CN100446479C - Method for measuring end-to-end available bandwidth based on network compact link position - Google Patents

Method for measuring end-to-end available bandwidth based on network compact link position Download PDF

Info

Publication number
CN100446479C
CN100446479C CNB2005100243112A CN200510024311A CN100446479C CN 100446479 C CN100446479 C CN 100446479C CN B2005100243112 A CNB2005100243112 A CN B2005100243112A CN 200510024311 A CN200510024311 A CN 200510024311A CN 100446479 C CN100446479 C CN 100446479C
Authority
CN
China
Prior art keywords
packet
link
measured
bag
tight link
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
Application number
CNB2005100243112A
Other languages
Chinese (zh)
Other versions
CN1658583A (en
Inventor
张大陆
吴烨
许健
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tongji University
Original Assignee
Tongji University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tongji University filed Critical Tongji University
Priority to CNB2005100243112A priority Critical patent/CN100446479C/en
Publication of CN1658583A publication Critical patent/CN1658583A/en
Application granted granted Critical
Publication of CN100446479C publication Critical patent/CN100446479C/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

The present invention relates to an end-to-end available bandwidth measuring method based on network tight link positioning. The method comprises the following steps: the hop number of a path to be measured is measured, and a tight link is positioned; a probe packet column is constructed according to the information of the tight link, the probe packet column is composed of a test packet column and a loading packet column, wherein the test packet column also comprises a forward packet column and a returning packet column, and a loading packet is greater than a test packet; the value of a life cycle domain in the IP packet head of each data packet in the loading packet column is set to be the hop number of the tight link to be measured; the value of a life cycle domain in the IP packet head of each data packet in the returning packet column is set to be a previous hop of the position of the tight link to measured; a sending terminal sends probe packet columns of different rates to a receiving terminal repeatedly for multiple times; the output rates and the input rates of multiple groups of links to be measured are calculated according to return information; consequently, the available bandwidth of the tight link is calculated. The method has the advantages that the tight link can be positioned at high speed, the network intrusion property is low, and the information of the available bandwidth of the link can be accurately given.

