CN100499529C - Jitter detection based congestion control method in city domain Ethernet - Google Patents

Abstract

The invention relates to a transmission control method of the multimedia data, in particular to a congestion control method based on the jitter measurement in the metro Ethernet network. The method utilizes the priority bit position of the C-VLAN Tag of the Q-in-Q the metro Ethernet network as the counter of the node time delay jitter. The accumulated time delay jitter of the multimedia data frame in the transmission process from end to end is followed. The data of the small jitter corresponds to the higher priority and the data of the higher jitter corresponds to the lower priority. The Ethernet data frame, the jitter value of which is larger than the user tolerate threshold value, is lost by the determination of the user tolerate threshold value. The invention can effectively reduce the average time delay jitter of the multimedia flow, which can improve the good throughput of the multimedia service and the service quality of the multimedia service.

Description

The jamming control method that detects based on shake in the Metro Ethernet

Technical field

The present invention relates to a kind of transfer control method of multi-medium data, be specially the jamming control method that detects based on shake in a kind of Metro Ethernet.

Background technology

Along with IPTV, video monitoring, the extensive use of Triple play etc., the multimedia video business becomes one of fastest-rising application on the current metropolitan area network.In current metropolitan area network field, Ethernet has become one of solution very widely.In fact, Ethernet service is implemented in current a lot of operations commercial city on metropolitan area network.How in Metro Ethernet better bearing multimedia to use be one of current Metro Ethernet urgent problem.

When transmitting on the video flowing of multimedia application or the network node of audio stream, keeping a constant delay relation between the general tentation data grouping, but in fact this is very difficult in packet switching.In the transmission course of reality, because the variation that the congested generation or the difference of transmission delay have caused packet to produce network delay when arriving the destination, just produce delay variation, caused the service quality of these multimedia data flow to reduce.When the multimedia data stream passing through network, if the delay variation that produces is excessive, even the user receives these data flow, it also will become unavailable concerning the user.This is because be difficult to adjustment between the sequential is carried out in the packet between a data flow or a plurality of data flow again at user side.[1] [2] in some multimedia systems have distributed a buffering area to solve this problem in client.Yet such buffering area only is to have increased the tolerance of network for delay variation, not from addressing this problem in fact.

If the delay variation that has a large amount of multi-medium datas to be grouped in accumulative total in the network transmission process has surpassed client's degrees of tolerance, it is bad more that so this situation will become.In this case, though these multi-medium datas are unavailable concerning the user, they are still taking the bandwidth of network, and transmit the Congestion Level SPCC that these useless data have increased network undoubtedly.This new congestion problems can not be by traditional queue management method, such as random early detection (red) [3] or lose tail algorithm (DropTail) [3] and solve.

The main representative of tradition jamming control method is random early detection (red) and loses tail algorithm (DropTail) [3].RED is a kind of active queue management mechanism that IETF proposes.Its operation principle is according to current queue length and queue size threshold value, and the grouping of coming in is carried out abandoning at random with certain probability.The validity of RED has been passed through the checking of some practices, but still has some defectives.Such as, the performance sensitive of RED algorithm can cause formation vibration, throughput to reduce and delay variation aggravation [4] under the particular network load state in design parameter and network condition.Generally speaking, RED does not provide any protection mechanism to TCP stream.When UDP stream and its shared queue, the flow of general UDP can be greater than the TCP flow.The friendly that how to reach TCP in the RED algorithm is a much-talked-about topic in the RED research field.In document [5], provided the mutation of a RED algorithm, it has considered that when abandoning the size of grouping is to reach better effect and fairness.In document [6], replace average queue length to calculate losing probability in RED with average retardation.In document [7], propose a kind of preferential drop policy and discerned data flow single, that take too much bandwidth, the speed by this these streams of strategy just is compressed in the scope of value of regulation.In order to improve and improve other mutation algorithms that deficiency that the RED algorithm exists and defective have also produced a large amount of RED, more influential WRED[8 arranged], Stabilized-RED[9], Self-configurationRED[10], Adaptive RED[11], FRED[12] and Balanced-RED[13], wherein FRED and Balanced-RED stress to solve the fairness problem that RED exists, and all the other all are in order to strengthen the stability of RED.These traditional jamming control methods all do not have to consider fully the feature of network traffics, have especially ignored some characteristics that multimedia service is different from other data services, such as the sensitiveness of multimedia service for delay variation.Though those are sent to client, accumulated the data of the delay variation that surpasses user's tolerance, in fact nonsensical for the user.Traditional jamming control method is helpless for this situation.

