CN1194504C - Network status updating method based on time-delay probability-distribution function - Google Patents
Network status updating method based on time-delay probability-distribution function Download PDFInfo
- Publication number
- CN1194504C CN1194504C CN 03146792 CN03146792A CN1194504C CN 1194504 C CN1194504 C CN 1194504C CN 03146792 CN03146792 CN 03146792 CN 03146792 A CN03146792 A CN 03146792A CN 1194504 C CN1194504 C CN 1194504C
- Authority
- CN
- China
- Prior art keywords
- delay
- distribution function
- probability distribution
- state
- delay probability
- 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.)
- Expired - Fee Related
Links
- 238000005315 distribution function Methods 0.000 title claims abstract description 51
- 238000000034 method Methods 0.000 title claims abstract description 36
- 230000011664 signaling Effects 0.000 claims abstract description 9
- 230000001960 triggered effect Effects 0.000 claims abstract description 6
- 230000008569 process Effects 0.000 claims description 11
- 238000013139 quantization Methods 0.000 claims description 8
- 238000004806 packaging method and process Methods 0.000 claims description 3
- 238000004891 communication Methods 0.000 abstract description 2
- 230000008859 change Effects 0.000 description 13
- 238000005259 measurement Methods 0.000 description 7
- 230000007246 mechanism Effects 0.000 description 7
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000003111 delayed effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
Images
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The present invention relates to a network state updating method based on a time-delay probability distribution function, which belongs to the technical field of communication. In the method, each network node first carries out statistics on the time-delay probability distribution function of the node for data packets in various service type queues in each T period; the time-delay probability distribution function of the queues in the statistical period is compared with a time-delay probability distribution function published during the last state update, and the state update is triggered when a certain value is exceeded; the interval between two continuous state updates ranges from KT to NT, wherein the K and the T are positive integers, and the N is larger than the K; the latest time-delay probability distribution function on which statistics are carried out by each queue is encapsulated in state updating signaling packets so as to be broadcast towards a whole network; a state information database of local nodes is updated by each network node according to the state information in the received state updating signaling packets. The method of the present invention comprehensively and accurately reflects changes in the network state, and enhances the accuracy of the state information and the stability of routes so that a basis is provided for the QoS guarantee based on the end-to-end time delay.
Description
The invention relates to a network state updating method based on a delay probability distribution function, in particular to a network state updating method based on the change of the probability distribution function of the delay of a data packet at a node, belonging to the technical field of communication.
Background art currently the main traffic of the Internet is data traffic. Data traffic does not have strict requirements on the quality of service of the network. With the advent of voice and real-time multimedia services, user services have increasingly high quality of service (hereinafter QoS) requirements for networks. The QoS requirements mainly include available bandwidth, delay, packet loss rate, etc. In order to meet the QoS requirements of the user services, the routing process needs to continuously know the current operating state of the network, and on this basis, in combination with a certain optimization objective, an optimal or at least available transmission path meeting the user requirements is selected.
The QoS route selection process consists of two parts: the first is a process of selecting a path for arriving a service and sending a data packet, which is called a path finding process; the first is the interaction process of state information between nodes. Referred to as a state update procedure. In the QoS network, the measurement parameters of the state of each node are collected, processed and transmitted through a state updating mechanism. The good state updating mechanism can provide timely and reliable network state information for the QoS route, and is an important precondition for the success of the QoS route. Meanwhile, inaccurate state information not only causes routing failure, but also greatly increases network overhead, and causes low overall performance of the QoS network.
The design of the network state updating method mainly comprises two aspects of content, namely, selecting measurement parameters and designing a trigger mechanism. The currently commonly used state metric parameter is the available bandwidth of the link, because in the traditional Internet network, single queue queuing is adopted, and no classified service is available, and in this way, other metric parameters, such as delay and packet loss rate, can be converted into a function of the available bandwidth. The trigger mechanism is based on time change, such as periodic trigger, or based on measurement parameter change, such as judgment value trigger, that is, several judgment values are preset according to the change range of the measurement parameter, and when the change of the measurement parameter reaches a certain judgment value, the state update is triggered. Existing arbitration value triggers are based on changes in available bandwidth.
