WO2011050542A1 - Router constructing method supporting hierarchy of video quality - Google Patents

Router constructing method supporting hierarchy of video quality Download PDF

Info

Publication number
WO2011050542A1
WO2011050542A1 PCT/CN2009/074738 CN2009074738W WO2011050542A1 WO 2011050542 A1 WO2011050542 A1 WO 2011050542A1 CN 2009074738 W CN2009074738 W CN 2009074738W WO 2011050542 A1 WO2011050542 A1 WO 2011050542A1
Authority
WO
WIPO (PCT)
Prior art keywords
cells
cell
output
input
line card
Prior art date
Application number
PCT/CN2009/074738
Other languages
French (fr)
Chinese (zh)
Inventor
李挥
李硕彦
李锐源
庄少然
李烽
陈曦
陈钦树
李亦宁
潘凯
Original Assignee
深圳市利德嘉实业有限公司
北京大学深圳研究生院
上海北京大学微电子研究院
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
Application filed by 深圳市利德嘉实业有限公司, 北京大学深圳研究生院, 上海北京大学微电子研究院 filed Critical 深圳市利德嘉实业有限公司
Priority to PCT/CN2009/074738 priority Critical patent/WO2011050542A1/en
Publication of WO2011050542A1 publication Critical patent/WO2011050542A1/en

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00Routing or path finding of packets in data switching networks
    • H04L45/60Router architectures

Definitions

  • the present invention relates to the field of communications, and more particularly to a router construction method for supporting video quality grading composed of a plurality of switching units for multi-level interconnection, which is proposed for the current high-speed video application and video grading coding technology of the Internet.
  • Video grading coding refers to the adjustability of the code rate, that is, the video data is compressed only once, but Multiple frame rates, spatial resolutions, or video quality are decoded to support a variety of different application requirements for multiple types of users.
  • MPEG-4 implements Temporal Scalability through the video object layer (VOL, Video Object Layer) data structure, Spatial Scalability and mixed grading of support time and airspace.
  • VOL Video Object Layer
  • Each hierarchical code has at least two layers of VOL, the lower layer is called the base layer, and the upper layer is called the enhancement layer.
  • the base layer provides basic information about the video sequence
  • the enhancement layer provides a higher resolution of the video sequence.
  • the present invention provides a router system and a construction method for supporting video quality grading of hyper-scale streaming media.
  • a further technical solution adopted by the present invention to solve the technical problem is as follows: In the step S2, when a cell contends for an internal port or an outgoing resource, the low priority cell will be blocked and discarded without being cached.
  • step S5 the central control unit performs resource reservation control, but does not perform input/output pairing scheduling for each input/output port per cell slot.
  • a further technical solution adopted by the present invention to solve the technical problem is: the input line card determines the classification according to the header information of the newly arrived data packet, and the video coding data of different levels of the same stream is marked as different levels of cells. .
  • the packet switching structure is composed of a multi-level interconnection network partially sorted from a routing hub, and the cells are sorted according to their target output port addresses and priorities, and each The level blocks cells that exceed their line group forwarding capabilities by priority.
  • the output line card completes reassembly of the data packet, and when the cell is exchanged to the output port, the addressing bit is cancelled during the addressing process. After consumption, the cell control information mainly has the remaining active flag bits and priority flag bits, and the corresponding processor in the output line card removes the remaining control information bits and reassembles the cells into data packets.
  • a further technical solution adopted by the present invention to solve the technical problem is to provide a router system supporting video quality grading, comprising a plurality of input line cards, a plurality of output line cards, a cell switching structure, and a central control unit.
  • the input line card sets a corresponding priority according to the type of the data packet on the input port, and then divides the video code stream data packets that have been hierarchically coded into cells for exchange, and according to the video hierarchical coded stream The level is given the corresponding cell priority.
  • the packet switching structure performs routing and forwarding according to the output destination address of the cell, and the output line card reassembles the cell to form a data packet and then continue to be forwarded by the line rate or sent by the link.
  • a further technical solution adopted by the present invention to solve the technical problem is: the input line card determines the classification according to the header information of the newly arrived data packet, and the video coding data of different levels of the same stream is marked as different levels of cells. .
  • the packet switching structure may be partially ordered from a multi-level interconnection network of a routing hub, and the cells are sorted according to their target output port addresses and priorities, and each The level blocks cells that exceed their line group forwarding capabilities by priority.
  • the output line card completes the reassembly of the IP data packet, and when the cell is exchanged to the output port, the addressing bit is consumed in the addressing process, and the cell control
  • the information mainly has active flag bits and priority flag bits, and the corresponding processor in the output line card removes the remaining control information bits and reassembles the cells into data packets.
  • the router system and the construction method supporting the video quality classification of the invention can realize a super-large-scale streaming media router structure, and the number of expandable ports is N>10000, and the bandwidth per port is lOOMbits per second.
  • the streaming media that supports video hierarchical coding immediately discards the advanced enhanced coding portion of the media stream when the burst load traffic overloads the router switch fabric, while completely retaining the lower level basic level code stream. Quasi-circuit-switched architecture, no queuing delay, zero jitter. Access control is only allowed for newly arrived flows, but not with matching scheduling per slot.
  • the cells are distributed and self-routing, and there is no resource bottleneck.
  • Figure 1 is a block diagram of a packet-cutting cell of a router construction method supporting video quality classification of the present invention.
  • FIG. 2 is an 8x8 network of a router construction method capable of supporting video quality classification according to the present invention.
  • 3 is a schematic diagram of cell reassembly of a router construction method supporting video quality grading according to the present invention. detailed description
  • S1 setting a priority according to the type of the data packet at the input line card, and dividing the data packets of the video code stream that has been hierarchically coded into the cells for packet switching, according to the video hierarchical coded stream The level of the corresponding cell priority encoding;
  • the packet switching structure performs routing and forwarding according to the output destination address of the cell
  • S5 The central control unit performs access control for long-lasting streams.
  • step S2 when the cell contends for the internal port or the outgoing resource, the low priority cell will be blocked and discarded without being cached.
  • the central control unit performs resource reservation control, but does not perform input/output pairing scheduling for each input/output port per cell slot.
