CN1416633A - Unified algorithm for frame scheduling and buffer management in differentiated services networks - Google Patents
Unified algorithm for frame scheduling and buffer management in differentiated services networks Download PDFInfo
- Publication number
- CN1416633A CN1416633A CN01805246A CN01805246A CN1416633A CN 1416633 A CN1416633 A CN 1416633A CN 01805246 A CN01805246 A CN 01805246A CN 01805246 A CN01805246 A CN 01805246A CN 1416633 A CN1416633 A CN 1416633A
- Authority
- CN
- China
- Prior art keywords
- frame
- type
- formation
- network environment
- scheduling
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/28—Flow control; Congestion control in relation to timing considerations
- H04L47/283—Flow control; Congestion control in relation to timing considerations in response to processing delays, e.g. caused by jitter or round trip time [RTT]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/24—Traffic characterised by specific attributes, e.g. priority or QoS
- H04L47/2441—Traffic characterised by specific attributes, e.g. priority or QoS relying on flow classification, e.g. using integrated services [IntServ]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/30—Flow control; Congestion control in combination with information about buffer occupancy at either end or at transit nodes
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/32—Flow control; Congestion control by discarding or delaying data units, e.g. packets or frames
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/50—Queue scheduling
- H04L47/62—Queue scheduling characterised by scheduling criteria
- H04L47/625—Queue scheduling characterised by scheduling criteria for service slots or service orders
- H04L47/6275—Queue scheduling characterised by scheduling criteria for service slots or service orders based on priority
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
A frame forwarding and discard architecture (100) in a Differentiated Services network environment. The architecture (100) comprises a discard logic (102) for discarding a frame from a stream (104) of incoming frames of the network environment in accordance with a discard algorithm, the frame being discarded if a predetermined congestion level in the network environment has been reached, and a predetermined backlog limit of a queue associated with the frame, has been reached. Scheduling logic (118) is also provided for scheduling the order in which to transmit one or more enqueued frames of the network environment.
Description
The application requires to enjoy the priority of submitting on February 24th, 2000, be called No. 60/184557 United States Patent (USP) provisional application of " frame scheduling in differentiated services networks and buffer management unified algorithm " according to the 119th (e) money in No. 35 bill of US Congress.
Technical field
The present invention relates to the network exchange technology, more specifically, the Frame retransmission technique that relates in network exchange to be adopted.
Background technology
Differentiated Services (DiffServ) is a relative newer notion that occurs in the internet arena, a kind of " soft " approach of its be considered to improve service quality (QoS).Differentiated Services is the one group of new technology that is proposed by IETF (the Internet engineering duty group), this technology allows Internet service merchant and other Internet Service Provider based on the IP agreement to provide level different service for each unique user of additional payment, and sends information flow to these users.Under this system, the frame head that enters into each Frame of network router all includes a mark, and this mark has indicated this network router should be given which kind of level of this Frame in transmission course service.Then, this network router is used corresponding different service to each Frame that enters into each port.Adopt the method for Differentiated Services, the service provider then just can provide the service (but this is not strict constant promise) of priority level according to suitable frame flag contained in the frame head to all data frame stream (traffic) of some specific user.The priority of the service that provides is high more, and then frame time-delay (being that frame is waited for duration) is short more.When the generation frame is congested, these Frames that have priority flag will be enjoyed preferential service.
Transmit mechanism for present differentiation, because under condition of severe the most, as reduce the requirement that usage factor of system resource just can not guarantee frame time-delay and bandwidth isolation simultaneously not seriously, so this mechanism is incomplete.Will satisfy frame time-delay assurance and bandwidth simultaneously and guarantee can have a kind of frame to transmit scheduling mechanism with regard to needs, this mechanism combines the buffer management technology and transmits dispatching technique.
Summary of the invention
Open and claimed one aspect of the present invention has been to provide a kind of and has been used in frame scheduling in the differentiated services network environment, abandoned framework (architecture) in this literary composition.Comprise one in this framework and abandon logical circuit, it is used for according to a kind of discard algorithm a Frame being abandoned away from network environment incoming frame stream, if just abandon this Frame when predetermined extent and the overstocked length of formation relevant with this Frame reach predetermined limits when the congestion level in the network environment reaches.Also be provided with the scheduling logic circuit, be used for the order of transmission of the one or more data queued frames of network environment is dispatched.
Description of drawings
In order a more complete understanding to be arranged to the present invention and advantage thereof, hereinafter do following description in conjunction with the accompanying drawings, in the accompanying drawings:
Fig. 1 is a The general frame, has represented a kind of frame repeater system according to disclosed embodiment;
Figure line among Fig. 2 has been represented formed congested of system shown in Figure 1;
Block representation among Fig. 3 according to the frame repeater system sample of table 1 design; And
Figure line among Fig. 4 has been represented inferior congested in a WRED type is used.
