区的内容封装成图7, 8描述的数据包, 通过诸如IP或以太网的分组网发送到接收端。 Packet module function is to transmit the content package into buffer 7, FIG. 8 described in the packet, transmitted to the receiving end through the packet network such as IP or Ethernet. 每个数据包净荷是32字节。 Each data packet is a 32 byte payload. 封装成的数据包是90字节:分别是以太网头14字节,IP头20字节,UDP 头8字节,RTP协议12字节,控制字4字节,信息净负荷32字节。 Encapsulated packet is 90 bytes: Ethernet header are 14 Bytes, IP header 20 bytes, UDP header of 8 bytes, 12 bytes of the RTP protocol, the control word 4 bytes, 32 bytes of payload.

由于引入FEC机制,TDMoIP的控制字中的RES字段,长度字段以及序列号字段都已有新的定义,参见前述帧结构的内容。 Since the introduction of FEC mechanism, TDMoIP the RES field in the control word, the length field and sequence number fields have new definitions, see the contents of the frame structure.

同样,由于以时隙组为单位封包,控制字中的序列号定义为二維的结构。 Also, since the slot packet group unit, the control sequence number is defined for the word two-dimensional structure. 主序列号10位,次序列号6位,次序列号的前k个编号为普通数据包使用, k+l—n区间的序列号给FEC数据包使用。 Main SEQ ID NO 10, SEQ ID NO 6 times, first k number of times the serial number used for the ordinary data packets, k + l-n interval FEC packet sequence number to use.

图15是封包模块的示意图。 FIG 15 is a schematic diagram of packet module.

生成以太网帧头时,使用的源MAC和目的MAC地址应与实际设备配置的相同。 When generating the Ethernet frame header, source and destination MAC addresses should be configured the same as the actual device.

生成IP包头时,使用的源IP地址和目的IP地址应与实际设备配置的相同。 Generating the IP header, the same source IP address and destination IP address should be configured to use the actual device.

FEC类型字段的内容极其重要,不能出错。 FEC content type field is extremely important, you can not go wrong. 序列号也非常重要,不能出错。 Serial number is also very important, you can not go wrong.

10)解包模块设计 10) Design unpacking module

图16是解包模块的流程图。 FIG 16 is a flowchart unpacked module.

为简化设计和减少处理带来的时延。 To simplify the design process and reduce the delay caused. 本模块不解析以太网数据包帧头和IP包头的内容。 This module does not resolve the Ethernet packet header and the IP header contents. 以太网帧头和IP包头仅用于在分组网上实现数据包转发。 Ethernet frame header and IP header only for forwarding packets in the packet network. 首先分析数据包的控制字中的FEC类型字段,得到FEC机制的n和k值。 Firstly, FEC type field of the control word data packet, the n and k values ​​to obtain FEC mechanism.

再取出数据包的控制字中的序列号,分解得到主序列号和次序列号。 Then remove the sequence number control word data packet, the sequence number and the decomposition of the main secondary sequence number. 主序列号MOD去抖緩冲区的时隙组数L得到当前时隙组在去抖緩冲区或纠错缓冲区的存储位置。 SEQ ID NO MOD main slot group number L debounce buffer memory location to obtain the current time slot group debounce buffer or correction buffer.

如次序列号〉n,则是非法数据包,需丢弃。 SEQ ID NO summarized as follows> n, then the packet is illegal, it need be discarded.

如!1>=次序列号〉k,则是FEC数据包,需存入纠错緩冲区的相应位置。 The! 1> secondary sequence number => K, is the FEC packet, a corresponding position error correction into the buffer required. 如次序列号《k,则是普通数据包,需存入去抖緩冲区的相应位置。 Summarized as follows sequence number "k, ordinary data packets are required to deposit the corresponding position jitter buffer. 进入緩冲区的数据不包含所有的包头信息。 Data into the buffer does not contain all of the header information.

11) 解码模块设计。 11) decoding module design.

图17是解码模块的流程图。 FIG 17 is a flowchart of a decoding module.

在进行解码之前,需统计已收到的数据包和应收到的数据包的值,并计算端到端丢包率,以便反馈给远端。 Before decoding, the statistics for an already received data packets and data packets to be received, and calculates the end PLR for feedback to the distal end.

当FORMID指示没有FEC机制时,不需要解码模块。 When FORMID indicates no FEC mechanism, it does not require a decoding module.

当检查去抖緩冲区并未发现信息丢失时,解码模块也不工作。 When the inspection did not find the time to shake the buffer information is lost, decoding module does not work.

与编码模块对应,解码模块实质也不能直接恢复丢失的数据包,而是逐 Corresponding to the encoding module, a decoding module to substantially can not recover the lost data packets directly, but by

字节恢复丟失的数据包包含的信息。 Byte recover lost information included in the packet.

也就是,将收到的数据包和FEC包的对应字节以及丟失数据包的次序列 That is, the sequence corresponding to the secondary-byte packets and FEC packets received and lost packets

号送交解码逻辑,通过运算得到丟失的字节,再送到緩沖区的特定位置。 No. sent decode logic, by calculation to obtain the missing bytes, and then sent to a specific location of the buffer.

12) 时钟恢复模块设计。 12) clock recovery module design.

在不使用同一的外部参考时钟直接作为两端业务时钟的情况下,时钟恢复主要有两种方法:差分时钟恢复方法和自适应时钟恢复方法。 In the case of no external reference clock directly as the ends of the same service clock, the clock recovery mainly two methods: differential clock recovery method and the adaptive clock recovery method.

本发明可通过配置确定为采用差分时钟恢复方法还是自适应时钟恢复方法。 The present invention may be configured as determined by differential clock recovery method is an adaptive clock recovery method.

差分时钟恢复方法要求在两端具有同一的参考时钟。 Differential clock recovery method requires the same reference clock on both ends. 本发明中设计一个32位业务时钟计数器对本地业务时钟计数,同时设计一个32位参考时钟计数器对外部参考时钟计数。 The present invention is designed to service a 32-bit clock counter for counting clock local service, while a 32-bit design of the external reference clock counter counts the reference clock.

在发送端,当一个完整的时隙进入緩冲区后,记录本地业务时钟计数器和外部参考时钟计数器的值,并计算本地业务时钟与外部参考时钟的差值, At the transmitting side, after a full slot into the buffer, the recording counter value of the local service clock and the external reference clock counter, and calculating a difference local service clock and the external reference clock,

并将这个差值通过数据包(在RTP协议时戳域携带)发送到接收端 This difference and the data packets (RTP protocol time stamp field carries) transmitted to a receiving end

在接收端,从数据包中(RTP协议时戳域)提取发送端业务时钟和与外 At the receiving end, and extracts the transmission clock with the outer side service from the packet (time stamp field RTP protocol)

部参考时钟的差值,并根据这个差值结合外部参考时钟得到本地业务时钟。 The difference between the reference clock unit, in accordance with this difference and the reference clock in conjunction with the external clock to obtain local traffic. 差分时钟恢复方法较为直接,但需要依赖同一的外部参考时钟,这里叙 Differential clock recovery method is more direct, but relies on the same external reference clock, where Syria

述从略。 Said omitted.