Description

The method of the end-to-end available bandwidth measurement of a kind of tight link Network Based location
Technical field
The present invention relates to a kind of method of measurement of available bandwidth.The technical field of performance test that belongs to computer network.
Background technology
Article one, network path is made up of some network links.For end to end initiatively for the method for testing, the available bandwidth of network link is meant that one group of probe data packet row can not have congestedly the maximum rate by link, and the available bandwidth of network path is meant that one group of probe data packet row can not have congested maximum rate by this path.Tight link in the network path is meant that the network link available bandwidth equals the link of network path available bandwidth.
Along with the fast development of computer networking technology, very big innovation has also taken place in corresponding applied in network performance test technology, and the localization method of the tight link of test path mainly contains following two kinds at present.
1. based on the Passive Location of SNMP
This method is to utilize using simple network management protocol (simple network management protocol, SNMP) read relevant information among the all-router MIB in the path to be measured, thereby obtain along the distribution situation of the link available bandwidth size in path to be measured.By quantitative judgement, application program can obtain the positional information of tight link easily.The advantage of these class methods is that method is simple, practical, workable, and is convenient to management.But the shortcoming of the method also is conspicuous.At first, the MIB of router might not open read right.And common situation is, the different route-maps that ISP had are privately owned, therefore, has intrinsic limitation based on the suitable environment of this localization method piecemeal.This limitation then causes this method not possess practical feasibility.Secondly, this method is means of a kind of passive monitoring by the means that read the distribution of router mib information architecture available bandwidth.The accuracy subject to confirmation of information that router provides, and this method does not provide relevant therewith affirmation scheme.Therefore, this method also has the accurate positioning problem of tight link.Use the software of these class methods that MRTG etc. is arranged.
2. positive location method end to end
This method uses end to end measuring technique initiatively to carry out the location of the tight link in path to be measured.End in path to be measured initiatively sends probe stream to the other end.The character that is shown in link by probe stream is determined the tight link information in the path to be measured.The advantage of the method is that method is simple, practical, have nothing to do, be convenient to management with intermediate equipment (as the router in the path).It is the instrument that first item uses end-to-end positive location method Measurement Network path available bandwidth bottleneck that Carnegie Mellon Univ USA researches and develops the Bfind instrument.At the test transmitting terminal, Bfind constantly sends load to receiving terminal, obtains the belt time-delay of each link in the network path by the method for traceroute simultaneously, and judges the position of network availability bandwidth bottleneck with this.But early stage end-to-end positive location method is bigger to the negative effect that network caused.Because it is bigger to send load, can influence other application in the network path, and lower by the belt time-delay of the link that method the obtained accuracy of traceroute.
The availability bandwidth measurement method of link to be measured mainly contains following two kinds.
1, in the passive measurement method of SNMP
The relevant information among the MIB that SNMP reads the all-router in the path to be measured is used in this method utilization, thereby obtains distributing along the link available bandwidth in path to be measured.By quantitative judgement, application program can obtain available bandwidth easily.There are identical pluses and minuses in this method and the above-mentioned Passive Location based on SNMP.
2, to the active method of measurement of holding
This method be with end to end initiatively measuring technique carry out the measurement of available bandwidth.Also there are identical pluses and minuses in the method with the above-mentioned method of positive location end to end.It is bigger that early stage (2003) BFind orientation tool generates load, and network intrusions is big; The DRPS orientation tool need carry out the measurement of available bandwidth earlier, and the phenomenon of convergence is slowly arranged, and promptly the testing time is longer.The Pathload instrument that uses the software of these class methods to also have Boston Univ USA to research and develop, it can only give the scope of outgoing link available bandwidth.And the IGI instrument that Carnegie Mellon Univ USA researched and developed, it has supposed that tight link position is identical with the link bandwidth bottleneck, this hypothesis usually is wrong in the actual network environment.
Summary of the invention
At the deficiency of network path available bandwidth bottleneck localization method and available bandwidth measurement, the present invention proposes the method for the end-to-end available bandwidth measurement of a kind of tight link Network Based location.It helps available bandwidth and place, bottleneck position in application program or the accurate awareness network of the network management personnel path, and helps to develop and implement the congestion control policy and the scheduling strategy of various novel bandwidth Network Based.
Probe bag described in the invention is listed as by test pack row to be formed with load bag row institute, and the test pack row have comprised the probe bag row of two kinds of character again: Bao Lie and return bag and be listed as advances.
Load bag row can be made of N equally spaced packet greater than 500 bytes.Advance bag or return bag of one of all continuous nonseptate transmission behind each load bag.The size of test pack is preferably less than 100 bytes.With life cycle territory (the Time To Live in each packet IP packet header of load Bao Liezhong, TTL) value is made as position i to be measured and jumps, because the router on the network path can subtract 1 with the life cycle thresholding after receiving packet, and when the life cycle thresholding is 0, packet will be routed device and abandon.Therefore, these load bags row that are provided with the life cycle thresholding will be routed device and abandon when marching to the i bar link of network path to be measured.
Returning bag row has from the packet that returns character by N/2 and is formed.For example, in the realization of TCP/IP network, having this packet from return attribute is the traceroute bag.The previous dive position that life cycle thresholding in each packet IP packet header of Bao Liezhong is provided with link to be measured position will be returned, promptly i-1 jumps, return so and time of advent between continuous two return data bags of Bao Liezhong will successfully write down the speed input inter packet gap that load bag row enter i bar network link at interval, according to the formula (1) of describing probe bag row speed, we have just obtained the input rate that the probe bag is listed in i bar link;
R = s δ (formula 1)
Wherein R represents the speed of probe bag row, and s represents the length of probe bag row, and δ represents the inter packet gap sum of probe bag row.
The bag row that advance are made up of N/2 packet, wherein, life cycle thresholding in each packet IP packet header uses as default (different operating system has different default values), thereby, the bag row that advance obtain measured receiving terminal, advance so and continuous two times of advent of advancing between the packet of Bao Liezhong will successfully write down at interval the speed output packet interbody spacer that load bag row enter i bar network link, same, according to formula (1), just obtain the probe bag and be listed in the output speed that i jumps.
If find input rate greater than output speed threshold value σ, then judge i jump out showed congested, otherwise show i jump do not take place congested.Corresponding to probe bag row initial rate R, the input rate and the output speed that the probe bag are listed as every link in the network path to be measured that is passed through are judged, the congestion link number that record is obtained.