Embodiment
Novel scheduling mechanism disclosed herein preferably is attached to a kind of being used in the integrated processes that differentiated service environment execution frame is transmitted with measurable Quality of Service criterion and buffer management, and quality of service criterion wherein for example is time-delay or bandwidth.
QoS is a notion that implication is very wide in range, and different people has different explanations to this.On overall, the method for raising QoS described herein is based on the supposition of several aspects: the Business Stream pattern that is provided is unknown; The professional right and wrong rule of inlet flow or (still,, then can make some other assurances) of shaping not to switching performance if inlet flow is rule or shaping; And network manager is known the application type (or stream type) that carries out and the relative importance of these application, the voice communication in this way of application examples wherein, file transfer or network browsing on this network.Term " shaping " or " shapingization " are defined as the control (being flow control) to the Business Stream flow, promptly by the Business Stream flow restriction is prevented that in the degree of the input tape tolerance that is in close proximity to upstream device overflowing from appearring in upstream device.Regularization and shapingization are similar, and still, in the regularization process, exceeding the Business Stream of setting flow is to be dropped, rather than buffer memory gets up.Have under the prerequisite of these working knowledges, network manager then just can be divided into several types with various application, and sets up a service level agreement for all types of application, and this service level agreement for example is made up of the bandwidth of each type application or the assurance of delaying time.
One type can provide the flow that surpasses the agreement bandwidth.One the flow that the service traffics that provide of rule type are no more than agreement is provided.In contrast, the stream business that type provided in violation of rules and regulations then can surpass the agreement flow.A violation type is to be gathered by little Business Stream of a plurality of violations to form.In order to realize the utilization of high link bandwidth, a violation type is allowed to use any redundancy bandwidth.But, must exceed not damage other QoS that follows the rule type to the tolerance of violation type.
Represented an example form about six types of streams in the following table 1, all kinds stream all has exclusive separately attribute and application in the form.
Six types of streams of table 1. show the sample form
| Type C nTarget | Obtain the total bandwidth of assurance | Low drop probabilities (low packet loss type) | High drop probabilities (high packet loss type) |
| Transmit the highest C of priority 3 | ? ??40Mbps | Use: telephone service (voIP), circuit simulation stand-by period: less than the 1ms packet drop: if C 3Do not transmitted, then do not have packet loss by excess | Use: the video teaching stand-by period: less than the 1ms packet drop: if C 3Do not transmitted, then do not have packet loss, otherwise, then at first be dropped by excess |
| Transmit the medium C of priority 2 | ? ??35Mbps | Use: interactive application, ecommerce stand-by period: less than the 4ms packet drop: if C 2Do not transmitted, then do not have packet loss by excess | Use: the non-key interactive application stand-by period: less than the 4ms packet drop: if C 2Do not transmitted, then do not have packet loss, otherwise, then at first be dropped by excess |
| The low C that transmits priority 1 | ? ??25Mbps | Use: Email, the document backup stand-by period: hope can be less than 16ms, but is not vital packet drop: if C 1Do not transmitted, then do not have packet loss by excess | Use: random network is browsed the stand-by period: hope can be less than 16ms, but is not vital packet drop: if C 1Do not transmitted, then do not have packet loss, otherwise, then at first be dropped by excess |
| Amount to | ??100Mbps |
As shown in table 1, various stream types (being that telephone service, circuit simulation, video teaching, key and non-key interactive application, ecommerce, Email, document backup and random network are browsed) have been divided into three type (C
1, C
2, C
3), each type all has been assigned with certain bandwidth assurance and time-delay gauge.Type C
3Be the highest transmission type of priority, all Frames all need be sent in the time less than 1ms, and taken 40Mbps (40%) among the port total bandwidth 100Mbps.Type C
2Be the type of a high medium priority, it has taken the 35Mbps (or being 35%) among the port total bandwidth 100Mbps, and needs all Frames all should send in the time less than 4ms.At last, C
1Type is the minimum type of priority, and it has taken the 25Mbps (or being 25%) among the port total bandwidth 100Mbps, and need all Frames all should be before packet loss takes place, in time, send less than 16ms.
In addition, each transmission type (C
1, C
2And C
3) all have two subgroups: high packet loss subgroup and low packet loss subgroup.The situation of frame losing should seldom appear in the user who follows rule.But user in violation of rules and regulations (being the user that the high flow of ether sends Frame) will lose some Frames, and what at first be dropped will be those Frames that meet high packet loss criterion.If do like this and also be not enough to alleviate congestion, then some Frame that meets low packet loss criterion also will be dropped, and in the worst case, all Frames all will be dropped.
As can be seen from Table 1: the application of type, corresponding attribute and time-delay and packet loss criterion can be arranged in pairs or groups according to any desirable mode.For example, random network is browsed and has been in high packet loss and allows in the stream type of long delay, and voip phone then has been in the stream type of low packet loss ratio, short time delay.