  • the input line card determines its classification according to the header information of the newly arrived data packet, and the video coding data of different levels of the same stream is marked as cells of different levels.
  • the packet switching structure is composed of a multi-level interconnection network partially ordered from a routing hub, the cells are sorted according to their target output port addresses and priorities, and the packets exceeding the forwarding capability of the line group are blocked by priority at each level. yuan.
  • the router input line card determines its classification based on the header related information of the newly arrived packet.
  • the output line card completes the reassembly of the IP data packet.
  • the addressing bit is consumed in the addressing process, and the cell control information mainly has the active flag bit and the priority flag bit.
  • the corresponding processor in the output line card removes these remaining control information bits and reassembles the cells into packets.
  • the present invention provides a router system that supports video quality grading, comprising a number of input line cards, a number of output line cards, a cell switching structure, and a central control unit.
  • the input line card sets a corresponding priority according to the type of the data packet on the input port, and the video stream data packet that has been hierarchically encoded is divided into video stream coded streams when they are divided into the cells for exchange.
  • the level is given to the corresponding cell priority, and the packet switching structure performs routing and forwarding according to the output destination address of the cell.
  • the output line card reassembles the cells to form a data packet and continues the line rate forwarding or transmission by the link.
  • the main function of the input line card is to set the corresponding priority according to the type of the data packet at the input port.
  • the video stream packets that have been hierarchically encoded are further divided into cells for exchange, and according to which The level of the video grading code stream is given the corresponding cell priority.
  • the input line card can divide a packet into a plurality of small cells, and add m+n+1 bits of control information Ipm-l ... p0dld2 before each input cell.
  • Dn where the first bit is the active bit, pm- ⁇ ... ⁇ is the priority bit, and dld2...dn is the addressing bit.
  • the packet priority is set to level 8.
  • the group priority list is as follows:
  • the proportion of high-priority data packets in the network is extremely low, and the occupied bandwidth is negligible.
  • Streaming media occupies the largest traffic in network transmission.
  • the basic layer data and audio data of the video are the most important information.
  • the biggest feature of the streaming media information is its strong real-time performance. Therefore, when designing the streaming router, priority will be given when the bandwidth is limited. Its high priority packet is sent.
  • the video base layer and audio are set to a higher priority, which of course is to preferentially guarantee the transmission of the base layer and audio information of the video packet.
  • the user can decode a video image having a general image quality through the received video base layer. When the bandwidth is sufficient, the user can receive the enhancement layer information (El, E2, E3) of the video image within the image decoding delay range, so that a clear high quality video image can be reconstructed.
  • the network implements self-routing exchange.
  • the input signal can be sorted using a 2x2 sequencer according to the following 2-bit sequence:
  • each input signal has a routing flag consisting of one active bit and the next n address bits.
  • Each address bit represents an output selection at a particular level of routing node and is consumed when passing through a routing node of a certain level.
  • each node is a 2Rx2R hub, and its input signals are sorted according to the following 2-bit sequence: 10 ⁇ 00 ⁇ 11 , the self-routing mechanism can be used to control signal routing. .
  • a routing unit takes the first two bits from each input cell as an in-band control signal. It is assumed that an active bit is generated on an active input port, followed by an address bit and a non-control bit. The address bits indicate the choice made between the two outputs of the unit.
  • the idle packet is a string of 0 bits. Therefore, the in-band control signals for the destination address of the 0 destination address, the packet to the 1 destination address, and the idle packet are 10, 11 and 00, respectively. In this case, the basic linear ordering of the routing unit translates to:
  • the following table shows the connection status of a routing unit.
  • each packet should carry an in-band control signal as its prefix when entering the switching node, so that the switching decision of the node depends only on the in-band control signal carried by the input packet.
  • an in-band control mechanism is self-routing, and control signals for packets entering a switching node are determined only by the I/O address of the packet.
  • control signals of packets entering a switching node need to contain as few bits as possible, so that the exchange decision process can be performed very quickly.
  • the basis of this mechanism is the self-routing destination address-oriented sequence of the network: for example, the Omega network is a monotonic sequence of 1, 2, ..., n.
  • the mechanism can be generalized as follows:
  • a target address is binary (dld2... dn)
  • the grouping uses the binary control stream lp ⁇ ...pod ⁇ d ⁇ ) ... d ⁇ ) as a prefix. Therefore, the format of the entire packet entering the switching network is: In-band control signal data
  • the bit stream Before the bit stream reaches the jth stage, the data segment (1 ⁇ (1)(1 ⁇ (2)...(1 ⁇ ( ⁇ -1) has been consumed, so the first control bit of the two cells is lpsPiPodyG), j The level exchange control reads only these two bits in the packet.
  • Switching control can be described by a finite state automaton with three states: initial, parallel, and crossed states. Its initial state is arbitrary, with a tentative connection state parallel, or cross.
  • the switching unit copies the active bit and the priority bit at the active output, and respectively latches the connection state into parallel or cross according to the state of the automaton parallel or cross. After that, the successive bits are only The single unit passes through the switching unit in the latched connection state.
  • Op 2 piPoO and 0 ⁇ ' 2 ⁇ 0 outputs 0 and 0;
  • the main function of the output line card is to complete the reassembly of the data packet.
  • the addressing bit is consumed in the addressing process.
  • the cell control information mainly has the active flag bit and the priority flag bit, and the corresponding processor in the output line card removes the remaining control.
  • the information bits then reassemble the cells into packets.
  • the router system and the construction method supporting the video quality classification of the invention can realize a super-large-scale streaming media router structure, and the number of expandable ports is N>10000, and the bandwidth per port is 100 Mbits per second.
  • the streaming media supporting video hierarchical coding immediately discards the advanced enhanced coding part of the media stream when the burst load traffic overloads the router switching structure, and completely retains the level enhanced code stream.
  • Quasi-circuit-switched architecture no queuing delay, zero jitter. Access control is only allowed for newly arrived flows, but not with matching scheduling per slot.
  • the cells are distributed and self-routing, and there is no resource bottleneck.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