这里重点讨论自适应时钟恢复模块。 Here we focus on adaptive clock recovery module.

由于分组网的时延问题,自适应时钟恢复方法恢复的时钟在相位上有一定延迟,但在锁定状态下,频率差异应在规定的范围内。 Since the packet network delay problem, an adaptive clock recovery method for recovering the clock of a certain delay in phase, but in the locked state, the frequency difference to be within a predetermined range.

自适应时钟恢复的基本原理类似于PLL的自动反馈调节机制。 The basic principle is similar to the adaptive clock recovery PLL automatic feedback regulation mechanism.

概括起来也就是三个主要功能块: That is summed up in three main functional blocks:

"鉴相"逻辑:这实际上并不等同于PLL的鉴相器,但输出在功能上与鉴相器的输出等价。 "Phase" Logic: This is actually not the same as the PLL phase detector, the output is functionally equivalent to the output of the phase detector. 也就是利用去抖緩冲区的使用情况,或利用时戳信息,得到本地业务时钟和远端业务时钟的差异表示。 I.e. using the time stamp information to the jitter buffer usage, or the use, to obtain the service clock and the local difference in clock teleservice FIG.

环路滤波逻辑:使用滤波逻辑消除分组网时延变化带来的抖动。 Loop filtering logic: using the filtered logic eliminate jitter caused by the packet network delay variation. 压控振荡器/数控振荡器:结合环路滤波逻辑的输出生成本地业务时钟。 VCO / numerically controlled oscillator: binding loop filtering logic output clock to generate a local traffic. El的标准频率是2.048 Mhz, ITU-T G.703对El频率精度的要求是2.048 Mhzi50ppm。 El standard frequency is 2.048 Mhz, ITU-T G.703 requirements for frequency accuracy is El 2.048 Mhzi50ppm. 因此,发送端的业务时钟频率ft应落在(2047897.6hz, 2048102.4hz)区间。 Therefore, the transmission side clock frequency ft traffic should fall (2047897.6hz, 2048102.4hz) interval. 同样接收端依据数据包恢复的时钟频率f;范围也应落在该区间。 Similarly the receiving end based on data packet recovered clock frequency F; should also fall within this interval range.

理想情况下,fr=ft 。 Ideally, fr = ft.

但由于分组网的特性,大多数情况下,该等式并不能满足。 However, due to the characteristics of the packet network, in most cases, this equation can not be satisfied. 定义Af-fr-fp 由上述可知,Af必然落在(-204.8 hz, 204.8 hz)区间, Definition of Af-fr-fp seen from the above, the inevitable fall of Af (-204.8 hz, 204.8 hz) interval,

而MAX (|Af|) =409.6 hz。 The MAX (| Af |) = 409.6 hz.

区间(-204.8 hz, 204.8 hz )就是接收端业务时钟围绕中心频率(2.048 Mhz ) Interval (-204.8 hz, 204.8 hz) is to receive service clock terminal around the center frequency (2.048 Mhz)

波动的范围。 Range fluctuations. 自适应恢复方法的目的就是要消除这种差异,使Af尽可能趋近于0。 The purpose of adaptive recovery method is to eliminate this difference, the Af close to zero as possible.

在实际应用时,接收端业务时钟的调整范围是可配置的。 In practice, the receiving end of the adjustment range of the service clock is configurable.

如果发送端业务时钟ft劣化到超出ITU-T G.703的规定,为使&跟踪fp 可以通过配置信息调大接收端时钟调整的范围。 If the transmitting side service degradation beyond predetermined clock ft ITU-T G.703, in order to make the & fp can be configured to track a large range of adjustment information receiving terminal clock adjustment.

同样,如果发送端业务时钟ft具有较高的精度、较好的漂移和抖动特性, 可以通过配置信息调d、接收端时钟调整的范围。 Similarly, if the transmission end the service clock ft with higher precision, better wander and jitter characteristics, configuration information can be adjusted by d, the receiving terminal clock adjustment range.

这里用w表示时钟调整的配置信息,其含义是本地业务时钟的精度为2.O48 Mhz土w ppm ,则本地业务时钟的频率范围为: Herein represents the clock adjustment configuration information w, which means that the accuracy of local service 2.O48 Mhz clock soil w ppm, the range of the local service clock frequency is:

((2.048M-2.048w) hz, ( 2.048 M + 2德w) hz ),调整幅度为4.096 whz。 ((2.048M-2.048w) hz, (2.048 M + 2 de w) hz), adjustments for 4.096 whz.

本发明的自适应恢复方法有两种模式:基于緩冲区填充级模式和基于RTP时戳模式。 Adaptive recovery method of the present invention has two modes: mode based buffer fill level and a timestamp of RTP-based mode.

(1)基于緩冲区填充级模式使用本模式对远端没有依赖,数据包不需要显式携带定时信息。 (1) based on the buffer fill level Mode This no dependency on the distal end, carrying the data packets do not require explicit timing information. 图18是基于緩冲区填充级模式的自适应时钟恢复模块的功能示意图。 FIG 18 is a functional schematic diagram of the buffer fill level adaptive clock recovery module based mode.

定义填充级Fz去抖緩沖区中所有未恢复成E1码流的字节数与整个去抖緩冲区大小的比值。 Define a fill-level debounce buffer Fz all bytes are not restored to the stream with the entire E1 debounce buffer size ratio.

假设整个去抖緩沖区的大小为Sl字节,则Sl可用下式计算得到: SI =L * k * 32 Debounce assumed that the entire size of the buffer byte Sl, Sl is calculated by the formula: SI = L * k * 32

如图19所示,假设写指针(1907)为Pll,读指针(1908 )为P12。 19, assuming that the write pointer (1907) for the Pll, read pointer (1908) to P12. Pll, P12都是相对于去抖緩冲区的逻辑起始位置的偏移量。 Pll, P12 are offset with respect to the logical starting position jitter buffer.

读指针P12可以是逐字节的移动,利用移位逻辑甚至可以做到是逐位移动。 P12 read pointer may be moved byte by byte, using the shift logic may even be done by bit. 这里把P12看作是逐字节移动 Here it is considered the byte-wise movement P12

写指针P11是以时隙(32字节)为单位移动的。 P11 is a slot write pointer (32 bytes) as the mobile. 它总是指向去抖緩冲区中最新到达的时隙组,并且指向该时隙组中到达的具有最大次序列号的时隙的最后一个字节。 It always points to the buffer slot group debounce newly arrived, and points to the last byte of time slot having the largest sequence number in the time slot group reached.

则F可用下式计算: Then F can be calculated:

F= (Pll +S1 -P12) /Sl F = (Pll + S1 -P12) / Sl

本发明的填充级偏移量FP定义为F相对于半填充状态(F=50% )的差值。 FP filling level shift amount F of the invention is defined with respect to the half-filled state a difference (F = 50%) of.

即: which is:

FP = F_0.5 FP = F_0.5

由于从分组网侧收到的数据包不断填充去抖緩沖区,而El侧不断将去抖緩冲区的数据发送出去,这样緩冲区的使用情况是不断变化的。 Because the received packet from the packet network side continuously filled debounce buffer, while El-side debounced constantly send out data buffer so that buffer usage is constantly changing. 因此填充级增量FD为时间的函数,随着时间T变化。 Thus the filling level as a function of time increments FD, T changes with time.