Except three above-mentioned type (C
1, C
2And C
3) outside, can also mark off more a plurality of transmission types with other time-delay gauge and minimum bandwidth assurance.In addition, (best-effort) stream that in another kind of variations, does the best possible can constitute a type that priority is minimum, this stream only other type all without any stream the time just can enjoy bandwidth.Can also increase a kind of higher type of priority that transmits, it has the absolute priority that is higher than other three (or more a plurality of) types; That is to say, if, then also will at first transmit it even the type has only a Frame to transmit.But should be noted in the discussion above that in this specific embodiment each 10/100Mbps port is supported three type (C altogether
1, C
2And C
3).
In the applied environment of a 1Gbps, because line speed is high more, desired QoS guarantees high more, so each port may be supported nearly eight type (C
8-C
1).For example, a kind of default configuration form may have six time-delay gauge formation Q
8-Q
1(correspond respectively to Type C
8-C
1) and two formation Q that do the best possible
2And Q
1(correspond respectively to Type C
2And C
1).For the situation of 1Gbps port, the setting of time-delay gauge for example is: C
8And C
7Be 0.16ms, C
6Be 0.32ms, C
5Be 0.64ms, C
4Be 1.28ms and C
3Be 2.56ms.Have only when needing service, just provide service the type Business Stream that does the best possible without any time-delay gauge stream.For the situation of this 1Gbps port, have two formations that do the best possible, and type high priority is big more more (that is to say C
2To C
1Has absolute priority).Description equally, above is a kind of sample situation.Should be noted that: the IETF classification of type that above-mentioned framework and the Internet engineering duty group propose is compatible.
For solving owing to the uncertainty that produces is not understood in the mixing to inlet flow, adopted a kind of time-delay to guarantee algorithm, dynamically adjust according to the required moment of formation occupy-place situation and formation HOL frame and dispatch criterion and packet loss criterion.The result is exactly: from reliability (confidence) access Frame, can both guarantee to reach time-delay gauge-even occurring under the congested situation of system-wide for all height.Time-delay guarantees that algorithm can identify type in violation of rules and regulations, and abandons some Frames intelligently, does not damage the purpose of following the rule type to reach.This algorithm can also distinguish high packet loss stream with RED (RandomEarly Detection the is earlier detection at random) method (WRED) of a weighting and hang down packet loss stream.The WRED method is used in the internet, is used for avoiding its appearance before becoming problem in that frame is congested.The RED algorithm is monitoring current load along a network on select location point, and when beginning to increase, abandons some Frames congested randomly.In response to abandoning the upper strata operation that Frame detects, the transmission speed of Frame has been slowed down.
Referring to Fig. 1, represented a block diagram among Fig. 1, this block diagram is a upper overview to disclosed embodiment.Novel forwarding mechanism disclosed herein comprises two aspects that combine: buffer management and transmission scheduling, buffer management wherein are used to determine the access of input data frame and abandon according to a kind of discard algorithm work; Transmit scheduling and then be used for the order that the transmission of specified data frame is left away.The importance of above-mentioned combination can be summarized as follows: the correlation between bandwidth, time-delay and the buffer memory by mathematics be expressed as the relational expression of " the bandwidth ∝ queue length of enjoying/experience time-delay ".This has utilized the merging mechanism of dispatching technique and buffer administrative skill to control delay duration and the queue length that is experienced.As the result of the above-mentioned fact and relationship, this associative mechanism has also been adjusted the bandwidth that each type is enjoyed.
Again referring to Fig. 1, frame repeater system 100 comprise one according to discard algorithm work abandon logical circuit 102, discard algorithm is wherein monitoring incoming bit stream 104.This output 106 that abandons logical circuit 102 flow into one or more formations 108,110 and 112 and (is also referred to as formation Q
1, Q
2..., Q
n) in, wherein each formation corresponding respectively each stream type C
1, C
2..., C
nFormation 108,110,112 is being stored these Frames according to each Frame appointed stream type of institute temporarily, and each formation outputs to Frame in the multiplexing logic circuit 114, finally to be input in the output queue 116, this output queue has the total bandwidth capacity of K Mbps.For example, transmit the minimum Type C of priority
1With a service level agreement S
1Be associated, this agreement is by time-delay gauge parameter (δ
1) and a bandwidth parameter (r
1) limit.(be also referred to as Q if be queued in formation 108
1) in the Frame number can not be by time-delay gauge parameter (δ
1) send in time of appointment, then just had and some Frame in the type will have been abandoned to prevent congested certain probability.Similarly, also expressed transmission priority time high Type C among the figure
2, the type and a service level agreement S
1Be associated, this agreement is by time-delay gauge parameter (δ
2) and a bandwidth parameter (r
2) limit.(be also referred to as Q if be queued in formation 110
2) in the Frame number how to can not be by time-delay gauge parameter (δ
2) send in time of appointment, then just had and some Frame in the type will have been abandoned to prevent congested certain probability.