In current metropolitan area network field, Ethernet has been one of the most popular technical solution.In fact, Ethernet service is implemented in current a lot of operations commercial city on metropolitan area network.Yet ethernet technology is extended to metropolitan area network must face and solve many stubborn problems, for example breaks through a series of problems such as the several restriction of 4096 VLAN (VLAN), the professional connection of transparent LAN, service quality assurance.For restriction and the professional transmission of realization transparent LAN that in Metro Ethernet, breaks through 4096 VLAN, the standard IEEE 802.1ad that IEEE defined Q-in-Q in 2005 (Provider Bridge) [14].Q-in-Q is also referred to as Stack VLAN, improves the autgmentability of network by the mode that increases one deck Qtag.In the Q-in-Q technology, outer field VLAN label is called S-VLAN, and this VLAN belongs to VLAN space in the carrier network scope, and the VLAN label of internal layer is called C-VLAN, is the VLAN space that belongs to client.When operator after the ethernet data frame that the user sends is received at the UNI place, just a S-VLAN Tag label is added to the Ether frame of user IEEE 802.1Q Tag label.The Q-in-Q Ethernet data frame format of IEEE802.1ad as shown in Figure 1.

S-VLAN Tag label is to be embedded in after Ethernet target MAC (Media Access Control) address and the source MAC, before C-VLAN.In 802.1Q VLAN, the priority among the VLAN has taken three, supports eight different priority, is used for distinguishing different services.When carrying out the forwarding of Frame in Metro Ethernet, only consider the priority setting among the S-VLAN usually, the setting of S-VLAN ID and S-VLAN CoS is usually used in identification service and service performance (CoS).And C-VLAN Tag label is not changed usually, and is also not accessed.

Summary of the invention

The objective of the invention is at the deficiencies in the prior art part, the jamming control method that detects based on shake in a kind of Metro Ethernet is provided, this method can reduce the average delay shake of media stream effectively, improve the effective throughput of multimedia service, thereby improve the service quality of multimedia service.

Technical scheme of the present invention is as follows: the jamming control method that detects based on shake in a kind of Metro Ethernet, this method utilizes priority bit position among the C-VLAN Tag of the Q-in-Q in the Metro Ethernet as the counter of node delay variation, follow the tracks of the delay variation that the multi-medium data frame is accumulated in transmission course end to end, the corresponding higher priority of data that the jitter value total_jitter that ethernet data frame experiences from the source to the present node is less, the lower priority of data correspondence that the jitter value total_jitter that ethernet data frame experiences from the source to the present node is bigger, by the user being tolerated the definite of threshold value, the jitter value total_jitter that ethernet data frame is experienced from the source to the present node tolerates that greater than the user ethernet data frame of threshold value abandons.

The jamming control method that detects based on shake in the aforesaid Metro Ethernet, wherein, the shake local_jitter that ethernet data frame experiences in switch obtains by following formula: local_jitter=delay-ave_delay, in the formula, delay is the time delay that ethernet data frame experiences in switch, and ave_delay is the average delay that ethernet data frame experiences in switch; Time delay delay calculates by following formula: delay=buf_oc*Packet_size/link_bw, and in the formula, buf_oc is the current cache occupation rate of output queue, and link_bw is available link bandwidth, and Packet_size is a data packet length; Average delay ave_delay calculates by the exponent-weighted average of the time delay of each ethernet data frame, and formula is ave_delay=ave_delay ^*(1-W _d)+delay ^*W _d, W _dBe weight.

The jamming control method that detects based on shake in the aforesaid Metro Ethernet, wherein, in when judgement of abandoning of carrying out packet, the jitter value total_jitter that ethernet data frame experiences from the source to the present node adds the value of the jittering counting device of the priority correspondence among the C-VLANTag that carries in the ethernet data frame for local jitter value, work as total_jitter〉user_threshold/2 or total_jitter＜-user_threshold/2, then this ethernet data frame just is dropped, (user_threshold/2 user_threshold/2) is user's tolerance threshold range; If total_jitter tolerates within the threshold value the user, so then upgrade jittering counting device counter, and the value of counter is converted to corresponding priorities among the C-VLAN Tag.