A router constructing method supporting hierarchy of video quality is provided in the present invention. The method includes the following steps: S1: priorities are set according to the kinds of data packets at an input line card, and for the video bit-stream data packets which have been hierarchically encoded, during they are cut into cells for packet switching, priority codes are allocated to the corresponding cells according to the levels of bit-streams encoded hierarchically by the video resolution; S2: a packet switching structure performs route forwarding according to the output destination addresses of the cells; S3: an output line card recombines the cells to form data packets, and then continue to forward them at wire-speed or transmit them by links; S4: N×N routing structure is formed, its inputs and outputs are divided into M line groups, and each line group is made up of G input or output ports, i.e. N=M×G; S5: a central control unit performs access admission control for the long-time-continued streams.

Description

支持视频质量分级的路由器构造方法  Router construction method supporting video quality grading
技术领域 Technical field
本发明涉及通信领域, 更特别地是, 针对目前互联网高速增长的视频 应用及视频分级编码技术提出的涉及用于大量交换单元经多级互连构成的 支持视频质量分级的路由器构造方法。  The present invention relates to the field of communications, and more particularly to a router construction method for supporting video quality grading composed of a plurality of switching units for multi-level interconnection, which is proposed for the current high-speed video application and video grading coding technology of the Internet.
背景技术 Background technique
互联网流量高速增长, 2006 年美国南方贝尔电话公司的首席网络架 构师 Henry Kafka在美国全国光纤通信工程师大会上表示, 目前宽带家庭 用户平均每月下载 2 吉字节 (GB )数据, 随着文件下载的流行会达到每 户每月 9 GB; 网络电视用户月均下载 224 GB数据, 而高清视频用户月均 下载约 1太字节数据 ( 1000G )。  Internet traffic is growing at a high rate. In 2006, Henry Kafka, chief network architect of BellSouth Telephone Company of the United States, said at the National Fiber Optic Communications Engineers Conference that broadband home users currently download an average of 2 gigabytes (GB) of data per month, with file downloads. The popularity will reach 9 GB per household per month; Internet TV users will download 224 GB per month, while HD video users will download about 1 terabyte of data (1000G) per month.
高速增长流量的构成背后, 是流量构成发生重大的变化。 从应用层和 传输层上来看, 2000年前后, 以万维网 HTTP访问, 电子邮件 SMTP等典 型的基于 TCP类型的流量占据约 90 %以上; 其它域名服务 DNS, 路由选 择协议 RIP, 网络管理 SNMP等典型的 UDP应用约占 5%左右; 少量网络 控制协议如 ICMP, IGMP, EGP, IGP, OPSF流量约占 3%左右。 到了今 天, 实时视频交互, 流媒体等应用开始主导网络的时候, UPD约占 50 % ; 可以预计到了网络电视, 高清视频,及 3G和后 3G的高速移动视频普及的 时代, 视频流将占据主导的地位, 就是 90 %以上的 UDP视频流量。  Behind the composition of high-speed growth traffic is a major change in the composition of the flow. From the application layer and the transport layer, around 2000, typical TCP-based traffic such as Web HTTP access, email SMTP, etc. accounted for more than 90%; other domain name services DNS, routing protocol RIP, network management SNMP, etc. The UDP application accounts for about 5%; a small number of network control protocols such as ICMP, IGMP, EGP, IGP, and OPSF traffic account for about 3%. Today, when real-time video interaction, streaming media and other applications start to dominate the network, UPD accounts for about 50%. It can be predicted that the era of Internet TV, high-definition video, and high-speed mobile video of 3G and 3G will dominate the video stream. The status is more than 90% of UDP video traffic.
总而言之, 实时及交互的音视频将是未来网络服务的主流。 数据文件 的比例在 5到 10年内将降到 10 %以内。  All in all, real-time and interactive audio and video will be the mainstream of future network services. The proportion of data files will drop to less than 10% within 5 to 10 years.
此外随着因特网业务的巨大增长, 在速率起伏 4艮大的 IP ( Internet Protocol )网络及具有不同传输特性的异构网络上进行视频传输的要求和应 用越来越多。 在这种背景下, 视频分级编码的重要性日益突出, 其应用非 常广泛, 此处视频编码的可分级性(Scalability )是指码率的可调整性, 即 视频数据只压缩一次,却能以多个帧率、空间分辨率或视频质量进行解码, 从而可支持多种类型用户的各种不同应用要求。 如 MPEG-4通过视频对象 层 (VOL, Video Object Layer )数据结构来实现的时域分级(Temporal Scalability ), 空域分级( Spatial Scalability )和支持时、 空域的混合分级。 每一种分级编码都至少有两层 VOL, 低层称为基本层, 高层称为增强层。 基本层提供了视频序列的基本信息, 增强层提供了视频序列更高的分辨率 发明内容 In addition, with the tremendous growth of Internet services, there are more and more requirements and applications for video transmission on IP network (Internet Protocol) networks with different rates and heterogeneous networks with different transmission characteristics. In this context, video grading coding is becoming more and more important, and its application is very extensive. Here, the Scalability of video coding refers to the adjustability of the code rate, that is, the video data is compressed only once, but Multiple frame rates, spatial resolutions, or video quality are decoded to support a variety of different application requirements for multiple types of users. For example, MPEG-4 implements Temporal Scalability through the video object layer (VOL, Video Object Layer) data structure, Spatial Scalability and mixed grading of support time and airspace. Each hierarchical code has at least two layers of VOL, the lower layer is called the base layer, and the upper layer is called the enhancement layer. The base layer provides basic information about the video sequence, and the enhancement layer provides a higher resolution of the video sequence.
本发明提供一种可以实现超大规模流媒体的支持视频质量分级的路由 器系统以及构造方法。  The present invention provides a router system and a construction method for supporting video quality grading of hyper-scale streaming media.
本发明解决技术问题所采用的技术方案是:提供一种支持视频质量分级 的路由器构造方法, 其包括以下步骤: S1 : 在输入线卡处根据数据包的种 类设定优先级, 对已经分级编码的视频码流数据包, 在把它们切分成用于 分组交换的信元时, 按其视频分级编码码流的级别给予对应的信元优先级 编码; S2: 分组交换结构根据信元的输出目标地址进行路由转发; S3: 输 出线卡对信元进行重组装,形成数据包后继续线速转发或由链路发送; S4: 构成 Ν χ Ν的路由结构, 其输入输出分为 Μ个线组每个, 每个线组由 G 个输入或输出端口构成, 即N = MxG; S5: 中央控制单元对长时间持续的 流进行接入允许控制。  The technical solution adopted by the present invention to solve the technical problem is to provide a router construction method for supporting video quality classification, which includes the following steps: S1: setting a priority according to the type of the data packet at the input line card, and having the hierarchical coding Video stream packets, when they are divided into cells for packet switching, are given corresponding cell priority coding according to the level of their video hierarchical code stream; S2: packet switching structure according to cell output destination The address is routed and forwarded; S3: The output line card reassembles the cell to form a data packet and continues to be forwarded by the line rate or sent by the link; S4: A routing structure constituting Ν Ν, whose input and output are divided into a line group Each line group consists of G input or output ports, ie N = MxG; S5: The central control unit performs access control for long-lasting streams.