The selection of available bandwidth as a metric parameter for network state changes has the following problems:
1. the available bandwidth of the network link is constantly changing, and the value of the available bandwidth issued as the state information is actually an average value counted in a certain period of time. It cannot reflect the change of network state comprehensively and accurately.
2. In a QoS network, different types of services have different quality of service requirements and need to be used. The available bandwidth is a value after aggregation of multiple services, and cannot reflect the service condition of a certain service specifically.
3. After different queuing and scheduling mechanisms are adopted for different types of services, other measurement parameters such as delay and packet loss rate are no longer functions of available bandwidth.
4. In the QoS network, with the emergence of the multipath technology, the concept of bandwidth is diluted, and the guarantee of different service qualities is finally summarized as the guarantee of delay, rather than the guarantee of bandwidth.
5. In the QoS network, because the service has strong burstiness, packet loss is inevitable, the guarantee of service quality is guaranteed in a probability sense, and the guarantee of bandwidth has no direct meaning any more.
The invention aims to provide a network state updating method based on a delay probability distribution function, which takes the delay probability distribution functions of different types of services at nodes as measurement parameters of network state change, designs a new state updating method applicable to a QoS network, adopts a trigger mechanism combining time trigger and parameter trigger, combines a proper routing mechanism, replaces bandwidth guarantee with delay guarantee, fully exerts the characteristics and advantages of packet switching, and provides different types of service quality guarantee for the QoS network.
The invention provides a network state updating method based on a delay probability distribution function, which comprises the following steps:
1. the network node takes T as a period and counts the delay probability distribution function of the data packets of various service type queues at the node;
2. when the delay probability distribution function of the queue counted in the period is compared with the delay probability distribution function issued by the last state update and exceeds a certain value, the state update is triggered, the interval between two continuous state updates is KT-NT, wherein K, N is a positive integer, and N is more than K;
3. packaging the latest delay probability distribution function counted by each queue in a state updating signaling packet and broadcasting the latest delay probability distribution function to the whole network;
4. and each network node updates the state information database of the local node according to the state information in the newly received state updating signaling packet.
The delay probability statistical process in the method may include the following steps:
(1) the node records the time of each data packet entering and leaving the service type queue, calculates the delay of each data packet, and sets the delay of the discarded data packet to be infinite;
(2) quantizing the calculated delay into discrete values, counting the number of data packets with the same delay quantization value by taking T1 as a period, and dividing the number by the total number of the data packets entering the queue in the T1 period to obtain the probability distribution of each delay quantization value.
The judgment value of the trigger state updating in the method is the direction divergence between the queue delay probability distribution function counted in the period and the delay probability distribution function released by the last state updating.
The network state updating method based on the delay probability distribution function has the following advantages that:
1. the delay probability distribution function of the data packets of different service types at the node is adopted as the state information, and the change of the network state is comprehensively and accurately reflected.
2. The change of the available bandwidth of the physical port of the node is refined into the change of delay probability distribution function of each service type queue using the same port, and a foundation is provided for QoS guarantee based on end-to-end delay.
3. And the probability distribution function of the sub-queue statistical delay provides statistical reference information for the QoS scheduling and routing of the sub-type and the priority.
4. The updating is triggered by the change of the delayed probability distribution function, and the stability is better.
5. The method combines time triggering and parameter triggering and adopts a triggering updating method, so that the accuracy of state information and the stability of routing are improved.
Drawings
FIG. 1 is a block flow diagram of the method of the present invention.