There are a plurality of types in an illustrated embodiment, the C that its medium priority is the highest
nWith a service level agreement S
nBe associated, this agreement is by time-delay gauge parameter (δ
n) and a bandwidth parameter (r
n) limit.(be also referred to as Q if be queued in formation 112
n) in the Frame number how to can not be by time-delay gauge parameter (δ
n) send in time of appointment, then just there is certain probability: abandon to prevent some Frame in the type congested.Output queue 116 is being stored from each type formation 108,110 and 112 temporarily and is being received the Frame of coming, and with each Type C
1, C
2... C
nOutput is to a port P (not shown).Multiplexer 114 is by 118 controls of a scheduling logic circuit, and this logical circuit has been determined the order that Frame is left away from each type formation 108,110 and 112.
Hereinafter be to the more description of generalization of this novel system.Suppose that port P provides service for n stream type of service, this n type is designated as C
1, C
2... C
nNetwork provider is each COS C
iAll discussed and decided a service level agreement S
i, this agreement is δ
iAnd r
IExpression formula, i.e. S
i=(δ
i, r
i), δ in the formula
iBe Type C
iIn any Frame allowed that the longest experience guarantees time-delay, and r
iBe meant and be Type C
iThe overtime minimum bandwidth that guarantees of distributing.These types are configured to like this: make Type C
1Maximum guarantee time-delay δ
1More than or equal to Type C
2Maximum guarantee time-delay δ
2, and Type C
2Maximum guarantee time-delay δ
2More than or equal to Type C
3Maximum guarantee time-delay δ
3, and (that is to say δ by that analogy
1〉=δ
2〉=... 〉=δ
n).Disclosed favourable part that should mechanism is: what kind of the flow pattern formula that no matter is provided is, can both satisfy service level agreement S simultaneously
iThe time-delay constraints and the bandwidth constraint condition of (i is an all values).
Hereinafter the discussion to the scheduling of time-delay gauge is to carry out in the environment of 10/100Mbps port, and port wherein has three time-delay gauge type (C
3, C
2And C
1).But, also can make up applied environment similarly with more a plurality of types.Under the 10/100Mbps port case in table 1, when the time-delay of gauge is dispatched, be drained into three scheduling queue Q
1-Q
3(belong to Type C respectively
1, C
2And C
3) in Frame all comprise and stab time of advent.When team's head (HOL) position in the Frame arrival formation, make the scheduling decision according to the timestamp of HOL Frame in each formation.In the exemplary rule that provides hereinafter, time-delay is defined as work (or Frame) and stabs and the time difference between the current time time of advent.Obviously,, transmission to be waited for, then just the type can not be selected without any Frame if for a certain particular type.
Referring to Fig. 2, this figure has represented congested 200 notion according to disclosed embodiment in Euclidean space.With each COS C
iThe middle formation of waiting for that output port P transmits is overstock length (is unit of measurement with the total bytes) and is made as Q
iIf λ
i=δ
1/ δ
i, and establish D=K δ
1(is unit of measurement with the byte number).Congested hyperplane 200 is by Vector Groups { Q
1, Q
2, Q
3... Q
nSurround, and limit by equation (1):
And
Q
i＞r
i·δ
1
(3)
It is congested that first condition (formula [2]) shows that system 100 has taken place, and that is to say that system 100 has exceeded congested 200.Second condition (formula [3]) shows Type C
iBacklog very big.Even one belongs to Type C
iFrame get permission to have inserted, but can not make any change, this constraints result that to be existing backlog guarantee with minimum bandwidth to other type to its situation that will meet the constraints of delaying time.Thereby the input data frame of the type i will be dropped.
Disclosed buffer management algorithm can be modified to the WRED technology that comprises, the existing in the literature very sufficient introduction of the benefit of WRED technology.The WRED technology adopts a kind of queue length of weighting, comes the judgement system when to become very congested and need to consider to abandon one or more Frames.This abandons rule must abandon abundant Frame, queue length is remained on congested below 200; If not, will abandon 100% Frame prevents congested.Because target is to distinguish high packet loss stream and hang down packet loss stream, so can not allow system 100 to arrive congested 200 critical condition, on congested, all Frames all will be dropped, and the precedence when not considering to abandon.Thereby, in this certain embodiments, defining congested of two Asias (level 1 and level 2), these two congested faces in Asia are designed to realize early stage congestion-prevention, like this, Frame just seldom runs into the severe condition with 100% drop probabilities.