本发明解决技术问题所采用的进一步技术方案是: 所述的步骤 S2中, 当出现信元争用内部端口或出线资源时, 低优先级的信元将被阻塞掉丟弃 而不予緩存。  A further technical solution adopted by the present invention to solve the technical problem is as follows: In the step S2, when a cell contends for an internal port or an outgoing resource, the low priority cell will be blocked and discarded without being cached.
本发明解决技术问题所采用的进一步技术方案是: 所述的步骤 S5中, 中央控制单元进行资源预留控制, 但不对输入输出端口按每信元时隙进行 输入输出配对调度。  A further technical solution adopted by the present invention to solve the technical problem is as follows: In the step S5, the central control unit performs resource reservation control, but does not perform input/output pairing scheduling for each input/output port per cell slot.
本发明解决技术问题所采用的进一步技术方案是: 所述的输入线卡根 据新到达数据包的头信息决定其分类, 同一个流的不同级别的视频编码数 据, 被标记为不同等级的信元。  A further technical solution adopted by the present invention to solve the technical problem is: the input line card determines the classification according to the header information of the newly arrived data packet, and the video coding data of different levels of the same stream is marked as different levels of cells. .
本发明解决技术问题所采用的进一步技术方案是: 所述的分组交换结 构以部分排序自路由集线器的多级互连网络构成, 信元按其目标输出端口 地址及优先级进行排序, 并在每级按优先级阻塞超过其线组转发能力的信 元。 本发明解决技术问题所采用的进一步技术方案是: 所述的输出线卡完 成数据包的重新组装, 当信元交换到输出端口, 寻址位在寻址过程中被消 耗完, 信元控制信息主要剩余的活跃标记位及优先级标记位, 输出线卡中 的相应处理器刨去这些剩余的控制信息位再将信元重新组装成数据包。 本发明解决技术问题所采用的进一步技术方案是: 提供一种支持视频 质量分级的路由器系统, 其包括若干输入线卡, 若干输出线卡, 一个信元 交换结构, 及一个中央控制单元。 所述的输入线卡在输入端口根据数据包 的种类设定相应优先级, 对已经分级编码的视频码流数据包再把它们切分 成用于交换的信元, 并按其中视频分级编码码流的级别给予对应的信元优 先级, 分组交换结构根据信元的输出目标地址进行路由转发, 输出线卡对 信元进行重组装, 形成数据包后继续线速转发或由链路发送。 A further technical solution adopted by the present invention to solve the technical problem is: the packet switching structure is composed of a multi-level interconnection network partially sorted from a routing hub, and the cells are sorted according to their target output port addresses and priorities, and each The level blocks cells that exceed their line group forwarding capabilities by priority. A further technical solution adopted by the present invention to solve the technical problem is: the output line card completes reassembly of the data packet, and when the cell is exchanged to the output port, the addressing bit is cancelled during the addressing process. After consumption, the cell control information mainly has the remaining active flag bits and priority flag bits, and the corresponding processor in the output line card removes the remaining control information bits and reassembles the cells into data packets. A further technical solution adopted by the present invention to solve the technical problem is to provide a router system supporting video quality grading, comprising a plurality of input line cards, a plurality of output line cards, a cell switching structure, and a central control unit. The input line card sets a corresponding priority according to the type of the data packet on the input port, and then divides the video code stream data packets that have been hierarchically coded into cells for exchange, and according to the video hierarchical coded stream The level is given the corresponding cell priority. The packet switching structure performs routing and forwarding according to the output destination address of the cell, and the output line card reassembles the cell to form a data packet and then continue to be forwarded by the line rate or sent by the link.
本发明解决技术问题所采用的进一步技术方案是: 所述的输入线卡根 据新到达数据包的头信息决定其分类, 同一个流的不同级别的视频编码数 据, 被标记为不同等级的信元。  A further technical solution adopted by the present invention to solve the technical problem is: the input line card determines the classification according to the header information of the newly arrived data packet, and the video coding data of different levels of the same stream is marked as different levels of cells. .
本发明解决技术问题所采用的进一步技术方案是: 所述的分组交换结 构可以部分排序自路由集线器的多级互连网络构成, 信元按其目标输出端 口地址及优先级进行排序, 并在每级按优先级阻塞超过其线组转发能力的 信元。 本发明解决技术问题所采用的进一步技术方案是: 所述的输出线卡完 成 IP数据包的重新组装, 当信元交换到输出端口,寻址位在寻址过程中被 消耗完, 信元控制信息主要剩下活跃标记位及优先级标记位, 输出线卡中 的相应处理器刨去这些剩余的控制信息位再将信元重新组装成数据包。 本发明的支持视频质量分级的路由器系统以及构造方法可以实现超大 规模流媒体路由器结构, 可扩展的端口数N>10000, 每端口带宽 lOOMbits 每秒。 支持视频分级编码的流媒体, 在突发负载流量使得路由器交换结构 过载时立即丟弃媒体流中高级增强的编码部分, 而完整保留较低级别的基 本等级码流。 准电路交换结构, 无排队时延, 零抖动。 只对新到达的流进 行接入允许控制, 但不必每时隙的匹配调度。 信元分布式自路由地转发, 不存在资源瓶颈。  A further technical solution adopted by the present invention to solve the technical problem is: the packet switching structure may be partially ordered from a multi-level interconnection network of a routing hub, and the cells are sorted according to their target output port addresses and priorities, and each The level blocks cells that exceed their line group forwarding capabilities by priority. The further technical solution adopted by the present invention to solve the technical problem is: the output line card completes the reassembly of the IP data packet, and when the cell is exchanged to the output port, the addressing bit is consumed in the addressing process, and the cell control The information mainly has active flag bits and priority flag bits, and the corresponding processor in the output line card removes the remaining control information bits and reassembles the cells into data packets. The router system and the construction method supporting the video quality classification of the invention can realize a super-large-scale streaming media router structure, and the number of expandable ports is N>10000, and the bandwidth per port is lOOMbits per second. The streaming media that supports video hierarchical coding immediately discards the advanced enhanced coding portion of the media stream when the burst load traffic overloads the router switch fabric, while completely retaining the lower level basic level code stream. Quasi-circuit-switched architecture, no queuing delay, zero jitter. Access control is only allowed for newly arrived flows, but not with matching scheduling per slot. The cells are distributed and self-routing, and there is no resource bottleneck.