The jamming control method that detects based on shake in the aforesaid Metro Ethernet, wherein, the thresholding that can tolerate according to the user is divided into 8 different grades with jitter value, on the occasion of 4 grades, 4 grades of negative value.

The jamming control method that detects based on shake in the aforesaid Metro Ethernet wherein, is represented the positive and negative of jitter value with the lowest order of three bits of the priority among the C-VLAN Tag, 0 expression shake on the occasion of, negative value is shaken in 1 expression.

The jamming control method that detects based on shake in the aforesaid Metro Ethernet, wherein, in the priority high two respectively expression shake on the occasion of or 4 grades of negative value, promptly from grade 0 to grade 3, the binary bits of grade 0 correspondence is 11, the binary bits of grade 1 correspondence is 10, and the binary bits of grade 2 correspondences is 0l, and the binary bits of grade 3 correspondences is 00.

Method provided by the present invention detects the delay variation of multimedia data stream in the network node experience, and the next counter as the node delay variation of the C-VLAN Tag that utilizes the Q-in-Q in the Metro Ethernet, tolerate that above the user Frame of threshold value improves the service quality of multimedia service by abandoning those.The result shows that this method is compared the friendly that not only can keep TCP with traditional jamming control method of RED and DropTail, and the shake of average delay that can more effective reduction media stream, and can improve the effective throughput of multimedia service.

Description of drawings

Fig. 1 is the Ether frame form that has Q-in-Q VLAN tag label.

Fig. 2 is the corresponding relation between C-VLAN Tag priority and the jitter value.

Fig. 3 is a l-G simulation test topological diagram of the present invention.

Fig. 4 is the throughput schematic diagram of TCP stream under different frame length situations.

Fig. 5 is the throughput schematic diagram of UDP stream under different frame length situations.

Fig. 6 is the UDP under different frame length situations and the ratio chart of TCP throughput.

Fig. 7 is the average delay shake comparison diagram of frame length 128 bytes.

Fig. 8 is the average delay shake comparison diagram of frame length 512 bytes.

Fig. 9 is the average delay shake comparison diagram of frame length 1024 bytes.

Figure 10 is the comparison diagram of the effective throughput of frame length 1024 bytes.

Fig. 1 l is for realizing the structure chart of the ply-yarn drill exchange chip in the S4606 switch of the present invention.

Figure 12 is for using the software system structure block diagram of the inventive method.

Embodiment

Below in conjunction with drawings and Examples the present invention is described in detail.

Based on the basic thought of shaking the jamming control method (JDCC) that detects is its delay variation of detection and tracking in the multi-medium data stream transmission procedure, if delay variation has surpassed user's tolerance threshold value, so just this packet is abandoned, can reduce unnecessary bandwidth waste like this.In order to keep following the tracks of the delay variation that the multi-medium data frame is accumulated in transmission course end to end, need a jittering counting device.Obviously certain field that this counter will be in Frame is recorded in the shake of being experienced in the whole transmission course.In current Metro Ethernet, the networking mode of Q-in-Q is very general, and in the networking mode of Q-in-Q, after entering carrier network, transmit decision and priority policy all at the S-VLANTag label, and three bits of the priority of C-VLAN Tag label are obsolete, only after arriving user side, C-VLAN Tag just works, and it is very little at user side congested possibility to have taken place basically, and this time, the effect of jittering counting device also can stop.Therefore, utilize the priority of C-VLAN Tag label to come in the method for the invention as the jitter value counter in this ethernet data frame transmission course.