Referring to Fig. 3, the block diagram among the figure has been represented a kind of exemplary repeater system according to table 1.This frame repeater system 300 (being similar to system 100) has the bandwidth of 100Mbps, and has adopted and abandon logical circuit 102 according to discard algorithm work disclosed herein.This abandons logical circuit 102 and is monitoring incoming bit stream 302, and based on predefined criterion the selected data frame 304 in the bit stream 302 is discarded in the dustbin 306 (shown in the figure only is in order to set off a discussion).The Frame of those accesses (307,309,311) then is discharged in the defeated queuing (308,310 and 312) of corresponding types subsequently.For example, input rank 308 is to belong to C
1Type (transmitting the minimum type of priority), the time-delay gauge of the type requires all Frames 307 all should send in the time less than 16ms, if the type C
1When becoming a violation type when having surpassed agreed 25Mbps, just exist to abandon some C owing to the flow that provides
1The probability of type data frame prevents congested with this.Input rank 310 is C
2The formation of type (transmitting the medium type of priority), the time-delay gauge of the type requires all Frames 309 all should send in the time less than 4ms, if the type C
2When becoming a violation type when having surpassed agreed 35Mbps, just exist to abandon some C owing to the flow that provides
2The probability of type data frame prevents congested with this.At last, input rank 312 is C
3The formation of type (transmitting the highest type of priority), the time-delay gauge of the type requires all Frames 311 all should send in the time less than 1ms, if the type C
3When becoming a violation type when having surpassed agreed 40Mbps, just exist to abandon some C owing to the flow that provides
3The probability of type data frame prevents congested with this.
Be drained in each formation (308,310 and 312) corresponding data frame (307,309 and 311) by a multiplexer 314 (it is similar to above-mentioned multiplexing logic circuit 114), with the flow that is no more than 100Mbps be multiplexed to one output bit stream 316 in, 100Mbps flow value wherein is the port output speed of system 300.But, there is a scheduling logic circuit 318 to be connected with multiplexer 314, be used for the transmission of the corresponding types Frame (307,309 and 311) that transmits from each type formation (308,310 and 312) is dispatched.As mentioned above, each data queued frame (307,309 and 311) all has been coupled with a timestamp when arriving corresponding formation (308,310 and 312).When the Frame (307,309 and 311) of a certain type arrives a team position (313,315 and 317) of its corresponding formation (308,310 and 312), make the scheduling decision according to stabbing the time of advent of each formation squadron Frame.
Referring to Fig. 4, the figure line among the figure has been represented congested and inferior congested.Should be noted that definable goes out congested of any number Asia.Inferior congested (being respectively 400 and 402) of level 1 and level 2 a certain proportion ofly highly abandons Frame and keeps in large quantities simultaneously and lowly abandon Frame and prevent congested by abandoning randomly.This becomes and low abandon the victim of Frame and just begin in early days to be dropped with regard to making height abandon Frame.In this example, when already present total formation backlog N 120 between 200KB the time, and type formation Q
1-Q
3In the Frame backlog that is buffered of any one formation meet or exceed corresponding queues limit A, B or C (is unit with KB), just have the certain probability that abandons Frame.In inferior congested 400 of level 1, there is 16Q
3+ 4Q
2+ Q
1〉=120KB, and formation Q
1-Q
3In any one or the overstocked degree of a plurality of formations surpassed corresponding limit (being respectively A, B and C), the scope of low packet loss and high packet loss be respectively 0 and 0-X% between.Similarly, in inferior congested 402 of level 2, there is 16Q
3+ 4Q
2+ Q
1〉=160KB, and formation Q
1-Q
3In any one or the overstocked degree of a plurality of formations surpassed corresponding limit (being respectively A, B and C), the scope of low packet loss and high packet loss is respectively between 0-Y% and 0-Z%.At last, among congested 200 of the 3rd level, this congested face is by 16Q
3+ 4Q
2+ Q
1The relational expression of 〉=200KB limits, and is that high packet loss rule or low packet loss rule all stipulate to want 100% ground to abandon Frame.
Table 2 pair adopts the packet loss rule of WRED to conclude, and in this table, each inferior congested face is that the port according to 100Mbps defines total maximum queue backlog N=200KB in this concrete example.
Table 2. is the packet loss rule that improves QoS under on the 10/100Mbps port, for the situation with three time-delay gauge types.
| ???C 3 | ???C 2 | ???C 1 | Low packet loss | High packet loss | |
| Level 1 N=120 | ? ???Q 3≥C ???KB | ? ???Q 2≥B ???KB | ? ???Q 1≥A ???KB | ????0％ | ????X％ |
| Level 2 N=160 | ????Y％ | ????Z％ | |||
| Level 3 N=200 | ????100％ | ????100％ |
It should be noted that: in having table 1, in the specific embodiment of time-delay gauge, work as 16Q
3+ 4Q
2+ Q
1During 〉=N KB, only used above-mentioned Frame (or packet loss) rule that abandons, content will go through hereinafter in this respect.
Table 3 has provided an example, and it combines above-mentioned drop mechanism with the WRED technology.