附图说明 DRAWINGS
图 1是本发明的支持视频质量分级的路由器构造方法的一个分组切割 成信元示意图。  BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 is a block diagram of a packet-cutting cell of a router construction method supporting video quality classification of the present invention.
图 2是本发明采用的可支持视频质量分级的路由器构造方法的 8x8网 络原理示意图。 图 3是本发明的支持视频质量分级的路由器构造方法的信元重新组装 的示意图。 具体实施方式 2 is an 8x8 network of a router construction method capable of supporting video quality classification according to the present invention. Schematic diagram of the network principle. 3 is a schematic diagram of cell reassembly of a router construction method supporting video quality grading according to the present invention. detailed description
以下内容是结合具体的优选实施方式对本发明所作的进一步详细说 明, 不能认定本发明的具体实施只局限于这些说明。 对于本发明所属技术 领域的普通技术人员来说, 在不脱离本发明构思的前提下, 还可以做出若 干筒单推演或替换, 都应当视为属于本发明的保护范围。  The following is a detailed description of the present invention in conjunction with the specific preferred embodiments. It is to be understood by those skilled in the art that the present invention can be delineated or replaced without departing from the spirit and scope of the invention.
随着上述应用和技术的变迁,网络的主要设备路由器必需从目前的"数 据" 路由器演变为 "流媒体" 路由器, 表 1是流媒体路由器与传统数据路 由器的比较。  With the changes in the above applications and technologies, the network's main device routers must evolve from current "data" routers to "streaming" routers. Table 1 compares streaming routers with traditional data routers.
Figure imgf000006_0001
Figure imgf000006_0001
表 1 "流媒体" 路由器与 "数据" 路由器的比较 本发明提供一种支持视频质量分级的路由器构造方法, 其包括以下步 骤:  Table 1 Comparison of "Streaming Media" Router and "Data" Router The present invention provides a router construction method supporting video quality grading, which includes the following steps:
S1 : 在输入线卡处根据数据包的种类设定优先级, 对已经分级编码的 视频码流的数据包, 在把它们切分成用于分组交换的信元时, 按其中视频 分级编码码流的级别给予对应的信元优先级编码;  S1: setting a priority according to the type of the data packet at the input line card, and dividing the data packets of the video code stream that has been hierarchically coded into the cells for packet switching, according to the video hierarchical coded stream The level of the corresponding cell priority encoding;
S2: 分组交换结构根据信元的输出目标地址进行路由转发;  S2: The packet switching structure performs routing and forwarding according to the output destination address of the cell;
S3 : 输出线卡对信元进行重新组装, 形成数据包后继续线速转发或由 链路发送; S3: The output line card reassembles the cells, and after forming the data packet, continues the line speed forwarding or Link transmission
S4: 构成 Ν χ Ν的路由结构, 其输入输出分为 M个线组每个, 每个线 组由 G个输入或输出端口构成, 即N = MxG;  S4: A routing structure constituting Ν Ν ,, whose input and output are divided into M line groups, each line group consisting of G input or output ports, that is, N = MxG;
S5: 中央控制单元对长时间持续的流进行接入允许控制。  S5: The central control unit performs access control for long-lasting streams.
所述的步骤 S2中, 当出现信元争用内部端口或出线资源时, 低优先级 的信元将被阻塞掉丟弃而不予緩存。  In the step S2, when the cell contends for the internal port or the outgoing resource, the low priority cell will be blocked and discarded without being cached.
所述的步骤 S5中, 中央控制单元进行资源预留控制, 但不对输入输出 端口按每信元时隙进行输入输出配对调度。  In the step S5, the central control unit performs resource reservation control, but does not perform input/output pairing scheduling for each input/output port per cell slot.
所述的输入线卡根据新到达数据包的头信息决定其分类,同一个流的不 同级别的视频编码数据, 被标记为不同等级的信元。  The input line card determines its classification according to the header information of the newly arrived data packet, and the video coding data of different levels of the same stream is marked as cells of different levels.
所述的分组交换结构以部分排序自路由集线器的多级互连网络构成,信 元按其目标输出端口地址及优先级进行排序, 并在每级按优先级阻塞超过 其线组转发能力的信元。 路由器输入线卡根据新到达数据包的头相关信息 决定其分类。  The packet switching structure is composed of a multi-level interconnection network partially ordered from a routing hub, the cells are sorted according to their target output port addresses and priorities, and the packets exceeding the forwarding capability of the line group are blocked by priority at each level. yuan. The router input line card determines its classification based on the header related information of the newly arrived packet.
所述的输出线卡完成 IP数据包的重新组装, 当信元交换到输出端口, 寻址位在寻址过程中被消耗完, 信元控制信息主要剩下活跃标记位及优先 级标记位, 输出线卡中的相应处理器刨去这些剩余的控制信息位再将信元 重新组装成数据包。  The output line card completes the reassembly of the IP data packet. When the cell is exchanged to the output port, the addressing bit is consumed in the addressing process, and the cell control information mainly has the active flag bit and the priority flag bit. The corresponding processor in the output line card removes these remaining control information bits and reassembles the cells into packets.
本发明提供一种支持视频质量分级的路由器系统, 其包括若干输入线 卡, 若干输出线卡, 一个信元交换结构, 及一个中央控制单元。 所述的输 入线卡在输入端口根据数据包的种类设定相应优先级, 对已经分级编码的 视频码流数据包, 在把它们切分成用于交换的信元时按其中视频分级编码 码流的级别给予对应的信元优先级, 分组交换结构根据信元的输出目标地 址进行路由转发。 输出线卡对信元进行重新组装, 形成数据包后继续线速 转发或由链路发送。  The present invention provides a router system that supports video quality grading, comprising a number of input line cards, a number of output line cards, a cell switching structure, and a central control unit. The input line card sets a corresponding priority according to the type of the data packet on the input port, and the video stream data packet that has been hierarchically encoded is divided into video stream coded streams when they are divided into the cells for exchange. The level is given to the corresponding cell priority, and the packet switching structure performs routing and forwarding according to the output destination address of the cell. The output line card reassembles the cells to form a data packet and continues the line rate forwarding or transmission by the link.
1.输入线卡 1. Input line card
输入线卡的主要功能是在输入端口根据数据包的种类设定相应优先 级, 特别地, 对已经分级编码的视频码流数据包, 再把它们切分成用于交 换的信元, 并按其中视频分级编码码流的级别给予对应的信元优先级。  The main function of the input line card is to set the corresponding priority according to the type of the data packet at the input port. In particular, the video stream packets that have been hierarchically encoded are further divided into cells for exchange, and according to which The level of the video grading code stream is given the corresponding cell priority.