Priority in the C-VLAN Tag label is obsolete in the transmission course of carrier network, so three priority bit positions among the JDCC method employing C-VLAN Tag are as the counter of delay variation.But when Frame entered the network node of user side, S-VLAN Tag was peeled off, and C-VLAN Tag works, and at this moment the priority among the C-VLAN Tag is again the counter of jitter value simultaneously.For multimedia data stream, the less data of those jitter values should preferentially be transmitted, so higher priority of their correspondences, and for the very big Frame of those jitter values, may be nonsensical to the user, when generation is congested, should at first abandon, so corresponding lower priority.The corresponding relation that can draw thus between jitter value and the priority is the bigger lower priority of jitter value correspondence, and the corresponding higher priority of lower jitter value.Regulation is represented the positive and negative of jitter value with the lowest order of three bits of jitter value among the present invention, and 0 expression shake is on the occasion of, 1 expression shake negative value.The thresholding that can tolerate according to the user is divided into 8 different grades with jitter value, on the occasion of 4 grades, and 4 grades of negative value.In the priority high two respectively expression shake on the occasion of or 4 grades of negative value, promptly from grade 0 to grade 3.High priority is low more more owing to jitter value, and jitter value low priority is high more more.So the binary bits of grade 0 correspondence is 11; The binary bits of grade 1 correspondence is 10, and the binary bits of grade 2 correspondences is 01, and the binary bits of grade 3 correspondences is 00.Such as: if jitter value in positive grade 0, corresponding priorities is so: high two lowest order is 0 for bit 1, and promptly corresponding priorities is 6.If jitter value is in the negative grade 0, team's priority of answering is 7 so.And if jitter value is higher, drop in the grade 3, corresponding priorities is 0 or 1 so.From top analysis as can be seen, the value of priority just in time is the negate computing of jitter value grade.The visible Fig. 2 of concrete corresponding relation.

If user data is before entering carrier network, itself just be with priority in the C-VLAN Tag label of user data, so can be in S-VLAN Tag label with the priority mapping in the user C-VLAN Tag label.Priority in the C-VLAN Tag label just can be used for identifying the grade of shake thresholding like this.If enter in the C-VLAN Tag label of the user data before the carrier network and be not with priority, adopt above-mentioned method to carry out the distribution of priority so.Concrete code is as follows:

Said procedure has been described the false code of JDCC method, variable delay is the time delay that ethernet data frame experiences in switch, and time delay can be estimated in the following manner: the current buffer occupation rate (buf_oc) of output queue and available link bandwidth (link_bw):

Delay=buf_oc*Packet_size/link_bw; (1) average delay that experiences in switch of Frame calculates by the exponent-weighted average of the time delay of each ethernet data frame, as shown in Equation (2):

Ave_delay=ave_delay* (1-W _d)+delay*W _d(2) by exponent-weighted average, the increase of short-term averaging time delay can effectively smoothly be fallen.The shake that ethernet data frame experiences in switch then provides by following formula:

Local_jitter=delay-ave_delay; (3) the jitter value total_jitter that experiences from the source to the present node of ethernet data frame is the value that local jitter value adds the jittering counting device of the priority correspondence among the C-VLAN Tag that carries in the ethernet data frame.If total_jitter has surpassed user's tolerance threshold value, that is:

total_jitter>user_threshold/2?or?total_jitter<-user_threshold/2

(4) then this ethernet data frame just is dropped.If total_jitter tolerates within the threshold value the user, so then upgrade jittering counting device counter, and counter is converted to corresponding priorities among the C-VLAN Tag according to the corresponding relation among Fig. 2.

The delay variation value of the packet that the performance of JDCC method can be received by measurement obtains, and adopts sampling average delay D, owing to use the sampled packet N of limited quantity _sBy using the Cheyshev inequality, can draw the probability upper limit P (drop) that abandons grouping.Suppose and receive grouping D _iTime delay between be separate, with standard deviation sigma _dWith distributing.Sampling average delay D is:

$\stackrel{\&OverBar;}{D}=\frac{1}{N_s}{\mathrm{\&Sigma;}}_{i=1}^N{D}_i---\left(5\right)$

The Chebyshev inequality can be used for obtaining a loose upper limit, can obtain abandoning the probability P (drop) of grouping, and the time delay D of receiving grouping _iWith average time delay D, adopt quantity N _sAnd abandon relation between the thresholding.Abandoning thresholding user_th is defined as:

user_th＝2α _jJ _max； (6)

J _MaxBe the maximum delay shake that allows, α _j∈ (0,1).Then drop probability P (drop) is:

$P\left(\mathrm{drop}\right)=P(|D_i-\stackrel{\&OverBar;}{D}|\&GreaterEqual;\frac{\mathrm{user}\_\mathrm{th}}{2})---\left(7\right)$

Can get according to the chebyshev inequality:

$P\left(\mathrm{drop}\right)\&le;\frac{4{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="d"\; filenames="C200710151204E00182.png,C200710151204E00201.png"\; state="\{\{state\}\}">d</figure-callout>}}^2}{\mathrm{user}\_{\mathrm{th}}^2}---\left(8\right)$

Formula (7) can be written as:

$P\left(\mathrm{drop}\right)=P\left\{(\stackrel{\&OverBar;}{D}\&GreaterEqual;D_i+\frac{\mathrm{user}\_\mathrm{th}}{2})\mathrm{or}(\stackrel{\&OverBar;}{D}\&le;D_i-\frac{\mathrm{user}\_\mathrm{th}}{2})\right\}\&le;\frac{4{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="d"\; filenames="C200710151204E00182.png,C200710151204E00201.png"\; state="\{\{state\}\}">d</figure-callout>}}^2}{\mathrm{user}\_{\mathrm{th}}^2}---\left(9\right)$

Can be got by formula (8), given user's maximum tolerance threshold value or maximum delay jitter value can obtain the upper limit of drop probability, and from formula (8) as can be seen, it is high more to abandon threshold value, and drop probability is then more little.Simultaneously, formula (8) gives the UPS upper performance score of drop probability.

Fig. 3 has provided the topological diagram of test.Main frame is by linking in the carrier network on user's access switch (the customer switch among the figure) in topological diagram.The data flow of sending from subscriber's main station is not to be with VLAN Tag label, after data flow enters subscriber exchange, stamping C-VLAN Tag label for data flow on the subscriber exchange with on the port that subscriber's main station links to each other, distribute C-VLAN according to port here.The VLAN ID of subscriber's main station A1～A4 is respectively 100-400 in the test, but is not with priority.On the first line of a couplet port of subscriber exchange, be configured to VLAN TRUNK pattern.The data flow of coming out from user's first line of a couplet port has just had C-VLAN Tag label like this.After entering the SP switch, the SP switch comes the label for distribution of flows S-VLAN Tag according to port, and customer traffic has just become the Q-in-Q pattern after entering the SP network.In the test, the SP VLAN of the port on the SP switch that two subscriber exchanges link to each other is respectively 1000 and 2000.Because JDCC method of the present invention moves in the SP network, mainly be the carrier network part so l-G simulation test relates to, i.e. SP Networks part among Fig. 3.

Send the data flow of TCP (adopting the stream of FTP) and UDP (multimedia stream, such as video or the audio frequency of MPEG-2, bit rate is 6Mb/s) on two subscriber's main stations that each subscriber exchange connects respectively.The setting subscriber exchange is 10Mb/s to the link capacity of SP switch, and transmission delay is 3ms.The sample of MPEG-2 video flowing adopts the VBR form.For the UDP stream of testing that adopts is simply in the following discussion represented video flowing.

Frame will pass a link that 3 SP switching nodes are arranged in test, arrive the user side switch of the other end then.Node between the SP switch is mutually interconnected by the bandwidth link of 10Mb/s each other, and the propagation delay between them is 2ms.Link between these 3 SP switches is the bottleneck on the whole link.RED and JDCC method realize on the SP switching node, are used for carrying out TCP stream and UDP stream are carried out congested control.That sets in test on each node Frame in the formation is restricted to 150; Test 256,512 in the test, 1,024 three kinds of frame lengths, the testing time was 10 seconds.The rate setting of the stream of each UDP and TCP is 6Mbps.The threshold setting of JDCC method is 0.1s, weights W _dBe set to 0.02; Parameter in the RED method is set to 100 frames of max threshold, and minimum is 30 frames.In test mainly from the friendly of TCP, three aspect comparative analysiss of effective throughput of delay variation and media stream the performance of JDCC method and traditional jamming control method RED, DropTail.