Table 3. combines the discarding method sample of WRED technology
| ???C 3 | ???C 2 | ???C 1 | Low packet loss | High packet loss | |
| Level 1 N=120 | ? ???Q 3≥5 ???KB | ? ???Q 2≥17.5 ???KB | ? ???Q 1≥50 ???KB | ????0％ | ????50％ |
| Level 2 N=160 | ????25％ | ????75％ | |||
| Level 3 N=200 | ????100％ | ????100％ |
Level 3 in the table 3 has been deferred to rule proposed above, and supposition has the time-delay gauge constraints of system among Fig. 1.For example, according to above-mentioned separate equation, if at (also only when under such condition): 16Q under such condition
3+ 4Q
2+ Q
1〉=200KB, and formation Q
2Having surpassed predetermined overstocked limit one is Q
2During 〉=17.5KB, a C
2The Frame of type will be dropped.Level 1 and level 2 have formed congested of two Asias (being respectively 400 and 402) of above-mentioned discussion.For example, if 120KB≤16Q
3+ 4Q
2+ Q
1＜200KB, and Q
2〉=17.5KB then carries out packet loss with certain probability.Can notice: on each WRED level, identify Frame and belong to high packet loss kind and still belong to low packet loss kind, and specify different drop probabilities in view of the above.
As shown in Figure 2, have three Type C at this
1, C
2And C
3Specific embodiment in, each point on congested 200 has all formed the three-dimensional coordinate (Q of a queue length
1, Q
2, Q
3), this coordinate points can be born, and that is to say, if each queue length (Q
1, Q
2, Q
3) on these numerical value, keeping stable, each then corresponding time-delay gauge all is satisfiable.For example, the one group of stable state queue length that can bear is that unit representation is (50,17.5,5) with KB.These numerical value are to derive like this: Q
3=(r
3) (δ
3)=(40Mbps) (1ms)=5KB; Q
2=(r
2) (δ
2)=(40Mbps) (4ms)=17.5KB; Q
1=(r
1) (δ
1)=(25Mbps) (16ms)=50KB.
For transmitting scheduling, Δ (F) is defined as the current stand-by period of Frame F.Then, the Frame F of type i is defined as has sluggish degree (slack) ψ
i(F), ψ wherein
i(F)=δ
i-A (F).The advantage of this transmission dispatching method is very simple relation: sluggish degree (lag time) is more little, and it is high more then to transmit priority.When the lag time that calculates is when equating for the formation of two or more types, then scheduling process at first is that formation to high priority (type that just has more strict delay requirement) transmits.
Although above preferred embodiment has been made detailed description, be to be understood that: need not surmount protection range of the present invention and design philosophy that appended claims limits, just can make change, replacement and the change of various ways this.
Claims (28)
1. method of in the differentiated services network environment, carrying out the data frame scheduling and abandoning, it comprises step:
Abandon logical circuit by one a Frame in the input data frame stream abandoned, wherein abandon logical circuit according to a kind of discard algorithm work; And
Order of transmission is dispatched the order when order of transmission wherein is meant the one or more data queued frames in the network environment are sent by a scheduling logic circuit.
2. method of in the differentiated services network environment, carrying out the data frame scheduling and abandoning, it comprises step:
If
Congestion level in network environment reaches predetermined extent, and
The backlog of the formation relevant with certain Frame in the input data frame stream reaches pre-
When fixed limit is spent,
Just abandon logical circuit and abandon this Frame by one;
Order of transmission is dispatched the order when order of transmission wherein is meant the one or more Frames in the network environment are sent by a scheduling logic circuit.
3. method according to claim 2 is characterized in that: in abandoning step, described Frame is associated uniquely with a data frame type, and this described unique data frame type also is associated with described formation in the network environment.
4. method according to claim 2 is characterized in that: in abandoning step, according to maximum delay limit and minimum bandwidth limit described formation is controlled.
5. method according to claim 2, it is characterized in that: comprise a plurality of formations in the network environment wherein, and according to the corresponding queues type these formations are classified, described a plurality of queue types prolong to transmitting the highest type of priority from transmitting the minimum type of priority.
6. method according to claim 5 is characterized in that: the maximum delay value of the minimum type of described transmission priority is greater than described maximum delay value with the highest transmission priority type.
7. method according to claim 2 is characterized in that: the scheduling step comprises a step of calculating lag time, in order to determine the order of transmission of described one or more data queued frames.
8. method according to claim 7 is characterized in that: described lag time is short more, and corresponding transmission priority is high more.
9. method according to claim 7, it is characterized in that: each described one or more data queued frame all is associated with a timestamp, and the difference of long slack time of the time when the described lag time of certain selected data frame is defined as that the selected data frame is discharged in the formation described in described one or more data queued frame in described one or more data queued frame and described corresponding formation.
10. method according to claim 7, it is characterized in that: in described scheduling step, if first lag time of a lower priority formation equals second lag time of a higher priority queues, then at first described higher priority queues is carried out and transmitted scheduling.
11. method according to claim 2 is characterized in that: in described scheduling step, when a selected data frame in described one or more data queued frames arrives team position in its corresponding formation, then this Frame is carried out and transmitted scheduling.
12. method according to claim 2 is characterized in that: in abandoning step, abandon described Frame according to a plurality of predetermined congestion level.