初始情况下,F=0.5。 Initially, F = 0.5.

当提取的时钟比发送端的业务快时,则填充级减少,填充级偏移量FP 是负值。 When the extracted clock is faster than the transmission service end, the reduced filling level, the filling level offset is negative FP.

当提取的时钟比发送端的业务慢时,则填充级增加,填充级偏移量FP 是正值。 When the extracted clock than the transmitting side service slow, the fill level increases, the filling level offset is a positive value FP.

FP经数模转换后得到一个控制信号V, V也是时间T的函数。 FP after digital to analog conversion function to obtain a control signal V, V T is the time.

这个V可以看作是上述PLL机制的鉴相电路的输出。 This output V can be regarded as the phase of the PLL circuit mechanism.

V经低通滤波后,消除掉网络时延变化等因素引入的随机抖动成份。 After low-pass filtering V, eliminate random jitter introduced by the network component delay variation and other factors.

由于V是相对较弱的信号,需以一定增益比放大得到Vc。 Since V is a relatively weak signal amplification required to obtain a constant gain ratio Vc. Vc作为压控 Vc as a voltage-controlled

振荡器的输入,使压控振荡器产生一定频率增量,与振荡器的中心频率(2.048 Input of the oscillator, voltage controlled oscillator to generate a frequency increment, with the center frequency of the oscillator (2.048

Mhz)合成得到El的发送时钟。 MHz) El synthesized in the transmit clock.

令Af'为本次时钟调整的频率增量,则Af'是填充级偏移量FP的函数。 Order Af 'oriented secondary clock frequency increment adjustment, the Af' is offset FP filling level function.

基本可以用下式表示: It can be substantially represented by the following formula:

Af、4馬w *FP/0.5 上述的压控振荡器的输入应能使压控振荡器的输出产生Af1的调整量。 Af, horses input 4 w * FP / 0.5 of the above-described voltage controlled oscillator VCO output should enable adjustment amount generated by Af1. 经过多次反馈调节后,填充级基本稳定的等于0.5,也就是半填充状态, 或以半填充状态为中心以极小的幅度振动。 After several feedback regulation, substantially stable filling level is equal to 0.5, that is, a half-filled state, or in a half-filled state with a minimum amplitude of the vibration center. 这时,系统时钟处于锁定状态,本地服务时钟与远端服务时钟基本匹配。 In this case, the system clock is locked, the local clock and the service clock substantially matching the distal end of the service. (2)基于RTP时戳模式 (2) based on the RTP time stamp mode

直观的说,该模式同样是利用适当机制,产生一个频率增量Af,该增量 Intuitively, this mode is also the use of a suitable mechanism to produce a frequency increment of Af, the delta

与压控振荡器的中心频率合成得到本地的业务时钟f;。 And the center frequency of the VCO is synthesized local service clock f ;.

图20是基于RTP时戳模式的自适应时钟恢复模块的功能示意图。 FIG 20 is a functional schematic diagram of an adaptive clock recovery RTP timestamp mode Module. 与基于缓沖区填充级模式不同的是,两端都需要对本地业务时钟的进行 And based on different buffer filling level mode is required at both ends of the local service clock

统计 statistics

以一个32位计数器对本地业务时钟计数。 In a 32-bit counter for counting clock local traffic.

当TDM时隙完整进入緩冲区后,记录本地业务时钟计数器的值以备生成数据包对应的时戳。 After complete TDM slots into the buffer, the recording clock counter value of the local traffic to prepare for generating a time stamp packet corresponds.

当发送端需要发送数据包(以时隙组为单位发送)时,在RTP协议时戳域填入前述记录的本地业务时钟计数值。 When the transmitting side needs to send a packet (a unit of transmission in slot group), when the RTP timestamp field clock count value filled in the local service record.

当接收端从分组网侧接收一个数据包时,从该数据包提取时戳信息,并将该时戳作为数据包发送(分组网侧)时戳按序列号记录到时戳緩冲区的相应位置。 When the receiving end receives the corresponding packet from the packet network side, timestamp information is extracted from the data packet, and time stamp when the stamp as a data packet transmission (packet network side) of the press stamp serial number is recorded to the buffer position.

以Tts (n)表示该时戳。 In Tts (n) represents the time stamp.

同时记录本地本地业务时钟计数器的值,作为该数据包的接收时戳。 Local traffic simultaneously recorded value of the local clock counter, as the reception time stamp of the data packet. 以Rts (n)表示该时戳。 In Rts (n) represents the time stamp.

与去抖环冲区相同,时戳緩沖区也是一个循环结构。 Debounce ring with the same buffer zone, the buffer is a time stamp cyclic structure. 可以存放T个发送时戳和T个接收时戳。 T stamp and a time stamp may be stored when receiving the T send. 序列号为n的时戳在时戳緩冲区的存储位置由n % T确定。 Timestamp buffer when the serial number n is determined by the storage location stamp n% T.

当本端接收到第l个数据包时,开始记录发送时戳信息。 When the present l receives the first data packet, timestamp information transmission start recording. 当去抖緩冲区由全空到半填充状态时,El P12读指针指向第1个有效 When dejitter buffer to completely empty the half-filled state, El P12 read pointer points to the first valid

时隙组的第1个时隙的第1个字节。 First slot in the first byte of slot group. 此时时戳缓冲区已存放了L/2个接收时 At this time, when the buffer has been stored for the timestamp L / 2 received

戳信息及L/2个发送时戳信息。 Time stamp information and L / 2 th timestamp message.

令t (n)表示P12读指针指向第n个有效时隙组的第n个时隙的第1个 Order t (n) represents the read pointer P12 of a n-th slot n-th active set of time slots

字节的时刻。 Byte moment.

令ARts (n) =Rts (n+l ) -Rts (n), 且ATts ( n ) = Tts ( n+l ) - Tts ( n )。 Order ARts (n) = Rts (n + l) -Rts (n), and ATts (n) = Tts (n + l) - Tts (n). 令At ( n ) =ATts ( n ) -ARts ( n ), 令At1 (n) =At (n) /ATts (n), Order At (n) = ATts (n) -ARts (n), so At1 (n) = At ​​(n) / ATts (n),

则At1 (n)可以看作是上述PLL机制的鉴相器输出,利用这个输出经环路滤波器,在控制压控振荡器产生本地的业务时钟。 The At1 (n) can be regarded as the phase of the PLL output mechanism, by using the output of the loop filter, a voltage controlled oscillator generating a local control of the service clock.

也可以这么说,根据这个At1 (n)产生Af1 (n),而这个Af"(n)与本地振荡器的中心频率合成得到t (n)时刻的本地业务时钟。 △f1 (n)与时戳的关系可以由下式确定: The same can be said that, to produce Af1 (n) according to this At1 (n), and this Af "(n) and the center frequency of the local oscillator of the local service synthesized clock t (n) time. △ f1 (n) when relations stamp can be determined by the following formula:

△f1 ( n ) = 4.096 w *At ( n ) /ATts ( n ) 滤波逻辑可以用加权移动平均算法或指数加权移动算法。 △ f1 (n) = 4.096 w * At (n) / ATts (n) filtering logic algorithm weighted moving or exponential weighted moving average algorithm.