如图 1所示, 在输入线卡可将一个分组切分成若干个小的信元, 在每 个输入信元前加入 m+n+1位的控制信息 Ipm-l ...p0dld2...dn,其中首位为 活跃位, pm-ΐ ...ρθ为优先级位, dld2...dn为寻址位。 以 m=3为例, 将分组优先级设为 8级, 分组优先级列表如下: As shown in FIG. 1, the input line card can divide a packet into a plurality of small cells, and add m+n+1 bits of control information Ipm-l ... p0dld2 before each input cell. Dn, where the first bit is the active bit, pm-ΐ ... ρθ is the priority bit, and dld2...dn is the addressing bit. Taking m=3 as an example, the packet priority is set to level 8. The group priority list is as follows:
Figure imgf000008_0001
在网络中高优先级的数据分组所占比例极低, 占用带宽可忽略不计。 而流媒体占据着网络传输中最大的流量。 在流媒体中又以视频的基本层数 据和音频数据为最重要的信息, 流媒体信息最大的特点是它的强实时性, 因此在设计流式路由器时, 在带宽有限的情况下将优先保证其高优先级的 分组发送。 视频基本层和音频设为较高的优先级, 理所当然是优先保证视 频分组的基本层和音频信息的传输。 用户通过接收到的视频基本层可以解 码出具有一般图像质量的视频图像。 而当带宽足够的话, 用户完全可以在 图像解码延迟范围内接收到视频图像的增强层信息 (El、 E2、 E3), 从而可 重构出清晰的高质量视频图像。
Figure imgf000008_0001
The proportion of high-priority data packets in the network is extremely low, and the occupied bandwidth is negligible. Streaming media occupies the largest traffic in network transmission. In the streaming media, the basic layer data and audio data of the video are the most important information. The biggest feature of the streaming media information is its strong real-time performance. Therefore, when designing the streaming router, priority will be given when the bandwidth is limited. Its high priority packet is sent. The video base layer and audio are set to a higher priority, which of course is to preferentially guarantee the transmission of the base layer and audio information of the video packet. The user can decode a video image having a general image quality through the received video base layer. When the bandwidth is sufficient, the user can receive the enhancement layer information (El, E2, E3) of the video image within the image decoding delay range, so that a clear high quality video image can be reconstructed.
2. 分组交换结构  2. Packet switching structure
若一个互连网络中的每个路由节点都使用带内控制来设置其连接状 态, 则该网络实现了自路由交换。 对于一个由 2x2路由节点组成的互连网 络, 可以根据如下 2比特序列, 使用 2x2排序器对输入信号排序:  If each routing node in an interconnected network uses in-band control to set its connection state, the network implements self-routing exchange. For an interconnect consisting of 2x2 routing nodes, the input signal can be sorted using a 2x2 sequencer according to the following 2-bit sequence:
10 < 00 < 11。  10 < 00 < 11.
在一个 2nx2n互连网络中, 每个输入信号具有路由标记, 它由一个活 跃比特, 以及其后的 n个地址比特组成。 每个地址比特均表示在某个特定 级的路由节点处的输出选择, 并且当通过某级的该路由节点时被消耗。 使 用这样的路由标记方式, 2x2排序器将其输入信号按照如下的等价方式排 序: "向 0目标地址" < "空闲" < "向 1 目标地址"。 In a 2nx2n interconnect network, each input signal has a routing flag consisting of one active bit and the next n address bits. Each address bit represents an output selection at a particular level of routing node and is consumed when passing through a routing node of a certain level. Using such a route tagging method, the 2x2 sequencer sorts its input signals in an equivalent manner as follows: "To 0 destination address"<"idle"<"to 1 destination address".
在线组大小为 R的多级互连网络中, 若每个节点均为 2Rx2R集线器, 并将其输入信号按照以下 2比特序列排序: 10 < 00 < 11 , 则自路由机制可 用于控制信号的路由。  In a multi-level interconnection network with an online group size of R, if each node is a 2Rx2R hub, and its input signals are sorted according to the following 2-bit sequence: 10 < 00 < 11 , the self-routing mechanism can be used to control signal routing. .
这样就达到了将最大数目的向 0目标地址(相应的, 向 1 目标地址) 信号路由到它的 R个 0-输出 (相应的, 1-输出) 的目的。  This achieves the goal of routing the maximum number of 0 target addresses (corresponding to 1 target address) to its R 0-outputs (corresponding, 1-output).
一个路由单元从每个输入信元中取前两个比特, 作为带内控制信号。 假定在一个活跃输入端口生成一个活跃比特, 其后是地址比特及非控制比 特。 地址比特指示了在单元的两个输出之间所做的选择。 与此同时, 空闲 分组为 0比特组成的串。 因此目标地址为向 0目标地址、 向 1 目标地址的 分组、 以及空闲分组的带内控制信号分别为 10, 11和 00。 这种情况下, 路由单元的基本线性排序就转化为:  A routing unit takes the first two bits from each input cell as an in-band control signal. It is assumed that an active bit is generated on an active input port, followed by an address bit and a non-control bit. The address bits indicate the choice made between the two outputs of the unit. At the same time, the idle packet is a string of 0 bits. Therefore, the in-band control signals for the destination address of the 0 destination address, the packet to the 1 destination address, and the idle packet are 10, 11 and 00, respectively. In this case, the basic linear ordering of the routing unit translates to:
"向 0目标地址" < "空闲" < "向 1 目标地址"。  "To 0 destination address" < "idle" < "to 1 destination address".
下表给出了一个路由单元的连接状态。  The following table shows the connection status of a routing unit.
Figure imgf000009_0001
Figure imgf000009_0001
当一个网络包含大量交换节点时, 整个网络中交换控制的速度取决于 其分布程度。 在分布式控制中, 每个分组在进入交换节点时, 都应携带一 个带内控制信号作为其前缀, 使得该节点的交换决策只取决于该输入分组 所携带的带内控制信号。 在分组交换网络中, 一个带内控制机制是自路由 的, 进入一个交换节点的分组的控制信号只由该分组的 I/O地址确定。  When a network contains a large number of switching nodes, the speed of switching control throughout the network depends on its degree of distribution. In distributed control, each packet should carry an in-band control signal as its prefix when entering the switching node, so that the switching decision of the node depends only on the in-band control signal carried by the input packet. In a packet switched network, an in-band control mechanism is self-routing, and control signals for packets entering a switching node are determined only by the I/O address of the packet.
在宽带应用中, 进入一个交换节点的分组的控制信号需要包含尽可能 少的比特, 使得交换决策过程可以很快的执行。  In broadband applications, the control signals of packets entering a switching node need to contain as few bits as possible, so that the exchange decision process can be performed very quickly.
当一个分组进入网络时, 其用于带内控制的前缀为其二进制目标地址 dld2...dn。 比特 dj指示第 j级的单元在两个输出之间应做的选择, 并被第 j 级交换控制所消耗。 交换单元的平行 /交叉状态仅由两个输入分组的首比 特确定。 当两个分组争用交换单元的同一个输出端口时, 其中一个分组将 被误路由或阻塞。 通常一个应用允许空的输入, 因此实际上带内控制信号 在最前部包含一个 "活跃比特",用于区别于无载荷数据的空闲分组 00…… 0, 因此带内控制信号通常为
Figure imgf000010_0001