13. method according to claim 12 is characterized in that: in reaching described a plurality of predetermined congestion level during the most very level of Congestion Level SPCC, all Frames relevant with reaching described predetermined described formation of overstocking limit all are dropped.
14. method according to claim 13, it is characterized in that: when reaching described predetermined congestion level, and when described predetermined congestion level was lower than described high congestion level, some Frames relevant with reaching described predetermined described formation of overstocking limit just were dropped.
15. a framework of carrying out the data frame scheduling and abandoning in the differentiated services network environment, it comprises:
Abandon logical circuit, be used for abandoning some selected data frame of the input data frame stream of network environment according to a kind of discard algorithm; And
The scheduling logic circuit is used for the order of transmission of the one or more data queued frames of network environment is dispatched.
16. a framework of carrying out the data frame scheduling and abandoning in the differentiated services network environment, it comprises:
Abandon logical circuit, be used for abandoning a Frame of input data frame stream according to a kind of discard algorithm, if:
Congestion level in network environment reaches predetermined extent, and
When the backlog of the formation that is associated with described Frame reaches predetermined limits, just abandon described Frame; And
The scheduling logic circuit is used for the order of transmission of the one or more data queued frames of network environment is dispatched.
17. framework according to claim 16 is characterized in that: described Frame is associated uniquely with a data frame type, and this described unique data frame type also is associated with described formation in the network environment.
18. framework according to claim 16 is characterized in that: the control to described formation is carried out according to maximum delay limit and minimum bandwidth limit.
19. framework according to claim 16, it is characterized in that: comprise a plurality of formations in the network environment, and according to the corresponding queues type these formations are classified, described a plurality of queue types prolong to transmitting the highest type of priority from transmitting the minimum type of priority.
20. framework according to claim 19 is characterized in that: the maximum delay value of the minimum type of described transmission priority is greater than described maximum delay value with the highest transmission priority type.
21. framework according to claim 16 is characterized in that: calculate lag time, in order to determine the order of transmission of described one or more data queued frames.
22. framework according to claim 21 is characterized in that: described lag time is short more, and corresponding transmission priority is high more.
23. framework according to claim 21, it is characterized in that: each described one or more data queued frame all is associated with a timestamp, and the difference of long slack time of the time when the described lag time of certain selected data frame is defined as that the selected data frame is discharged in the formation described in described one or more data queued frame in described one or more data queued frame and described corresponding formation.
24. framework according to claim 21 is characterized in that:, then at first described higher priority queues is carried out and transmitted scheduling if first lag time of a lower priority formation equals second lag time of a higher priority queues.
25. framework according to claim 16 is characterized in that: when a selected data frame in described one or more data queued frames arrives team position in its corresponding formation, then this Frame is carried out and transmitted scheduling.
26. framework according to claim 16 is characterized in that: abandon described Frame according to a plurality of predetermined congestion level.
27. framework according to claim 26 is characterized in that: in reaching described a plurality of predetermined congestion level during the most very level of Congestion Level SPCC, all Frames relevant with reaching described predetermined described formation of overstocking limit all are dropped.
28. framework according to claim 27, it is characterized in that: when reaching described predetermined congestion level, and when described predetermined congestion level was lower than described high congestion level, some Frames relevant with reaching described predetermined described formation of overstocking limit just were dropped.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US18455700P | 2000-02-24 | 2000-02-24 | |
| US60/184,557 | 2000-02-24 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1416633A true CN1416633A (en) | 2003-05-07 |
| CN100568847C CN100568847C (en) | 2009-12-09 |
Family
ID=22677393
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB018052460A Expired - Fee Related CN100568847C (en) | 2000-02-24 | 2001-02-16 | Frame scheduling in differentiated services networks and buffer management unified algorithm |
Country Status (6)
| Country | Link |
|---|---|
| EP (1) | EP1258115A1 (en) |
| KR (1) | KR20020079904A (en) |
| CN (1) | CN100568847C (en) |
| AU (1) | AU2001237043A1 (en) |
| TW (1) | TW490964B (en) |
| WO (1) | WO2001063858A1 (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103067968A (en) * | 2011-10-19 | 2013-04-24 | 鼎桥通信技术有限公司 | Method of preventing frame protocol frames from being out-of-frame |