13 )远端类型确定 13) determining the type of the distal end

此步骤是用来确定与之通信的远端是否具有FEC机制。 This step is used to determine whether communication with the distal end of FEC mechanism.

由于控制字的两个保留字段分别定义为FEC类型字段和FEC使能字段, 所以判定远端是否具有相同的FEC机制比较方便。 Since the control word is two reserved fields are defined as a type field and an FEC FEC enable field, it is determined whether the distal end of the same FEC mechanism is convenient.

当FEC使能字段为OO,且FEC类型字段为OOOO时,可以判断远端没有相应的FEC机制。 When the FEC enable field of OO, and FEC type field is OOOO, the distal end can be determined without a corresponding FEC mechanism. 那样,本端也需要禁用FEC机制,需要作以下的处理: As this end also need to disable FEC mechanisms required for the following process:

封包时FEC使能字段设为00,且FEC类型字段设为0000。 FEC enable field is set to 00 packets and the FEC type field to 0000.

长度字段设为全O。 Full length field is set to O.

不生成FEC包。 FEC packets are not generated.

序列号不再分成两级。 SEQ ID no longer divided into two.

封包时,不以时隙组为单位,以时隙为单位。 When the packet is not in slot group as a unit, in units of slots. 本发明的应用实例 Examples of application of the present invention

武汉烽火网络有限责任公司研制的多业务分组平台产品M8000/T设备, 完成了在以太网上高效可靠的传送E1的功能。 Wuhan beacon network limited liability company developed the multi-service packet platform products M8000 / T equipment, completed over Ethernet function efficiently and reliably transmit E1.

图21、是M8000/T设备的正面轮廓,有4个百兆以太网光口, 4个百兆以太网电口以及4个千兆以太网光口。 FIG 21 is a M8000 / T device front profile, has four optical Fast Ethernet ports, Fast Ethernet ports 4 and four Gigabit Ethernet optical port. 这些以太网口可用于与远端设备互联。 The Ethernet port may be used to interconnect the distal end of the device.

图22是M8000/T设备的背面轮廓,有1-32个El接口。 FIG 22 is a M8000 / back surface profile of the device T, there are interfaces 1-32 El.

如图23所示,将3台M8000/T搭成线性拓朴的网络。 23, the three M8000 / T shelter of a linear network topology.

图23中,2301是El接口,连4妻2M误码仪和M8000/T设备。 In Figure 23, the interface 2301 is El, even his wife 2M BERT and 4 M8000 / T device.

2302是M8000/T设备。 2302 is the M8000 / T equipment.

2203是背景流量输入接口,从千兆接口输入,最大可输入2G的流量。 2203 is the background flow input interface, input from Gigabit interfaces, the maximum input flow of 2G.

2304是M8000/T设备的千兆接口。 2304 is the M8000 / T Gigabit Interface devices.

2305表示用千兆接口将两台设备连接起来。 2305 is represented by two Gigabit interfaces to connect devices.

图9是本发明的系统结构描述, FIG 9 is a system configuration according to the present invention will be described,

在PCI总线之上是本设备的控制平面,包括CPU, RAM存储器,ROM 存储器,FLASH存储器,带外以太网端口PHY等等。 On the PCI bus is the control plane of this device, including CPU, RAM memory, ROM memory, FLASH memory, and the like band Ethernet PHY.

在PCI总线之下的是本设备的数据平面,包括以太网交换芯片,RAM存储器,PHY接口芯片,FPGA, El收发器等等。 Under the PCI bus is the data plane of the equipment, including Ethernet switches, the RAM memory, a PHY interface chip, FPGA, El transceiver and the like.

从E1收发器进入的E1码流在FPGA完成向以太网数据包的映射,包括实现FEC功能。 Entering from the code stream E1 E1 transceiver to complete the mapping Ethernet packets in FPGA, including implement FEC functionality. 经过以太网交换芯片再从指定的以太网口进入以太网网络中, 再经过网络到达另一端。 After re-entry from the Ethernet switch port specified Ethernet Ethernet networks, through the network to the other end.

在另一端,El数据包和FEC包从以太网口进入交换芯片,再到达FPGA, 经FPGA处理后还原成E1码流,再从E1收发器发送出去。 At the other end, El and FEC packets from the packet enters the Ethernet switching chip port, and then reaches the FPGA, after reduction of E1 FPGA processing stream, and then transmitted from the transceiver E1.

图IO是本发明的功能结构描述。 FIG IO is a functional structure of the present invention will be described.

发送端带有发送緩冲区。 Sending end with a transmit buffer.

接收端带有去抖緩冲区和纠错緩冲区。 A receiver equipped with an error correction buffer and the buffer debounce.

中间传输的有普通数据包和FEC包。 Intermediate transfer ordinary data packets and FEC packets.

图11是本发明的主要数据流描述。 FIG 11 is a main data stream described in the present invention.

实现箭头表示功能结构中个功能模块的处理顺序。 The arrow indicates the processing sequence implemented in the functional configuration of functional modules.

虛线部分表示El信息在整个系统中的流向。 El dotted line represents the flow of information throughout the system.

发送时,E1信息从上端的E1接口进入,流经需要的功能模块后,变换成以太网数据包,再从下端的以太网口发送出去。 Sending, from the information E1 E1 interface into the upper end, after passing through the functional modules required, converted into Ethernet packets, and then sent out from the lower end of the Ethernet port.

接收时,携带E1信息的数据包从下端的以太网口进入,流经需要的功能模块后,从以太网数据包还原E1信息,再从上端的El接口发送出去。 Upon receipt, the data packets carrying information E1 from the lower end into the Ethernet port after flowing through the functional modules required to restore information from the E1 Ethernet packets, and then sent out from the upper end of the El interface.

测试设计如图23所示。 Test design shown in Figure 23.

从中间的M8000/T设备中的两个千兆以太网端口输入背景流量,以产生一定程度的拥塞,流量从Smartbits仪表中输入,可控制背景流量的大小。 From the middle of the M8000 / T Gigabit Ethernet devices two input ports background traffic, to a certain degree of congestion, traffic input from Smartbits instrument, to control the size of background traffic.

三台设备用千兆以太网口互联,中间的设备使用了两个千兆以太网口, 而两边的设备则个使用了一个千兆以太网口。 Gigabit Ethernet ports with three devices interconnected, the intermediate device uses two Gigabit Ethernet ports, and is a device used on both sides of a Gigabit Ethernet port.

El从误码仪输入并且从另一端输出到误码仪。 El outputted from the error detector and the input from the other end to the BER tester. 两路话机通过协议转换器连接到设备中两端的El接口。 Two-way telephone connection to the ends of the El interface device by a protocol converter.

首先测试再禁用FEC机制的情况下,El在以太网上传输的端到端性能。 The case of the first test and then disable FEC mechanism, El transmission over Ethernet-end performance. 然后测试再在使用FEC机制的情况下,El在以太网上传输的端到端性 Then tested again in the case of end-use FEC mechanism, El transmitted on the Ethernet

测试结果如下表所示:<table>table see original document page 37</column></row> <table>总结 The test results shown in the following table: <table> table see original document page 37 </ column> </ row> <table> summary

El在当前的网络中占有相当大的比重,其代表技术SDH已有几十亿美元的应用规模。 El occupies a large proportion in the current network, which represents the size of the existing technology SDH applications billions of dollars.