When a packet enters the network, its prefix for in-band control is its binary destination address dld2...dn. Bit dj indicates the choice of the jth level unit between the two outputs, and is The j-level exchange control is consumed. The parallel/cross state of the switching unit is determined only by the first bit of the two input packets. When two packets contend for the same output port of the switching unit, one of the packets will be misrouted or blocked. Usually an application allows for an empty input, so the in-band control signal actually contains an "active bit" at the forefront for distinguishing between idle packets 00... 0 of unloaded data, so the in-band control signal is usually
Figure imgf000010_0001
这种机制的基础是网络的自路由目标地址导向序列:例如 Omega网络 为单调序列 1 , 2, ……, n。对于一个导向序列为 γ(1)=1, γ(2)=2, … , γ(η)=η 的一般 2ηχ2η榕树类型网络, 该机制可推广如下, 一个目标地址为二进制 (dld2… dn)的分组使用二进制控制流 lp^ ...pod^d^) ... d^)作为前缀。因此 进入交换网络的整个分组的格式为: 带内控制信号数据  The basis of this mechanism is the self-routing destination address-oriented sequence of the network: for example, the Omega network is a monotonic sequence of 1, 2, ..., n. For a general 2ηχ2η榕tree type network with γ(1)=1, γ(2)=2, ..., γ(η)=η, the mechanism can be generalized as follows: A target address is binary (dld2... dn) The grouping uses the binary control stream lp^ ...pod^d^) ... d^) as a prefix. Therefore, the format of the entire packet entering the switching network is: In-band control signal data
lp2plPod(1)d(2) 例如一个 ATM单元或一个 IP分组 Lp 2 p lPo d (1) d (2) such as an ATM unit or an IP packet
在比特流到达第 j级之前, 数据段 (1γ(1)(1γ(2)...(1γ(ΐ- 1)已经被消耗, 因 此两个信元的首控制比特为 lpsPiPodyG), 第 j级交换控制仅读取分组中的 这两些比特。 Before the bit stream reaches the jth stage, the data segment (1γ(1)(1γ(2)...(1γ(ΐ-1) has been consumed, so the first control bit of the two cells is lpsPiPodyG), j The level exchange control reads only these two bits in the packet.
交换控制可以通过一个具有三个状态: 初始, 平行和交叉状态的有限 状态自动机进行描述。其初始状态是任意的,有一个暂定的连接状态平行, 或交叉。 自动机的激励信号包含来自两个同步输入数据中每个数据的首五 个比特 (= Ορ2ριροΟ, 1ρ2ριροΟ or 1ρ2ριρ01 两个数据输入共组成 9*23=81种 不同的激励。 根据该激励, 交换单元在活跃输出处复制活跃比特和优先级 位,并根据自动机的状态平行或交叉,分别将连接状态锁存为平行或交叉。 在此之后, 相继的比特只是筒单的在锁存的连接状态下通过交换单元。 当 两个输入信号的首尾两个比特均为 "10"或 "11"时, 就产生了输出争用。 输 出争用可以采用不同的仲裁方式解决, 或是偏转(即误路由)、或阻塞其中 的一个分组。 自动机的一个附加激励来自非数据输入的帧时钟。 它将单元 重置为初始状态, 下表显示了自路由交换单元中的自动机行为。 激励 ^犬态 =初始 状态 =平行 状态 =交叉 Switching control can be described by a finite state automaton with three states: initial, parallel, and crossed states. Its initial state is arbitrary, with a tentative connection state parallel, or cross. The automaton's excitation signal contains the first five bits from each of the two synchronized input data (= Ορ 2 ριροΟ, 1ρ 2 ριροΟ or 1ρ 2 ριρ 0 1 Two data inputs are combined to form 9*2 3 = 81 different According to the excitation, the switching unit copies the active bit and the priority bit at the active output, and respectively latches the connection state into parallel or cross according to the state of the automaton parallel or cross. After that, the successive bits are only The single unit passes through the switching unit in the latched connection state. When both the first and last two bits of the input signal are "10" or "11", the output contention is generated. The output contention can be differently arbitrated. The solution is to either deflect (ie misroute) or block one of the packets. An additional stimulus for the automaton comes from the frame clock that is not data input. It resets the unit to its initial state, and the following table shows the self-routing unit. Automaton behavior in . Excitation ^ dog state = initial state = parallel state = cross
Op2piPoO和 0 ρ'2ρΊρΌ0 输出 0 和 0; Op 2 piPoO and 0 ρ' 2 ρΊρΌ0 outputs 0 and 0;
0 Ρ2Ρ1Ρ0Ο和 1 ρ'2ρΊρΌ0 输出 1和 0; V 0 Ρ2Ρ1Ρ0Ο and 1 ρ' 2 ρΊρΌ0 outputs 1 and 0; V
0 P2P1P0O和 1 ρ'2ρ'ιρ'01 输出 0和 1; V 0 P2P1P0O and 1 ρ' 2 ρ'ιρ' 0 1 Output 0 and 1; V
1 P2P1P0O和 0 ρ'2ρΊρΌ0 输出 1和 0; V 1 P2P1P0O and 0 ρ' 2 ρΊρΌ0 outputs 1 and 0; V
1 P2P1P0O和 1 ρ'2ρΊρΌ0 输出争用仲裁 Ρ2ΡιΡο< ρ'2ρΊρΌ Ρ2ΡιΡο > ρ'2ρΊρΌ 1 P2P1P0O and 1 ρ' 2 ρΊρΌ0 Output contention arbitration Ρ2ΡιΡο<ρ'2ρΊρΌ Ρ2ΡιΡο >ρ'2ρΊρΌ
1 P2P1P0O和 1 ρ'2ρ'ιρ'01 输出 0 和 1; V 1 P2P1P0O and 1 ρ' 2 ρ'ιρ' 0 1 outputs 0 and 1; V
1 P2P1P0I和 0 ρ'2ρΊρΌ0 输出 0和 1; V 1 P2P1P0I and 0 ρ' 2 ρΊρΌ0 outputs 0 and 1; V
1 P2P1P0I和 1 Ρ'2Ρ'ιΡ'οΟ 输出 1和 1; V 1 P2P1P0I and 1 Ρ'2Ρ'ιΡ'οΟ Outputs 1 and 1; V
1 P2P1P0I和 1 Ρ'2Ρ'ιΡ'ο1 输出争用仲裁 Ρ2ΡιΡο> ρ'2ρΊρΌ P2P1P0 <Ρ'2Ρ'ιΡ'ο 帧时钟 状态 <~初始; 此处输出争用仲裁采用输出高优先级信元,丟弃低优先级信元的策略。 在单元控制的实际 VLSI实现中, 为了筒化逻辑, 对于来自一个输入 的两个首比特, 倾向于每次只处理其中之一。 1 P2P1P0I and 1 Ρ'2Ρ'ιΡ'ο1 Output contention arbitration Ρ2ΡιΡο> ρ'2ρΊρΌ P2P1P0 <Ρ'2Ρ'ιΡ'ο Frame clock state <~ initial; here the output contention arbitration uses the output high priority cell, A policy for dropping low priority cells. In the actual VLSI implementation of cell control, in order to bin the logic, for the first bit from one input, there is a tendency to process only one of them at a time.
3.输出线卡  3. Output line card
输出线卡的主要功能是完成数据包的重组。 当信元交换到输出端口,寻 址位在寻址过程中被消耗完, 信元控制信息主要剩下活跃标记位及优先级 标记位, 输出线卡中的相应处理器刨去这些剩余的控制信息位再将信元重 新组装成数据包。 本发明的支持视频质量分级的路由器系统以及构造方法可以实现超大 规模流媒体路由器结构, 可扩展的端口数 N>10000, 每端口带宽 100Mbit 每秒。 支持视频分级编码的流媒体, 在突发负载流量使得路由器交换结构 过载时立即丟弃媒体流中高级增强的编码部分, 而完整保留较 级别的增 强码流。 准电路交换结构, 无排队时延, 零抖动。 只对新到达的流进行接 入允许控制, 但不必每时隙的匹配调度。 信元分布式自路由地转发, 不存 在资源瓶颈。 The main function of the output line card is to complete the reassembly of the data packet. When the cell is switched to the output port, the addressing bit is consumed in the addressing process. The cell control information mainly has the active flag bit and the priority flag bit, and the corresponding processor in the output line card removes the remaining control. The information bits then reassemble the cells into packets. The router system and the construction method supporting the video quality classification of the invention can realize a super-large-scale streaming media router structure, and the number of expandable ports is N>10000, and the bandwidth per port is 100 Mbits per second. The streaming media supporting video hierarchical coding immediately discards the advanced enhanced coding part of the media stream when the burst load traffic overloads the router switching structure, and completely retains the level enhanced code stream. Quasi-circuit-switched architecture, no queuing delay, zero jitter. Access control is only allowed for newly arrived flows, but not with matching scheduling per slot. The cells are distributed and self-routing, and there is no resource bottleneck.