The present invention need be provided with the transmitting terminal and the receiving terminal of probe data packet row at the two ends of network path to be measured respectively.The present invention is made up of following three testing procedures:
Step 1 is measured the jumping figure (or number of links) in path to be measured, determines the initial rate of probe bag row.The method of measuring path to be measured jumping figure has multiple, for example, sends Internet Control Message Protocol (Internet Control MessagePacket, ICMP) packet with traceroute; Or send one group of H UDP (User Datagram Protocol, UDP) packet, and the life cycle thresholding in its IP packet header is set to P, and the life cycle thresholding in the IP of the UDP message bag that receiving terminal obtained bag row is Q so, and jumping figure is P-Q.The initial rate of probe bag row also has several different methods to obtain, and it only needs to guarantee that the available bandwidth greater than tight link in the path gets final product.
Step 2 is located tight link.Send probe bag row with the initial rate that step 1 was obtained along network path to be measured, as shown in Figure 2,, judge that then this unique congestion point is tight link position if the congestion link number that takes place equals 1.If the congestion link number that takes place is greater than 1, then each jumps transmission probe stream along the path once more, the speed that sends for last judged near the input rate of the congestion link of receiving terminal, judge again whether the congestion link number equals 1, if 1, this unique congestion point is tight link position, otherwise this process of repetition, up to the location of finishing tight link.For example, Fig. 3 has explained once tight chain position fixing process.Transmitting terminal has generated two groups of probe bag row altogether in this process, for the first time be to be that initial rate sends probe bag row with 10Mbps, the result obtains 5 congestion links, the speed 3.047Mbps that for the second time probe bag row speed is adjusted near the congestion link of receiving terminal surveys, the result obtains 1 congestion link, and this link is tight link.
Step 3 is measured available bandwidth.Finish the location of tight link, repeat to send the probe bag row of M different rates, obtain M input rate and output speed for tight link.Because input rate and output speed satisfy formula (2), according to formula (3), obtain the available bandwidth A of this tight link.
R O = R I R I < A C T R I R I + &lambda; R I &GreaterEqual; A (formula 2)
A=C T-λ (formula 3)
R wherein IWith R OInput and the output speed of representing tight link position respectively, C TRepresent tight link capacity, λ is for being present in the Mean Speed of the offered load in the tight link, and A is the available bandwidth of tight link.
The method of the end-to-end available bandwidth measurement of tight link Network Based provided by the invention location is convenient to the application of web application end to end.This method need not to know the private information of router, for web application is more friendly end to end.In addition, at present, increasing network application is to (the using or the like as Streaming Media) of network performance sensitivity.And the ever-increasing relatively application demand of the intrinsic bandwidth of network path is insufficient, the place, position of the bottleneck (being tight link) by awareness network path actual available bandwidth and available bandwidth, can effectively optimize the performance of various network applications end to end, thereby optimize the whole network performance.Tight link locating speed of the present invention is fast, network intrusions is little, can provide accurate link available bandwidth information.
Description of drawings
Further specify the present invention below in conjunction with drawings and Examples.
Fig. 1 is a probe Bao Lieyi example structure schematic diagram.
Fig. 2 is an embodiment emulation experiment environment schematic diagram.
When Fig. 3 met tight link i for probe bag row, probe bag column information changed schematic diagram.
Fig. 4 is emulation experiment is located tight link under the load of CBR background a process.
Fig. 5 is emulation experiment is measured available bandwidth with linear regression method under the load of CBR background a process.
Embodiment
The method of the end-to-end available bandwidth measurement of a kind of tight link Network Based location, the linkage length of the simulated environment of present embodiment is 10 (H=10).
As shown in the figure, Fig. 1 is a probe bag row example structure schematic diagram, and wherein white square is the load bag, and the black bag is for returning bag, and the square that has horizontal line is the bag that advances.Fig. 2 is an embodiment emulation experiment environment schematic diagram, and wherein Snd is a transmitting terminal, and Rcv is a receiving terminal, R iAnd S iBe the load receiving terminal and the load transmitting terminal of i group background stream.
When Fig. 3 met tight link i for probe bag row, probe bag column information changed schematic diagram.Fig. 4 is that (Constant Bit Rate, CBR) process of tight link is located in the background load down to emulation experiment at constant bit rate.Wherein sent probe stream altogether twice, the speed of probe stream 1 is 10Mbps, and the speed of probe stream 2 is 3.073Mbps, and as we know from the figure, tight link is a link 3.Fig. 5 is emulation experiment is measured available bandwidth with linear regression method under the load of CBR background a process.Inverse for ingress rate uses linear regression can obtain available bandwidth with the inverse that leaves speed.
After finishing the location of tight link, repeat to send probe bag row No. 15 times for tight link, 2% of the transmission rate of preceding 7 groups flow velocity during by the tight link in last location increases progressively, 2% of the transmission rate of 7 groups the flow velocity in back during by the tight link in last location successively decreased, and the transmission rate of last group be the transmission rate when locating tight link for the last time.Obtain the input rate and the output speed of 15 pairs of tight link positions like this.The utilization linear regression can obtain the capacity of tight link and the Mean Speed that is present in the offered load in the tight link as shown in Figure 5.
Emulation is divided into four groups in the present embodiment: the background stream in first group be constant bit rate (Constant Bit Rate, CBR), tight link is a link 3; Background stream in second group is CBR, and tight link is link 3 and link 7; The 3rd group background stream is Plato (Pareto) and distributes, and tight link is a link 3; The 4th group background stream is index stream, and tight link is a link 3.These four groups experiments respectively are made up of 7 experiments, and the background stream size of the background stream size of tight link and non-tight link is followed successively by: 3,5,3,7,1,3,1,2,3,9,5,9,7,9, and the Mbps of unit.
The checking that different background flows down
In first group of experiment, the tight link location of 7 experiments is all accurate, and the measurement relative error of available bandwidth is followed successively by: 1.5 ,-1.0,0.9 ,-1.3,1.7,1.8,1.6.The present invention is first group of its correctness of experimental verification.Fig. 4 and Fig. 5 are the process prescription of the 7th experiment, and Fig. 4 has described the tight link position fixing process of the 7th experiment, and Fig. 5 has described the measuring process of the available bandwidth of the 7th experiment.
In second group of experiment, the tight link location of 7 experiments is all accurate, and the measurement relative error of available bandwidth is followed successively by: 2.0,2.5 ,-1.7 ,-1.9,1.9,2.1,2.5.The present invention is second group of its correctness of experimental verification.
In the 3rd group of experiment, the tight link location of 7 experiments is all accurate, and the measurement relative error of available bandwidth is followed successively by: 7.8,8.1 ,-7.9 ,-3.4,10.9,8.1,9.3.The present invention is the 3rd group of its correctness of experimental verification.
In the 4th group of experiment, the tight link location of 7 experiments is all accurate, and the measurement relative error of available bandwidth is followed successively by: 8.3 ,-8.9,9.4 ,-10.2,7.7,9.2,10.1.The present invention is the 4th group of its correctness of experimental verification.
This shows that the present invention can correctly locate tight link, measures available bandwidth exactly.