但现有网络向分组网络演进是大势所趋。 However, the existing network to a packet network evolution is the trend. 分组网传送E1也是必然和必要的。 Packet network transmission E1 is inevitable and necessary.

分组网对QOS支较弱与El对QOS的高要求是一个难以调和的矛盾。 El packet network is weak and the high requirements for QOS is difficult to reconcile a contradiction QOS support. This

个矛盾严重阻碍了分组网传送E1的发展。 Paradox seriously hindered the development of packet network transmission E1.

本发明致力于在利用FEC技术改善分组网传送端到端E1的性能表现。 The present invention is directed to end E1 transmission performance in improving the packet network using FEC techniques. 在网络QOS不能保证的情况下,利用两端的FEC功能,对网络引起的数据包丢失进行纠错和恢复。 In the case of network QOS can not be guaranteed, the use of both ends of the FEC function, the network packet loss due to error correction and recovery. 从而确保,即使在网络性能较差的情况下,也能完成E1在分组网上的正确传输。 Thereby ensuring that, even in the case of poor network performance, it can be completed correctly transmit E1 in the packet network.

同时,本发明确保能与不具有FEC机制的同类技术兼容。 Meanwhile, the present invention ensures compatibility with similar technology having no FEC mechanism. 换句话说,具有FEC机制的接口可以与不具有FEC机制的同类接口互通互连。 In other words, the interface FEC-FEC mechanism may have similar mechanisms and without interconnection and interfaces.

在本发明的实施和应用过程中,测试表明,本发明的FEC机制具有良好的性能。 In the embodiment and application of the present invention, tests show, the FEC mechanism of the present invention have good performance.

尽管参考本发明的优选实施例具体展示和描述了本发明,但是本领域一般技术人员应该明白,在不脱离所附权利要求限定的本发明的原理和范围内的情况下,可以对其进行形式和细节上的具体修改。 Although reference to the preferred embodiment of the present invention has been particularly shown and described embodiment of the present invention, those skilled in the art should understand that in the case of the principles and scope of the invention without departing from the appended claims may be made in the form of and specific modifications of detail.

Claims (32)