Detailed Description
The flow chart of the network state updating method based on the delay probability distribution function is shown in figure 1, firstly, a network node counts the delay probability distribution function of data packets of various service type queues at the node by taking T1 as a period; when the delay probability distribution function of the queue counted in the period is compared with the delay probability distribution function issued by the last state update and exceeds a certain value, the state update is triggered, the interval between two continuous state updates is KT-NT, wherein K, N is a positive integer, and N is more than K; packaging the latest delay probability distribution function counted by each queue in a state updating signaling packet and broadcasting the latest delay probability distribution function to the whole network; and each network node updates the state information database of the local node according to the state information in the newly received state updating signaling packet.
The probability statistical process in the method is as follows: firstly, recording the time of each data packet entering and leaving a service type queue by a node, calculating the time delay of each data packet, and setting the time delay of the discarded data packet to be infinite; and quantizing the calculated delay into discrete values, counting the number of data packets with the same delay quantization value by taking T1 as a period, and dividing the number by the total number of the data packets entering the queue to obtain the probability distribution of each delay quantization value.
The judgment value of the trigger state updating in the method is the direction divergence between the queue delay probability distribution function counted in the period and the delay probability distribution function released by the last state updating.
The state updating process of the present invention has three clock control processes, as shown in fig. 1:
statistical clock T1: and the nodes count the probability distribution of the delay of each service type queue by taking T1 as a period.
Issue clock T2: when the probability distribution of the delay of each service type queue is constant all the time, the node issues the delay probability distribution function of each service type queue counted in the last T1 period at intervals of T2. T2 is an integer N times T1. T2 determines the maximum time interval between two status updates. It ensures that the node is in an active state, and if the state information of a certain node is not received at the interval T2, the node is considered to be in a failure state.
Clamp clock T3: when the network state changes frequently, in order to prevent the routing instability and excessive network overhead caused by too fast change of the state updating information, the interval of two consecutive state updates is not less than T3. T3 is an integer K times T1, K < N.
The judging process of the trigger state updating in the method of the invention is as follows: and continuously comparing the delay probability distribution function of each service type queue counted in the latest T1 period with the state information released by the last state update, and triggering the update as long as the delay probability distribution function of one service type queue changes and exceeds a preset judgment value.
Triggering updating: and (4) taking the probability distribution function of the delay of each service type queue counted by the last T1 period as the content of the state update.
And (3) state release: and encapsulating the probability distribution function of the delay of each service type queue counted in the last T1 period into a state updating signaling packet to be broadcast to the whole network.
The comparison of the variation of the probability distribution function according to the present invention is a problem that is worthy of intensive study. The method provides a solution based on the information theory, namely, the change of two probability distribution functions is measured by the directional divergence and is used as a judgment value for triggering state updating.
The probability distribution function of the delay of a certain type of data packet counted in the current T1 period is:
wherein (a)1，a2…ak) Is a quantized value of the packet delay. p is a radical of2(a1)，p2(a2)…p2(ak) Is the probability of the delay quantization value of the data packet counted in the period.
The probability distribution function of the delay of the type of data packet issued last time is:
wherein (a)1，a2…ak) Is a quantized value of the packet delay. p is a radical of1(a1)，p1(a2)…p1(ak) Is the probability of the delay quantization value of the data packet issued by the last state update.
The directional divergence calculation formula of the probability distribution function of the delay of the data packets of the type counted twice is as follows:
J(p2，p1；X)＝I(p2，p1；X)+I(p1，p2；X)
wherein,
divergence in direction J (p)2，p1(ii) a X) is a measure of the difference between the two probability distribution functions that can be used as a basis for determining the trigger state update.