Claims (2)

1, the method for the end-to-end available bandwidth measurement of a kind of tight link Network Based location may further comprise the steps:
(1) measures the jumping figure in path to be measured and locate tight link;
(2) the tight link information structure probe bag row that obtain according to (1), described probe bag is listed as by test pack row and load bag row to be formed, the test pack row have comprised that again the Bao Lie that advances wraps row with returning: wherein, life cycle thresholding in each packet IP packet header that the bag that advances is listed as uses as default, the load bag is greater than test pack, and the value in the life cycle territory in each packet IP packet header of load Bao Liezhong is made as tight link to be measured place jumping figure; The value of returning the life cycle territory in each packet IP packet header of Bao Liezhong is made as the place jumping figure of the previous dive position of tight link to be measured position;
(3) repeat repeatedly to send the probe bag row of different rates to receiving terminal by transmitting terminal, the input inter packet gap that enters link to be measured the time of advent at interval as probe bag row of continuous two the return data parlors of Bao Liezhong will be returned, and continuous two packets that advance of the Bao Liezhong that will advance time interval of arriving receiving terminals leave the output packet interbody spacer of link to be measured as the probe bag, calculate the input and output rate of many groups link to be measured, thereby calculate the available bandwidth of this tight link.
2, the method for the end-to-end available bandwidth measurement of a kind of tight link Network Based according to claim 1 location is characterized in that: described structure probe Bao Liezhong, advance bag or return bag of one of all continuous nonseptate transmission behind each load bag.
CNB2005100243112A 2005-03-10 2005-03-10 Method for measuring end-to-end available bandwidth based on network compact link position Expired - Fee Related CN100446479C (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CNB2005100243112A CN100446479C (en) 2005-03-10 2005-03-10 Method for measuring end-to-end available bandwidth based on network compact link position

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CNB2005100243112A CN100446479C (en) 2005-03-10 2005-03-10 Method for measuring end-to-end available bandwidth based on network compact link position

Publications (2)

Publication Number Publication Date
CN1658583A CN1658583A (en) 2005-08-24
CN100446479C true CN100446479C (en) 2008-12-24

Family

ID=35007847

Family Applications (1)

Application Number Title Priority Date Filing Date
CNB2005100243112A Expired - Fee Related CN100446479C (en) 2005-03-10 2005-03-10 Method for measuring end-to-end available bandwidth based on network compact link position

Country Status (1)

Country Link
CN (1) CN100446479C (en)