  1. 1.一种采用前向纠错FEC机制在分组网上实现可靠的端到端E1传送的方法,其特征在于,该机制的实现包含以下步骤: A.在E1进入分组网的源节点一端,将E1封装成一个数据包序列,并根据这个数据包序列生成一个FEC数据包组; B.将这个数据包序列和FEC数据包组通过分组网发送到E1离开分组网的目的节点一端; C.在所述目的节点,如果存在数据包丢失,则根据FEC包组和数据包序列再生丢失的数据包,然后提取或再生E1净载荷并还原E1码流。 A former enables robust forward error correction FEC packet network end E1 mechanism of transmission, wherein, to realize the mechanism comprises the steps of: A. in the packet network into the E1 end the source node, the E1 encapsulated into a sequence of packets, and generates an FEC packet groups according to the packet sequence; B. this data packet sequence and the FEC data packet group to the destination end node E1 away from the packet network by packet network; in C. the destination node, if there is packet loss, the FEC packet group according to the data packet sequence, and reproducing the lost packet, then extract or reproducing payload and reduced E1 E1 stream.
  2. 2. 根据权利要求1所述的分组网上实现可靠的端到端El传送的方法, 其特征在于,能够后向兼容不含FEC机制的同类技术,具有FEC机制的源节点可以和传统的目的节点互通互联。 The packet network according to a reliable end-implemented method as claimed in claim El transmission, wherein, after the same technology can be free to FEC mechanism is compatible with the source node and the FEC mechanism can be a conventional destination node interconnection.
  3. 3. 根据权利要求2所述的分组网上实现可靠的端到端El传送的方法, 其特征在于,以自协商的方式确定目的节点是否具备相同的FEC机制,如目的节点不具备相同的FEC机制,则将源节点自动设置成非FEC工作模式。 3. A packet network according to claim 2 implemented method according to El-end reliable transmission, wherein, in a self-negotiation to determine whether the destination node have the same FEC mechanism, such as the destination node does not have the same FEC mechanism , the source node will be automatically set to the non FEC operation mode.
  4. 4. 根据权利要求3所述的分组网上实现可靠的端到端El传送的方法, 其特征在于,在分组网上传送的是时分复用TDM的一次群,速率为2.048Mbps 。 Packet network according to claim 3 implemented method according to El-end reliable transmission, wherein the packet network is transmitted in a time division multiplexed TDM groups of a rate of 2.048Mbps.
  5. 5. 根据权利要求4所述的分组网上实现可靠的端到端El传送的方法, 其特征在于,分组网的封装层采用了实时RTP协议,RTP协议用以携带映射的数据包的时戳信息。 5. A packet network according to claim 4, wherein a reliable end-implemented method of El transmitted, characterized in that the encapsulation layer packet network using a real-time RTP protocol, the RTP protocol is used when the data packet carrying the timestamp information is mapped .
  6. 6. 根据权利要求5所述的分组网上实现可靠的端到端El传送的方法, 其特征在于,分组网的封装层的控制字的FORMID采用扩充定义,并用以标识FEC机制使用的规则。 6. A packet network according to claim 5, wherein a reliable end-implemented method of El transmitted, wherein the control word encapsulation layer packet network using expansion FORMID defined, and the rules for identifying the FEC mechanism used.
  7. 7. 根据权利要求6所述的分组网上实现可靠的端到端El传送的方法, 其特征在于,分组网的封装层的控制字的长度字段用以标识远端测量的单向端到端丟包率。 The packet network according to claim 6 reliable end-implemented method of El transmitted, wherein the control word packet network encapsulation layer length field to identify the remote end measurement unidirectional lost package rates.
  8. 8. 根据权利要求7所述的分组网上实现可靠的端到端El传送的方法, 其特征在于,分组网的封装层的控制字的序列号定义成二维结构:主序列号和次序列号两级。 Packet network according to claim 7 implemented method according to El-end reliable transmission, wherein the control word packet network encapsulation layer sequence number of two-dimensional structure is defined as: the primary and the secondary sequence number SEQ ID NO levels.
  9. 9. 根据权利要求8所述的分组网上实现可靠的端到端El传送的方法, 其特征在于,分组网的分组交换网络PSN层采用IP协议和帧格式。 Packet network according to claim 8 implemented method according to El-end reliable transmission, wherein the packet-switched network PSN to the packet network layer using the IP protocol and frame format.
  10. 10. 根据权利要求9所述的分组网上实现可靠的端到端El传送的方法, 其特征在于,分组网的数据链路层采用以太网链路层协议和帧格式。 10. A packet network according to claim 9 implemented method according to El-end reliable transmission, wherein the packet network using an Ethernet data link layer and link layer protocol frame format.
  11. 11. 根据权利要求10所述的分组网上实现可靠的端到端E1传送的方法, 其特征在于,设计一个32位业务时钟计数器对本地业务时钟计数,同时设计一个32位参考时钟计数器对外部参考时钟计数。 11. The packet network according to claim 10 a reliable end-implemented method of E1 transmission, wherein a design service clock counter 32 counts the clock to the local service, while a 32-bit reference design of the external reference clock counter clock count.
  12. 12. 根据权利要求11所述的分组网上实现可靠的端到端E1传送的方法, 其特征在于,映射数据包时,将E1码流按时隙分组,当一个完整的时隙进入緩冲区后,记录业务时钟计数器和参考时钟计数器的值以用于生成本时隙的时戳。 Claim 12. Once a complete packet slot line 11 enter the buffer requirements of the end-implemented method of E1 reliable transmission, wherein, when the mapping data packet, the packet time slots E1 stream, recording the value of the service clock and the reference clock counter to timestamp the time counter used to produce the slots.
  13. 13. 根据权利要求12所述的分组网上实现可靠的端到端E1传送的方法, 其特征在于,所述A步骤还进一步包括以下子步骤:映射数据包时,时隙组映射生成的所有的数据包中RTP字段需携带确定的时戳信息,如果配置成自适应时钟模式,则时戳信息依据记录的业务时钟计数值生成;如果配置成差分时钟恢复方式,则时戳信息依据记录的业务时钟计数值和外部参考时钟计数值,并根据一定规则生成。 When all of the map data packet, the map generation slot group: 13. The packet network according to claim 12 a reliable end E1 transmission method, wherein said step A further comprises the substeps of when RTP packet timestamp field determining information required to carry, if configured to adaptively clock mode, the timestamp information generated based on business value records clock count; if configured as a differential clock recovery mode, based on the timestamp information recorded traffic clock count value and the external reference clock count value, and generated according to a certain rule.
  14. 14. 根据权利要求13所述的分组网上实现可靠的端到端El传送的方法, 其特征在于,所述A步骤还进一步包括以下子步骤:将一组E1时隙集中映射成分组数据包,并且数据包头携带一组E1时隙包含的时隙数k等信息,其中El时隙组中包含的El时隙的个数可配置成8, 16或32。 Packet network according to claim 13 implemented method according to El-end reliable transmission, wherein said step A further comprises the substeps of: focus mapping a set of data packets component group E1 timeslot, and the packet header carries a group of information slots E1 timeslot number k and the like contained in, wherein the number of slots contained in the group El El timeslot may be configured to 8, 16 or 32.
  15. 15. 根据权利要求14所述的分组网上实现可靠的端到端E1传送的方法, 其特征在于,所述A步骤还进一步包括以下子步骤:将一组El时隙集中映射成分组数据包时,可以配置使用每个时隙独立映射成包的独立映射方式,也可选择从一组El时隙中的每个时隙选取部分数据,再封装成数据包的合成映射方式。 When the focus mapping component group El slot packet group: a packet network according to claim 14 implements a method according to a reliable end to end transmission of E1, wherein said step A further comprises the substeps of , each slot can be configured using separate mapping mode independently mapped into packets, also choose from a set of selected portions of the data in the El time slots each time slot, and then packaged into a composite data packet mapping method.
  16. 16. 根据权利要求15所述的分组网上实现可靠的端到端E1传送的方法, 其特征在于,所述A步骤还进一步包括以下子步骤:将一组E1时隙集中映射成分组数据包时,使用合成映射方式时,从每个时隙的每一字节选择对应的位拼成字节或字,再组成帧,这种映射方法可以用于对信息安全要求较高的应用。 When the focus mapping component group of data packets E1 timeslot group: a packet network according to claim 15, said end-implemented method of E1 reliable transmission, wherein said step A further comprises the substeps of when using synthetic mapping mode, selects a corresponding bit from each byte of each slot makes up a byte or word, then the composition of the frame, this method may be used for mapping high information security requirements of the application.
  17. 17. 根据权利要求16所述的分组网上实现可靠的端到端E1传送的方法, 其特征在于,所述A步骤还进一步包括以下子步骤:将一组El时隙集中映射成分组数据包后,根据生成的分組数据包序列生成FEC包组。 After the focus mapping component group El slot packet group: a packet network according to claim 16, said end-implemented method of E1 reliable transmission, wherein said step A further comprises the substeps of generates a set of FEC packets according to the packet sequence of packets generated.