UDP is a kind of connectionless protocol, and it does not have corresponding flow control mechanism, can not make respective handling to packet loss.When UDP stream and TCP stream shared bandwidth, this specific character of UDP stream can be greedy takies a large amount of bandwidth, and reduces the throughput that TCP flows.Therefore, it is very important that the JDCC method that the present invention proposes has definite TCP friendly, and it should not reduce the throughput of TCP stream, to reach and traditional RED the friendly that the DropTail method is identical.In test, TCP stream and UDP are banished before the fifo queue, use different queue management strategies respectively at them.Result of the test among Fig. 4 has provided TCP stream at the different pieces of information frame sign, the average throughput under the situation of different queue management strategy.Come as can be seen from result of the test, than traditional jamming control method, the JDCC method can reach the TCP throughput with RED, DropTail level of approximation.And provided the throughput of UDP stream under the situation of different pieces of information frame sign among Fig. 5, employing be with Fig. 4 in identical test.Adopt the average throughput of JDCC method UDP to reduce along with the increase of Frame, it is less than DropTail method and RED method simultaneously.This is because some UDP grouping has accumulated a large amount of delay variations and has been dropped.The TCP friendly of method is directly relevant with the ratio of UDP throughput and TCP throughput.Fig. 6 has provided the ratio between UDP under the different frame lengths and TCP throughput.As can be seen, the JDCC method is compared with other ripe jamming control methods and is being close aspect the TCP friendly from test.

Delay variation is a very important parameter index for multimedia service, guarantees that the less time delay shake is an important means that improves QoS in Multimedia Service.The delay variation of multimedia data stream during with the RED method is defined as follows: J _i=| delay-ave_delay|.Fig. 7, Fig. 8, Fig. 9 have provided the comparison diagram of the mean value of the delay variations of three kinds of methods under different frame length situations respectively.As can be seen from the figure, the JDCC method is under the situation of various frame lengths, and the average delay jitter value all will be starkly lower than traditional RED and DropTail method.It can also be seen that from figure three kinds of methods comprise DropTail, RED has identical curvilinear trend with the JDCC method, and this periodic trend mainly is because the flow control mechanism of Transmission Control Protocol itself causes.Result of the test among the figure shows that the JDCC method that the present invention proposes can reach lower delay variation than DropTail and RED method.

Multimedia UDP stream is very sensitive for delay variation.When the delay variation of grouping than the big situation of user's tolerance threshold value under, even grouping by success be sent to client, in fact nonsensical for the user.If there are enough bandwidth may increase so congested possibility not taking place, transmits the available bandwidth of these delay variations bigger grouping will minimizing transmission channel on switching node.And can cause other some Frames to increase delay variation, cause the more data frame to become data nonsensical concerning the user.So in the JDCC method, will tolerate that the Frame of threshold value abandons above the user.

In fact in the transmission course of data, what be concerned about concerning the user is singly not to be throughput to its useful output.Here " effectively " means satisfies the shake tolerance when multi-medium data frame arrives client, the Frame that just arrives user side will satisfy jitter＜user_threshold.The multi-medium data frame of " effectively " can satisfy the position that arrived it before playing to the user.The user tolerates that threshold value is arranged in test in the 0.03-0.1 scope of second.Result of the test is shown in Figure 10, and result of the test shows that the JDCC method can reach higher average effective throughput than DropTail method and RED method.Whole trend can see very clearly that in Figure 10 this explanation result of the test is consistent with theory analysis.

The hardware platform that the JDCC method adopts is mainly based on the S4606 switch.The JDCC method is mainly finished on LC-48FE hundred megaline cards (48x100-BASET ply-yarn drill).

JDCC method of the present invention mainly is to realize on the exchange chip of LC-48FE hundred megaline cards, the structure of exchange chip as shown in figure 11, ply-yarn drill among S4606 exchange core mainly contains following a few part and forms:

1) packet resolver (Packet Parser): major function is that the packet of coming in is resolved;

2) modifier of packet (Packet Modifier): mainly be to be used for to the packet that L3 the transmits functions such as next-hop MAC of making amendment;

3) CPU management interface controller (CMIC:CPU Management InterfaceController): mainly be to be responsible for and CPU interface partly; Comprising pci interface (PCIInterface) and management counter (Management counters);

4) Packet Filtering processor (Packet Filter Processor): the filtering function of mainly finishing packet;

5) memory management unit (Memory Management Unit): mainly be responsible for cache management and traffic management in the exchange chip; Be divided into two parts in two partial buffering administrative units (BufferManagement) and traffic management unit (Traffic Management).Memory management unit (Buffer Management): the JDCC algorithm of mainly finishing functions such as the management of the buffer queue in the exchange chip and access control is just finished in this part; Traffic management unit (Traffic Management): mainly finish the speed limit of flow, shaping, functions such as the management of formation and scheduling.