| CN102036398B (en) * | 2009-09-29 | 2015-06-03 | 中兴通讯股份有限公司 | Relay node (RN) and method thereof for transmitting data |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030223442A1 (en) * | 2002-05-29 | 2003-12-04 | Huang Anguo T. | Buffer memory reservation |
| US7283536B2 (en) * | 2002-06-11 | 2007-10-16 | Nokia Corporation | Multimode queuing system for DiffServ routers |
| US7177274B2 (en) | 2002-06-19 | 2007-02-13 | Telefonaktiebolaget Lm Ericsson (Publ) | Methods of transmitting data packets without exceeding a maximum queue time period and related devices |
| CN100359888C (en) * | 2003-11-27 | 2008-01-02 | 华为技术有限公司 | A data poll dispatching method |
| EP1619839A1 (en) * | 2004-07-21 | 2006-01-25 | Siemens Mobile Communications S.p.A. | Method of and apparatus for scheduling transmission of multimedia streaming services over the radio channel of wireless communication systems |
| KR100745682B1 (en) * | 2005-12-08 | 2007-08-02 | 한국전자통신연구원 | I/o packet control device and method of line card in packet exchange system |
| US20100195492A1 (en) * | 2007-07-23 | 2010-08-05 | Janos Harmatos | Controlling Traffic in a Packet Switched Communications Network |
| KR101013668B1 (en) * | 2008-10-17 | 2011-02-10 | 엘에스산전 주식회사 | Coil holding apparatus of switching mechanism for air circuit breaker |
| US10904313B2 (en) * | 2017-06-20 | 2021-01-26 | Telefonaktiebolaget Lm Ericsson (Publ) | Apparatuses, methods, computer programs, and computer program products for live uplink adaptive streaming |
-
2001
- 2001-02-16 CN CNB018052460A patent/CN100568847C/en not_active Expired - Fee Related
- 2001-02-16 AU AU2001237043A patent/AU2001237043A1/en not_active Abandoned
- 2001-02-16 KR KR1020027011090A patent/KR20020079904A/en not_active Application Discontinuation
- 2001-02-16 WO PCT/US2001/005014 patent/WO2001063858A1/en not_active Application Discontinuation
- 2001-02-16 EP EP01909268A patent/EP1258115A1/en not_active Withdrawn
- 2001-02-23 TW TW090104184A patent/TW490964B/en not_active IP Right Cessation
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102036398B (en) * | 2009-09-29 | 2015-06-03 | 中兴通讯股份有限公司 | Relay node (RN) and method thereof for transmitting data |
| CN103067968A (en) * | 2011-10-19 | 2013-04-24 | 鼎桥通信技术有限公司 | Method of preventing frame protocol frames from being out-of-frame |
Also Published As
| Publication number | Publication date |
|---|---|
| EP1258115A1 (en) | 2002-11-20 |
| KR20020079904A (en) | 2002-10-19 |
| CN100568847C (en) | 2009-12-09 |
| WO2001063858A1 (en) | 2001-08-30 |
| TW490964B (en) | 2002-06-11 |
| AU2001237043A1 (en) | 2001-09-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hong et al. | Congestion control and prevention in ATM networks | |
| JP2957062B2 (en) | Switching system using priority and method for limiting jitter in this system | |
| US20050163048A1 (en) | Method and system for providing committed information rate (CIR) based fair access policy | |
| US20010051992A1 (en) | Unified algorithm for frame scheduling and buffer management in differentiated services networks | |
| EP0920235A2 (en) | Congestion management in a multi-port shared memory switch | |
| US20060045009A1 (en) | Device and method for managing oversubsription in a network | |
| JPS63176045A (en) | Method and apparatus for width control type packet exchange | |
| JPS63176046A (en) | Method and apparatus for width control type packet exchange | |
| CN100568847C (en) | Frame scheduling in differentiated services networks and buffer management unified algorithm | |
| CN101834790A (en) | Multicore processor based flow control method and multicore processor | |
| CN109873773B (en) | Congestion control method for data center | |
| CN103023806A (en) | Control method and control device of cache resource of shared cache type Ethernet switch | |
| CN101567851A (en) | Method and equipment for shaping transmission speed of data traffic flow | |
| CN102461093A (en) | Method of managing a traffic load | |
| Yin et al. | Implication of dropping packets from the front of a queue | |
| CN108667746B (en) | Method for realizing service priority in deep space delay tolerant network | |
| CN116980342B (en) | Method and system for transmitting data in multi-link aggregation mode | |
| Bonald et al. | Scheduling network traffic | |
| CN111988815A (en) | Dynamic multi-path routing algorithm based on PDMR performance | |
| Movahhedinia et al. | A slot assignment protocol for indoor wireless ATM networks using the channel characteristics and the traffic parameters | |
| US7088738B1 (en) | Dynamic fragmentation of information | |
| CN114398174B (en) | Dynamic energy-saving scheduling method and system for data center network in SDN environment | |
| JPH07131465A (en) | Atm cell multiplex control system | |
| JPH11252097A (en) | Device and method for sending packet | |
| Usha et al. | Pushout policy in active queue management to support quality of service guaranties in IP routers |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| ASS | Succession or assignment of patent right |
Owner name: CONEXANT SYSTEMS, INC. Free format text: FORMER OWNER: GALINK SEMICONDUCTOR V.N. INC Effective date: 20070629 |
|
| C41 | Transfer of patent application or patent right or utility model | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20070629 Address after: American California Applicant after: Conexant Systems Inc. Address before: American California Applicant before: Mitel Semiconductor V.N. Inc. |
|
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20091209 Termination date: 20140216 |