  18. 18. 根据权利要求17所述的分组网上实现可靠的端到端E1传送的方法, 其特征在于,所述A步骤还进一步包括以下子步骤:生成FEC包组,采用Reed-Solomon (n, k)编码。 18. The packet network according to claim 17 E1 reliable end to end transmission method, wherein said step A further comprises the substeps of: generating a set of FEC packets using Reed-Solomon (n, k )coding.
  19. 19. 根据权利要求18所述的分组网上实现可靠的端到端E1传送的方法, 其特征在于,所述A步骤还进一步包括以下子步骤:生成FEC包组,采用Reed-Solomon (n, k)编码时,k可配置成8, 16 或32, n可通过自动协商设置成k + 4, k+8, k+16, k+32等值,且数据包头中包含配置信息:n和k的值。 19. The reliable packet network A method according to claim 18 E1 transmission end, wherein said step A further comprises the substeps of: generating a set of FEC packets using Reed-Solomon (n, k ) encoding, k may be configured to 8, 16 or 32, n may be to k + 4, k 8, k + 16, k + 32 + equivalence set by auto negotiation, and the packet header contains the configuration information: n and k value.
  20. 20. 根据权利要求19所述的分组网上实现可靠的端到端E1传送的方法, 其特征在于,所述A步骤还进一步包括以下子步骤:生成FEC包组,将包含k个时隙的时隙组中每个时隙的第i字节组合在一起得到k字节序列,并根据生成多项式生成nk个字节的纠错码,分别作为第1,2...nk个FEC包的第i字节。 Generating FEC packets group, k slots comprising: 20. The packet network according to claim 19 a reliable end-implemented method of E1 transmission, wherein said step A further comprises the substeps of i-th byte slot group for each slot together to provide a combination of a byte sequence k, and nk bytes of polynomials based on the generated error correction code, respectively, as the first ... the nk 1,2 FEC packet i bytes.
  21. 21. 根据权利要求1所述的分組网上实现可靠的端到端El传送的方法, 其特征在于,所述B步骤还进一步包括以下子步骤:发送数据包时,普通数据包的发送时间取决于数据包或其携带的时隙对应的业务时钟计数值,也就是说,时隙组内两个相邻数据包的发送时间间隔与两个相邻业务时钟计数值的差值成比例。 21. The packet network according to a reliable transmission method as claimed in claim end El, wherein said step B further comprises the sub-steps of: sending a packet, the packet transmission time depends on the normal or a packet service corresponding to the slot carried by clock count value, that is, transmission time slots within the groups of two adjacent packets interval proportional to the difference clock count value to the service of two adjacent.
  22. 22. 根据权利要求21所述的分组网上实现可靠的端到端E1传送的方法, 其特征在于,所述B步骤还进一步包括以下子步骤:发送数据包时,在两个普通数据包的发送间隔内,发送FEC数据包,每个时间间隔发送的FEC数据包个数,根据n和k的取值确定,发送FEC数据包的时间确定原则为:尽量使相邻数据包发送的时间间隔均匀。 22. The packet network according to claim 21 a reliable end-implemented method of E1 transmission, wherein said step B further comprises the sub-steps of: sending data packets sent in two packets of normal interval, FEC packet transmission, the number of packets transmitted in each time interval FEC, according to the values ​​of n and k is determined, the FEC data packet transmission time determination principle: try to make the time interval of adjacent data packets transmitted uniformly .
  23. 23. 根据权利要求10所述的分组网上实现可靠的端到端E1传送的方法, 其特征在于,所述C步骤还进一步包括以下子步骤:在接收端为控制时延设置接收定时器,溢出周期等于k*125us。 23. The packet network according to claim 10 a reliable end-implemented method of E1 transmission, wherein said step C further comprises the sub-steps of: receiving end disposed to receive control delay timer overflow period is equal to k * 125us.
  24. 24. 根据权利要求23所述的分组网上实现可靠的端到端El传送的方法, 其特征在于,所述C步骤还进一步包括以下子步骤:将接收到的数据包去掉数据包头后根据包序号存入去抖緩沖区, 如杲配置成基于时戳的自适应恢复模式,则需要将数据包携带的RTP协议中的时戳信息存入緩冲区的远端时戳存储空间,同时利用本地业务时钟生成时戳,并将这个时戳存入緩冲区的本地时戳存储空间;如果配置成差分时钟恢复模式,则以这个时戳结合外部时钟得到本地业务时钟。 24. The packet network according to claim 23 a reliable end-implemented method of El transmitted, wherein said step C further comprises the sub-steps of: receiving a data packet after removing the packet headers according to the packet number timestamp stored in the storage space debounce buffer, such as adaptive Gao arranged to time stamp based on the recovery mode, it is necessary to time stamp information carried in RTP packets that are stored in the distal end of the buffer, while the use of local a clock generation timestamp service, and time stamp when the stamp into the storage space when the local buffer; if configured as a differential clock recovery mode, places the external clock timestamp combined to give the local service clock.
  25. 25. 根据权利要求24所述的分组网上实现可靠的端到端El传送的方法, 其特征在于,所述C步骤还进一步包括以下子步骤:将接收到的FEC包去掉数据包头后根据包序号存入纠错緩冲区。 25. The packet network according to claim 24 a reliable end-implemented method of El transmitted, wherein said step C further comprises the sub-steps of: receiving FEC packets based on the packet header after removal No. error correction into the buffer.
  26. 26. 根据权利要求25所述的分组网上实现可靠的端到端E1传送的方法, 其特征在于,所述C步骤还进一步包括以下子步骤:当接收定时器溢出时,检查去抖緩冲区的数据,如存在信息丢失,则利用纠错信息再生丟失的信息。 26. The packet network according to claim 25 a reliable end-implemented method of E1 transmission, wherein said step C further comprises the sub-steps of: when receiving the timer overflow check debounce buffer data, such as the presence information is lost, the error correction information using the information reproducing lost.
  27. 27. 根据权利要求26所述的分组网上实现可靠的端到端El传送的方法, 其特征在于,所述C步骤还进一步包括以下子步骤:当接收定时器溢出时,丢失信息恢复完成后,则进行时钟恢复,并且时钟恢复的方法可以根据是否具有同一参考时钟,配置成自适应方式或差分方式。 27. The packet network according to claim 26 a reliable end-implemented method of El transmitted, wherein said step C further comprises the sub-steps of: when receiving the timer overflow, information loss after recovery is complete, the clock recovery, clock recovery and the method may have the same depending on whether the reference clock and configured to adaptively or differential mode.
  28. 28. 根据权利要求27所述的分组网上实现可靠的端到端E1传送的方法, 其特征在于,所述C步骤还进一步包括以下子步骤:采用自适应方式恢复时钟时,可以配置成基于去抖緩冲区的填充级或基于RTP协议携带的时戳两种模式。 28. The packet network according to claim 27 a reliable end-implemented method of E1 transmission, wherein said step C further comprises the sub-steps of: when an adaptive clock recovery mode, may be configured to based poke two modes shaking buffer fill level or RTP protocols carried on.
  29. 29. 根据权利要求28所述的分组网上实现可靠的端到端E1传送的方法, 其特征在于,所述C步骤还进一步包括以下子步骤:采用自适应方式恢复时钟时,可以配置恢复的业务时钟精度范围。 29. The packet network according to claim 28 a reliable end-implemented method of E1 transmission, wherein said step C further comprises the sub-steps of: when an adaptive clock recovery mode, may be configured to restore service clock precision.
  30. 30. 根据权利要求29所述的分组网上实现可靠的端到端El传送的方法, 其特征在于,所述C步骤还进一步包括以下子步骤:如果配置成自适应方式恢复时钟时,且配置成使用緩沖填充级时,则根据緩沖区的使用情况恢复本地业务时钟; 如果配置成自适应方式恢复时钟时,且配置成使用时戳时,则根据緩冲区存储的时戳信息恢复本地业务时钟。 30. The packet network according to claim 29 a reliable end-implemented method of El transmitted, wherein said step C further comprises the sub-steps of: if the recovered clock adaptively configured and arranged to when using a buffer fill level, the local service clock recovery based on the use of a buffer; if the recovered clock adaptively configured and arranged time stamp, the time stamp information according to the recovery clock when the local service time to use the buffer memory .
  31. 31. 根据权利要求30所述的分组网上实现可靠的端到端E1传送的方法, 其特征在于,所述C步骤还进一步包括以下子步骤:时钟恢复完成后,依据业务时钟,结合緩冲区的数据恢复E1信息流。 31. The packet network according to claim 30 a reliable end-implemented method of E1 transmission, wherein said step C further comprises the sub-steps of: after completion of clock recovery, according to the service clock, binding buffer data recovery E1 traffic.
  32. 32. —种应用权利要求1-31中任一项所述的方法以在分组网上实现可靠的端到端E1传送的设备。 32. - Method according to any one of claims 1-31 in order to achieve a reliable packet network transmission apparatus of claim end E1 of applications.
CN 200610163606 2006-12-01 2006-12-01 Method for reliable E1 transmission based on forward error correction mechanism in packet network CN101193060B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 200610163606 CN101193060B (en) 2006-12-01 2006-12-01 Method for reliable E1 transmission based on forward error correction mechanism in packet network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 200610163606 CN101193060B (en) 2006-12-01 2006-12-01 Method for reliable E1 transmission based on forward error correction mechanism in packet network