6) two layers of forwarding logic (L2 Switching): mainly be two layers of forwarding finishing packet;

7) three layers of forwarding logic (L3 Switching): mainly be the forwarding of finishing three layers (IP layers) of packet.

Software platform of the present invention is an Embedded operating system and at the protocol stack of the TCP/IP on upper strata; Embedded OS adopts VxWorks 5.5 kernels; The main USP of Firberhome Networks company exploitation that adopts of protocol stack part unifies software platform; Software Development Platform adopts the version of Tornado2.2 for PPC.The system configuration of whole software as shown in figure 12.Relate to the explanation of english abbreviation among Figure 12:

ARP (Address Resolution Protocol): address resolution protocol;

BGP (Border Gateway Protocol): border webmaster Routing Protocol;

CLI (Command Line Interface): order line;

Console: control desk serial ports;

DCP (Distributed Communication Protocol): distributed communication agreement;

DHCP (Dynamic Host Configuration Protocol): DHCP;

FTP (File Transmission Protocol): file transfer protocol (FTP);

GARP (General Attribute Register Protocol): General Attribute Registration Protocol

GMRP (GARP Multicast Register Protocol): GARP Multicast Registration Protocol;

GVRP (GARP Vlan Register Protocol): GARP Vlan log-in protocol;

HA (High Availability): high availability;

IDB (Information Database): information database;

ICMP (Internet Control Message Protocol): Internet control messages agreement;

IGMP (Internet Group Management Protocol): Internet igmpinternet;

IP (Internet Protocol): Internet protocol;

MSTP (Multiple Spanning Tree Protocol): Multiple Spanning Tree Protocol

OSPF (Open Shortest Path First): OSPF;

PIM/SM (Protocol Independent Multicast/Sparse Mode): sparse mode is independent

Multicast protocol;

RIP (Routing Information Protocol): routing information protocol;

RSTP (Rapid Spanning Tree Protocol): RSTP;

SNMP (Simple Network Management Protocol): Simple Network Management Protocol;

STP (Spanning Tree Protocol): Spanning-Tree Protocol;

TCP (Transmission Control Protocol): transmission control protocol;

Telnet: remote terminal protocol;

TFTP (Trivial File Transmission Protocol): TFTP;

UDP (User Datagraph Protocol): User Datagram Protoco (UDP).

List of references

[1]S.P.?Chan?and?C.W.Kok，“Protocol?and?buffer?design?for?multimedia-on-demand?system，”inProc.IEEE?PCM，Oct.2001，pp.1010-1015.

[2]，“Bitrate?adaptation?flow?control?for?multimedia-on-demand，”in?Proc.IEEE?ICC，May?2002，pp.2503-2507.

[3]S.Floyd?and?V.Jacobson，“Random?early?detection?gateways?for?congestion?avoidance，”IEEE/ACM?Trans.Netw.，vol.1，no.4，pp.397-413，Aug.1993.

[4]S.Floyd，A?report?on?some?recent?developments?in?TCP?congestion?control.IEEECommunication?Magazine，April?2001.

[5]S.De?Cnodder，O.Elloumi，and?K.Pauwels，“RED?behavior?with?different?packet?sizes，”inProc.5th?IEEE?Symp.Computers?Commun.，2000，pp.793-799.

[6]J.B.Pippas?and?I.S.Venieris，“A?RED?variation?for?delay?control，”in?Proc.ICC，vol.1，2000，pp.475-479.

[7]S.Floyd?and?R.Mahajan，“Controlling?High-Bandwidth?Flows?at?the?Congested?Router，”，Nov.20，2000.Tech.Rep.of?ACIRI.

[8]Cisco?system.Distributed?weighted?random?early?detection.http://cco.cisco.com

[9]Ott?Teunis?J，Lakshman?T?V，Wong?Larry?H.SRED：Stabilized?RED.In?IEEE?INFORCOM’99New?York，USA，1999.1346～1355.