Claims (2)
1. A network state updating method based on a delay probability distribution function is characterized by comprising the following steps:
(1) the network node takes T as a period and counts the delay probability distribution function of the data packets of various service type queues at the node;
(2) when the delay probability distribution function of the queue counted in the period is compared with the delay probability distribution function issued by the last state update and exceeds a certain value, the state update is triggered, the interval between two continuous state updates is KT-NT, wherein K, N is a positive integer, and N is more than K;
(3) packaging the latest delay probability distribution function counted by each queue in a state updating signaling packet and broadcasting the latest delay probability distribution function to the whole network;
(4) each network node updates the state information database of the local node according to the state information in the latest received state updating signaling packet;
the delay probability statistical process comprises the following steps:
(1) the node records the time of each data packet entering and leaving the service type queue, calculates the delay of each data packet, and sets the delay of the discarded data packet to be infinite;
(2) and quantizing the calculated delay into discrete values, counting the number of data packets with the same delay quantization value by taking T1 as a period, and dividing the number by the total number of the data packets entering the queue to obtain the probability of each delay quantization value.
2. The method of claim 1, wherein the decision value for triggering state update is a directional divergence between the queue delay probability distribution function counted in the current period and the delay probability distribution function issued by the last state update.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 03146792 CN1194504C (en) | 2003-07-11 | 2003-07-11 | Network status updating method based on time-delay probability-distribution function |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 03146792 CN1194504C (en) | 2003-07-11 | 2003-07-11 | Network status updating method based on time-delay probability-distribution function |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1472923A CN1472923A (en) | 2004-02-04 |
| CN1194504C true CN1194504C (en) | 2005-03-23 |
Family
ID=34156121
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 03146792 Expired - Fee Related CN1194504C (en) | 2003-07-11 | 2003-07-11 | Network status updating method based on time-delay probability-distribution function |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1194504C (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9438436B2 (en) * | 2006-08-14 | 2016-09-06 | Alcatel Lucent | Broadcast anchor availability indication |
| US8982725B2 (en) | 2012-02-03 | 2015-03-17 | Mediatek Inc. | Method and apparatus for collecting and providing diverse traffic information in cellular networks |
-
2003
- 2003-07-11 CN CN 03146792 patent/CN1194504C/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN1472923A (en) | 2004-02-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7551556B2 (en) | Monitoring traffic in packet networks using the sliding window procedure with subwindows | |
| CN107171842B (en) | Multipath transmission protocol congestion control method based on reinforcement learning | |
| US20050030952A1 (en) | Call admission control/session management based on N source to destination severity levels for IP networks | |
| US7430169B2 (en) | Retro flow control for arriving traffic in computer networks | |
| US20200169479A1 (en) | Efficient metric calculation with recursive data processing | |
| CN101969408B (en) | Active queue management method based on packet DSCP (Differentiated Services Code Point) marks | |
| CN111683393A (en) | Adaptive congestion control method for dynamically adjusting gain coefficient | |
| CN1933431A (en) | Method for detecting QoS | |
| GB2478277A (en) | Controlling packet transmission using variable threshold value in a buffer | |
| CN101056260A (en) | ECN mechanism-based congestion control method in the mixed network | |
| CN1297098C (en) | System flow controlling method | |
| CN1188984C (en) | Selecting method based on path-time delay probability distribution | |
| CN113064738B (en) | Active queue management method based on summary data | |
| CN112995048A (en) | Blocking control and scheduling fusion method for data center network and terminal equipment | |
| CN1200538C (en) | Method for selecting route by accounting and controlling node status | |
| CN1194504C (en) | Network status updating method based on time-delay probability-distribution function | |
| CN110324255B (en) | Data center network coding oriented switch/router cache queue management method | |
| US20040179473A1 (en) | Random early packet discard (RED) | |
| CN114844837B (en) | Congestion control method and device based on time delay in multi-service queue environment | |
| CN116170380A (en) | ECN marking strategy and queue management method and system based on congestion prediction | |
| CN107770082B (en) | Transmission control method based on task flow characteristics in data center network | |
| Irawan et al. | Performance evaluation of queue algorithms for video-on-demand application | |
| Kozlovskiy et al. | Development of a modified method of network traffic forming | |
| Wang et al. | TCP-friendly congestion control schemes in the Internet | |
| Shihada et al. | Threshold-based TCP Vegas over optical burst switched networks |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| 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: 20050323 |