Publications (2)

Publication Number Publication Date
CN101193060A true true CN101193060A (en) 2008-06-04
CN101193060B CN101193060B (en) 2010-09-01

Family

ID=39487821

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 200610163606 CN101193060B (en) 2006-12-01 2006-12-01 Method for reliable E1 transmission based on forward error correction mechanism in packet network

Country Status (1)

Country Link
CN (1) CN101193060B (en)

Cited By (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102137112A (en) * 2011-04-20 2011-07-27 中兴通讯股份有限公司 Method and device for converting media streams
CN102377678A (en) * 2011-11-28 2012-03-14 瑞斯康达科技发展股份有限公司 Data transmission and processing method and device
CN101511042B (en) 2009-03-23 2012-05-16 中兴通讯股份有限公司 Data synchronization method and device
CN103036958A (en) * 2012-12-06 2013-04-10 陕西维德科技股份有限公司 Server cluster high-speed memory system
CN101877620B (en) 2009-04-30 2013-08-14 华为技术有限公司 Method, apparatus and system for forward error correction
CN103650399A (en) * 2011-07-14 2014-03-19 微软公司 Adaptive generation of correction data units
US8711695B2 (en) 2008-12-19 2014-04-29 Panasonic Corporation Transmitter and receiver for streaming media including packet padding
CN103957023A (en) * 2014-05-16 2014-07-30 江苏新瑞峰信息科技有限公司 Device and method for transceiving of 3G signals
CN104081797A (en) * 2012-07-10 2014-10-01 英特尔公司 Network system configured for resolving forward error correction during a data mode
CN104796735A (en) * 2014-01-17 2015-07-22 中国移动通信集团公司 Video data transmission method and device
CN105392160A (en) * 2015-10-08 2016-03-09 江苏天智互联科技股份有限公司 Embedded e-commerce system and interactive method thereof
CN105843772A (en) * 2016-04-12 2016-08-10 华为技术有限公司 Device and method for processing data
WO2018014588A1 (en) * 2016-07-19 2018-01-25 华为技术有限公司 Data transmission method and apparatus

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102684806A (en) * 2012-05-10 2012-09-19 Ut斯达康通讯有限公司 High-accuracy self-adaptive clock recovery method of packet transport network (PTN) simulation service

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2001086153A (en) 1999-09-09 2001-03-30 Canon Inc Data communication equipment, data communication system, data communication method and storage medium
CN1256832C (en) 2002-09-13 2006-05-17 北京威速科技有限公司 Forward error correction in internet multimedia real time communication
CN1216476C (en) 2002-11-14 2005-08-24 北京润光泰力科技发展有限公司 Data interlaced device and method
JP2006217530A (en) 2005-02-07 2006-08-17 Mitsubishi Electric Corp Data transmission system, transmitting side terminal device, receiving side terminal device, data transmission program, and data reception program
CN100450187C (en) 2005-10-17 2009-01-07 华为技术有限公司 Multimedia data network realtime transfer method for supporting error elasticity

Cited By (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102257775B (en) 2008-12-19 2014-10-22 松下电器产业株式会社 Transmitter terminal and a receiver terminal
US8711695B2 (en) 2008-12-19 2014-04-29 Panasonic Corporation Transmitter and receiver for streaming media including packet padding
CN101511042B (en) 2009-03-23 2012-05-16 中兴通讯股份有限公司 Data synchronization method and device
CN101877620B (en) 2009-04-30 2013-08-14 华为技术有限公司 Method, apparatus and system for forward error correction
CN102137112A (en) * 2011-04-20 2011-07-27 中兴通讯股份有限公司 Method and device for converting media streams
CN103650399B (en) * 2011-07-14 2017-01-18 微软技术许可有限责任公司 Adaptive correction data generating means
CN103650399A (en) * 2011-07-14 2014-03-19 微软公司 Adaptive generation of correction data units
CN102377678B (en) * 2011-11-28 2014-12-10 瑞斯康达科技发展股份有限公司 Data transmission and processing method and device
CN102377678A (en) * 2011-11-28 2012-03-14 瑞斯康达科技发展股份有限公司 Data transmission and processing method and device
CN104081797A (en) * 2012-07-10 2014-10-01 英特尔公司 Network system configured for resolving forward error correction during a data mode
CN104081797B (en) * 2012-07-10 2018-07-31 英特尔公司 It is configured for analyzing a network system for forward error correction before during the data mode
CN103036958B (en) * 2012-12-06 2016-02-24 陕西维德科技股份有限公司 A server cluster high-speed storage system
CN103036958A (en) * 2012-12-06 2013-04-10 陕西维德科技股份有限公司 Server cluster high-speed memory system
CN104796735A (en) * 2014-01-17 2015-07-22 中国移动通信集团公司 Video data transmission method and device
CN104796735B (en) * 2014-01-17 2018-06-15 中国移动通信集团公司 Kinds of video data transmission method and apparatus
CN103957023A (en) * 2014-05-16 2014-07-30 江苏新瑞峰信息科技有限公司 Device and method for transceiving of 3G signals
CN105392160A (en) * 2015-10-08 2016-03-09 江苏天智互联科技股份有限公司 Embedded e-commerce system and interactive method thereof
CN105843772A (en) * 2016-04-12 2016-08-10 华为技术有限公司 Device and method for processing data
WO2018014588A1 (en) * 2016-07-19 2018-01-25 华为技术有限公司 Data transmission method and apparatus

Also Published As

Publication number Publication date Type
CN101193060B (en) 2010-09-01 grant

Similar Documents

Publication Publication Date Title
US6331978B1 (en) Generic label encapsulation protocol for carrying label switched packets over serial links
US7042848B2 (en) System and method for hierarchical policing of flows and subflows of a data stream
US20040252688A1 (en) Routing packets in frame-based data communication networks
US20080198755A1 (en) Probing-based mechanism to reduce preemption perturbation caused by higher priority tunnel establishment in a computer network
US6801530B1 (en) Communication system and method in a communication system
US6594279B1 (en) Method and apparatus for transporting IP datagrams over synchronous optical networks at guaranteed quality of service
US20060153179A1 (en) Techniques for transmitting and receiving traffic over advanced switching compatible switch fabrics
US7433301B2 (en) Method of transferring packets and router device therefor
US20030021287A1 (en) Communicating data between TDM and packet based networks
US20040085904A1 (en) Method for flow control of packets aggregated from multiple logical ports over a transport link
US20040184450A1 (en) Method and system for transport and routing of packets over frame-based networks
US6842455B1 (en) Conversion of contiguous concatenation of virtual concatenation in a synchronous digital communication network
US20070097865A1 (en) Method for measuring end-to-end delay in asynchronous packet transfer network, and asynchronous packet transmitter and receiver
US20070189327A1 (en) Apparatus and method for increasing reliability of data sensitive to packet loss
Jha et al. Engineering Internet QoS
US20050174941A1 (en) Methods and apparatus for controlling the flow of multiple signal sources over a single full duplex ethernet link
US6819679B1 (en) Multiprotocol packet framing technique
US20030133461A1 (en) Circuit emulation service over an internet protocol network
US20090003383A1 (en) Node Device, Packet Control Device, Radio Communication Device, and Transmission Control Method
CN101719843A (en) Method of LSP linear protection switching in PTN
US20060140226A1 (en) Techniques for processing traffic transmitted over advanced switching compatible switch fabrics
CN102158371A (en) Link performance detection method and device based on BFD (Bidirectional Forwarding Detection)
US6931025B1 (en) Optical adaptation layer frame for aggregating IP flows at edge nodes for sonet transport
CN1635754A (en) An Ethernet transmission equipment and method having characteristics of rapid protection and fairness
US20040022268A1 (en) Method and apparatus for end-to-end connection between an RPR and an MPLS network

Legal Events

Date Code Title Description
C06 Publication
C10 Entry into substantive examination
C14 Grant of patent or utility model
TR01