Claims

权利要求书 Claim
1.一种支持视频质量分级的路由器构造方法, 其包括以下步骤: A router construction method supporting video quality grading, comprising the steps of:
S1 : 在输入线卡处根据数据包的种类设定优先级, 对已经分级编码的 视频码流数据包, 在把它们切分成用于分组交换的信元时, 按其中视频分 级编码码流的级别给予对应的信元优先级编码; S1: setting a priority according to the type of the data packet at the input line card, and dividing the video code stream data packets that have been hierarchically coded into the cells for packet switching, according to the video hierarchical coded code stream The level is given a corresponding cell priority encoding;
S2: 分组交换结构根据信元的输出目标地址进行路由转发;  S2: The packet switching structure performs routing and forwarding according to the output destination address of the cell;
S3: 输出线卡对信元进行重组装, 形成数据包后继续线速转发或由链 路发送;  S3: The output line card reassembles the cells, and after forming the data packet, continues the line speed forwarding or is sent by the link;
S4: 构成 N X N的路由结构, 其输入输出分为 M个线组, 每个线组由 G个输入或输出端口构成, 即N = MxG;  S4: A routing structure constituting N X N, whose input and output are divided into M line groups, each line group consisting of G input or output ports, that is, N = MxG;
S5: 中央控制单元对长时间持续的流进行接入允许控制。  S5: The central control unit performs access control for long-lasting streams.
2. 如权利要求 1 所述的支持视频质量分级的路由器构造方法, 其特征 在于, 所述的步骤 S2 中, 当出现信元争用内部端口或出线资源时, 低优 先级的信元将被阻塞掉丟弃而不予緩存。  2. The method for constructing a router for supporting video quality grading according to claim 1, wherein in the step S2, when a cell contends for an internal port or an outgoing resource, a low priority cell is to be used. Block the drop and not cache.
3. 如权利要求 1 所述的支持视频质量分级的路由器构造方法, 其特征 在于, 所述的步骤 S5 中, 中央控制单元进行资源预留控制, 但不对输入 输出端口按每信元时隙进行输入输出配对调度。  3. The method for constructing a router for supporting video quality grading according to claim 1, wherein in the step S5, the central control unit performs resource reservation control, but does not perform input/output ports per cell slot. Input and output pairing scheduling.
4. 如权利要求 1 所述的支持视频质量分级的路由器构造方法, 其特征 在于, 所述的输入线卡根据新到达数据包的头信息决定其分类, 同一个流 的不同分辨率级别的视频编码数据, 被标记为不同等级的信元。  4. The method for constructing a router for supporting video quality grading according to claim 1, wherein the input line card determines its classification according to the header information of the newly arrived data packet, and the video of different resolution levels of the same stream. Encoded data, marked as cells of different levels.
5. 如权利要求 1 所述的支持视频质量分级的路由器构造方法, 其特征 在于, 所述的分组交换结构可以部分排序自路由集线器的多级互连网络构 成, 信元按其目标输出端口地址及优先级进行排序, 并在每级按优先级阻 塞超过其线组转发能力的信元。  5. The method according to claim 1, wherein the packet switching structure is partially ordered from a multi-level interconnection network of a routing hub, and the cell outputs a port address according to its destination. And prioritize the cells, and block the cells that exceed their line group forwarding capabilities by priority at each level.
6. 如权利要求 1 所述的支持视频质量分级的路由器构造方法, 其特征 在于, 所述的输出线卡完成 IP数据包的重构, 当信元交换到输出端口, 寻 址位在寻址过程中被消耗完, 信元控制信息主要剩下活跃标记位及优先级 标记位, 输出线卡中的相应处理器单元刨去这些剩余的控制信息位再将信 元重新组装成数据包。  6. The method of constructing a router for supporting video quality grading according to claim 1, wherein said output line card performs reconstruction of an IP data packet, and when the cell is exchanged to an output port, the addressing bit is addressed. After the process is exhausted, the cell control information mainly has active flag bits and priority flag bits, and the corresponding processor unit in the output line card removes the remaining control information bits and reassembles the cells into data packets.
7.—种支持视频质量分级的路由器系统, 其特征在于: 其包括若干输入 线卡, 若干输出线卡, 一个信元交换结构, 及一个中央控制单元, 所述的 输入线卡在输入端口根据数据包的种类设定相应优先级, 对已经按视频质 量分级编码的视频码流数据包, 在把它们切分成用于交换的信元时, 按其 视频分级编码码流的级别给予对应的信元优先级, 分组交换结构根据信元 的输出目标地址进行路由转发, 输出线卡对信元进行重组装, 形成数据包 后继续线速转发或由链路发送。 7. A router system supporting video quality grading, characterized in that it comprises: a plurality of input line cards, a plurality of output line cards, a cell switching structure, and a central control unit, said input line card being at the input port according to The type of the packet is set to the corresponding priority, and the video quality has been The size-coded video stream packets are given a corresponding cell priority according to the level of the video-graded code stream when they are sliced into cells for exchange, and the packet switching structure is based on the output destination address of the cell. Route forwarding is performed, and the output line card reassembles the cells to form a data packet and then continues to be forwarded by the line rate or sent by the link.
8. 如权利要求 7 所述的支持视频质量分级的路由器系统,其特征在于, 所述的输入线卡根据新到达数据包的头信息决定其分类, 同一个流的不同 级别的视频编码数据, 被标记为不同等级的信元。  8. The router system for supporting video quality grading according to claim 7, wherein the input line card determines its classification according to the header information of the newly arrived data packet, and the video coding data of different levels of the same stream, Cells marked as different levels.
9. 如权利要求 7 所述的支持视频质量分级的路由器系统,其特征在于, 所述的分组交换结构以部分排序自路由集线器的多级互连网络构成, 信元 按其目标输出端口地址及优先级进行排序, 并在每级按优先级阻塞超过其 线组转发能力的信元。  9. The router system for supporting video quality grading according to claim 7, wherein said packet switching structure is formed by a multi-level interconnection network partially ordered from a routing hub, and the cell outputs a port address according to a target thereof. The priorities are sorted, and cells that exceed their line group forwarding capabilities are blocked by priority at each level.
10. 如权利要求 7 所述的支持视频质量分级的路由器系统, 其特征在 于, 所述的输出线卡完成数据包的重新组装, 当信元交换到输出端口, 寻 址位在寻址过程中被消耗完, 信元控制信息主要剩下活跃标记位及优先级 标记位, 输出线卡中的相应处理器刨去这些剩余的控制信息位再将信元重 新组装成数据包。  10. The router system for supporting video quality grading according to claim 7, wherein said output line card completes reassembly of the data packet, and when the cell is exchanged to the output port, the addressing bit is in the addressing process. After being consumed, the cell control information mainly has active flag bits and priority flag bits, and the corresponding processor in the output line card removes the remaining control information bits and reassembles the cells into data packets.
PCT/CN2009/074738 2009-10-31 2009-10-31 Router constructing method supporting hierarchy of video quality WO2011050542A1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2009/074738 WO2011050542A1 (en) 2009-10-31 2009-10-31 Router constructing method supporting hierarchy of video quality

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2009/074738 WO2011050542A1 (en) 2009-10-31 2009-10-31 Router constructing method supporting hierarchy of video quality

Publications (1)

Publication Number Publication Date
WO2011050542A1 true WO2011050542A1 (en) 2011-05-05

Family

ID=43921269

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2009/074738 WO2011050542A1 (en) 2009-10-31 2009-10-31 Router constructing method supporting hierarchy of video quality

Country Status (1)

Country Link
WO (1) WO2011050542A1 (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1728677A (en) * 2004-07-28 2006-02-01 国家数字交换系统工程技术研究中心 Method and system for carrying video stream media through network in multiple velocities
CN1866927A (en) * 2006-05-08 2006-11-22 国家数字交换系统工程技术研究中心 Information switching realizing system and method and scheduling algorithm
WO2008066257A1 (en) * 2006-11-30 2008-06-05 Electronics And Telecommunications Research Institute Apparatus and method for hierarchical modulation transmission and reception of scalable video bitstream
CN101291546A (en) * 2008-06-11 2008-10-22 清华大学 Switching structure coprocessor of core router

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1728677A (en) * 2004-07-28 2006-02-01 国家数字交换系统工程技术研究中心 Method and system for carrying video stream media through network in multiple velocities
CN1866927A (en) * 2006-05-08 2006-11-22 国家数字交换系统工程技术研究中心 Information switching realizing system and method and scheduling algorithm
WO2008066257A1 (en) * 2006-11-30 2008-06-05 Electronics And Telecommunications Research Institute Apparatus and method for hierarchical modulation transmission and reception of scalable video bitstream
CN101291546A (en) * 2008-06-11 2008-10-22 清华大学 Switching structure coprocessor of core router

Similar Documents

Publication Publication Date Title
AU681062B2 (en) Network having secure fast packet switching and guaranteed quality of service
EP2060067B1 (en) Ethernet switching
US7688725B2 (en) Content-aware congestion control system
US6115373A (en) Information network architecture
CN109005471B (en) Scalable video stream multicast method based on QoS intelligent perception under SDN environment
JP2006506845A (en) How to select a logical link for a packet in a router
US7623453B2 (en) Aggregation switch apparatus for broadband subscribers
JP6055553B2 (en) Queue scheduling method, apparatus and system
US8031682B2 (en) Apparatus and method for aggregating and switching traffic in subscriber network
US8213421B2 (en) Methods and systems for efficient multicast across a mesh backplane
Szymanski et al. Internet multicasting of IPTV with essentially-zero delay jitter
EP1187402A2 (en) Switching apparatus, communication apparatus, and communication system
Kamiyama et al. Cache replacement based on distance to origin servers
Wu Packet forwarding technologies
WO2019200568A1 (en) Data communication method and device
WO2011050542A1 (en) Router constructing method supporting hierarchy of video quality
WO2014085972A1 (en) Basic self-routing unit and construction method of semi-cleaner, sorting unit, network hub and multicast exchange network thereof
US7009973B2 (en) Switch using a segmented ring
Szymanski et al. Design of an IPTV multicast system for internet backbone networks
Chen et al. Multicast Switching Fabric Based on Network Coding and Algebraic Switching Theory
Ushijima et al. Scalable Internet Backbone Using Multi-Gigabit ATM-Based Connectionless Switching
Li System architecture and hardware implementations for a reconfigurable MPLS router
Durković et al. Packet Switch Architecture For Efficient Unicast and Multicast Traffic Switching
Nananukul et al. Controlling short-term packet loss ratios using an adaptive pushout scheme
Smit et al. Multicast and Broadcast in the Rattlesnake ATM Switch.

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 09850748

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 09850748

Country of ref document: EP

Kind code of ref document: A1