[10]Feng?W，Kandlur?D，Saha?D，Shin?K.A?self-configuring?RED?gateway.In?Proc.IEEEINFOCOM’99.New?York，USA，1999.1320～1328

[11]Floyd?S，Gummadi?R，Shenker?S.Adaptive?RED：An?algorithm?for?increasing?the?robustnessof?RED’s?active?queue?management.http://www.cs.berkeley.edu/

[12]Lin?D，Morris?R.Dynamics?of?random?early?detection.In?Proceedings?of?the?ACMSIGCOMM?Conference?on?Applications，Technologies，Architectures，and?Protocols?for?ComputerCommunications.New?York，USA，1997.127～138

[13]Anjum?F，Tassiulas?L.Balanced-RED：Analgorithm?to?achieve?fairness?in?Internet.InProceedings?of?IEEE?INFOCOM’99.New?York，USA，1999.

[14]IEEE?Std?802.1ad ^TM-2005，IEEE?Standard?for?Local?and?Metropolitan?Area?Networks：VirtualBridged?Local?Area?Networks-Amendment?4：Provider?Bridges.

Claims (7)

1. the jamming control method that detects based on shake in the Metro Ethernet, it is characterized in that: this method utilizes priority bit position among the C-VLAN Tag of the Q-in-Q in the Metro Ethernet as the counter of node delay variation, follow the tracks of the delay variation that the multi-medium data frame is accumulated in transmission course end to end, the corresponding higher priority of data that the jitter value total_jitter that ethernet data frame experiences from the source to the present node is less, the lower priority of data correspondence that the jitter value total_jitter that ethernet data frame experiences from the source to the present node is bigger, by the user being tolerated the definite of threshold value, the jitter value total_jitter that ethernet data frame is experienced from the source to the present node tolerates that greater than the user ethernet data frame of threshold value abandons.

2. the jamming control method that detects based on shake in the Metro Ethernet as claimed in claim 1, it is characterized in that: the shake local_jitter that ethernet data frame experiences in switch obtains by following formula: local_jitter=delay-ave_delay, in the formula, delay is the time delay that ethernet data frame experiences in switch, and ave_delay is the average delay that ethernet data frame experiences in switch.

3. the jamming control method that detects based on shake in the Metro Ethernet as claimed in claim 2, it is characterized in that: time delay delay calculates by following formula: delay=buf_oc*Packet_size/link_bw, in the formula, buf_oc is the current cache occupation rate of output queue, link_bw is available link bandwidth, and Packet_size is a data packet length; Average delay ave_delay calculates by the exponent-weighted average of the time delay of each ethernet data frame, and formula is ave_delay=ave_delay* (1-W _d)+delay*W _d, W _dBe weight.

4. the jamming control method that detects based on shake in the Metro Ethernet as claimed in claim 1, it is characterized in that: in when judgement of abandoning of carrying out packet, the jitter value total_jitter that ethernet data frame experience from the source to the present node adds the value of the jittering counting device of the priority correspondence among the C-VLAN Tag that carries in the ethernet data frame for the jitter value of this locality.

5. the jamming control method that detects based on shake in the Metro Ethernet as claimed in claim 4, it is characterized in that: if total_jitter user_threshold/2 or total_jitter＜-user_threshold/2, then this ethernet data frame just is dropped, (user_threshold/2 user_threshold/2) is user's tolerance threshold range; If total_jitter tolerates within the threshold value the user, so then upgrade jittering counting device counter, and the value of counter is converted to corresponding priorities among the C-VLAN Tag.

6. the jamming control method as detecting based on shake in claim 1 or 2 or 3 or the 4 or 5 described Metro Ethernets, it is characterized in that: the thresholding that can tolerate according to the user is divided into 8 different grades with jitter value, on the occasion of 4 grades, 4 grades of negative value.

The jamming control method that detects based on shake in 7 Metro Ethernets as claimed in claim 6 is characterized in that: the lowest order with three bits of the priority among the C-VLAN Tag is represented the positive and negative of jitter value, 0 expression shake on the occasion of, negative value is shaken in 1 expression.

8. the jamming control method that detects based on shake in the Metro Ethernet as claimed in claim 7, it is characterized in that: in the priority high two respectively expression shake on the occasion of or 4 grades of negative value, promptly from grade 0 to grade 3, the binary bits of grade 0 correspondence is 11, the binary bits of grade 1 correspondence is 10, the binary bits of grade 2 correspondences is 01, and the binary bits of grade 3 correspondences is 00.

Images