Families Citing this family (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN100463422C (en) * 2006-09-08 2009-02-18 中山大学 A link, path, and network availability bandwidth measurement method
CN100559762C (en) * 2007-07-05 2009-11-11 中国科学技术大学 A kind of effective bandwidth measuring method based on PGM
CN100596081C (en) * 2007-10-15 2010-03-24 中国电信股份有限公司 Band width testing method and system for asynchronous transfer mode LAN simulation business
CN101299687B (en) * 2008-05-07 2010-10-27 中国联合网络通信集团有限公司 Method and system for measuring available bandwidth of wireless LAN
CN101340318B (en) * 2008-08-15 2010-12-22 同济大学 Available bandwidth measuring method of rate lowered bougie package
CN101729306B (en) * 2008-10-28 2012-05-23 中兴通讯股份有限公司 Method and system for testing downlink broadband of access line of broadband network user
CN101958811B (en) * 2009-07-14 2012-07-18 中兴通讯股份有限公司 Method and system for measuring end-to-end available bandwidth
CN101753457B (en) * 2009-12-24 2011-12-07 中国人民解放军理工大学指挥自动化学院 Method and system for predicating available bandwidth of end-to-end route between nodes at Internet end
CN101895417B (en) * 2010-07-06 2013-02-13 同济大学 Positioning method for tight link based on available bandwidth of subpaths
CN102055628B (en) * 2011-01-11 2012-07-04 同济大学 Method of available bandwidth measurement and close link circuit positioning based on single pack column
CN104113446A (en) * 2014-07-03 2014-10-22 南京航空航天大学 Method for measuring single-end available bandwidth in unexpected background flow environment
CN113852496B (en) * 2021-09-13 2023-07-25 电子科技大学 High-precision network bandwidth measurement system capable of diagnosing tight link position
CN114884846A (en) * 2022-04-26 2022-08-09 清华大学 End-to-end available bandwidth monitoring method and system

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6614763B1 (en) * 1999-02-04 2003-09-02 Fujitsu Limited Method of and apparatus for measuring network communication performances, as well as computer readable record medium having network communication performance measuring program stored therein
CN1488214A (en) * 2001-04-19 2004-04-07 �ʼҷ����ֵ������޹�˾ Method and device for robust real-time estimation of the bottleneck bandwidth in the internet
CN1543157A (en) * 2003-03-03 2004-11-03 Controlling admission of data streams onto a network based on end-to-end measurements
CN1545277A (en) * 2003-11-19 2004-11-10 北京邮电大学 An end-to-end network bottleneck bandwidth measuring method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6614763B1 (en) * 1999-02-04 2003-09-02 Fujitsu Limited Method of and apparatus for measuring network communication performances, as well as computer readable record medium having network communication performance measuring program stored therein
CN1488214A (en) * 2001-04-19 2004-04-07 �ʼҷ����ֵ������޹�˾ Method and device for robust real-time estimation of the bottleneck bandwidth in the internet
CN1543157A (en) * 2003-03-03 2004-11-03 Controlling admission of data streams onto a network based on end-to-end measurements
CN1545277A (en) * 2003-11-19 2004-11-10 北京邮电大学 An end-to-end network bottleneck bandwidth measuring method

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
An Empirical Evaluation of Wide-Area Internet Bottlenecks. Aditya Akella, Srinivasan Seshan, Anees Shaikh.IMC03. 2003 *
IP网络可用带宽测量技术的研究. 倪县乐，梁永芳，丁炜.电信科学. 2004 *
Locating the tightest link of a network path. D Zhang, W Huang, C Lin.ACM SIGMETRICS Performance Evaluation Review. 2004 *

Also Published As

Publication number Publication date
CN1658583A (en) 2005-08-24

Similar Documents

Publication Publication Date Title
CN100446479C (en) Method for measuring end-to-end available bandwidth based on network compact link position
Cáceres et al. Multicast-based inference of network-internal loss characteristics
JP4348124B2 (en) Method and communication device for estimating QoS
US8045478B2 (en) Performance measurement in a packet transmission network
JP4769609B2 (en) Switch device
JP2004289791A (en) Measurement architecture for obtaining packet loss and delay of one-way for every hop in multi-class service network
CN105245399B (en) The measure of bottleneck link in a kind of end to end network communication path
EP2936741B1 (en) Probing a network
CN103634157B (en) parallel message routing detection method
CN106656643A (en) Measuring method of segmental calculation of network delay
Aida et al. A scalable and lightweight QoS monitoring technique combining passive and active approaches
Suárez et al. Computer networks performance modeling and simulation
Salcedo Morillo et al. Overhead in available bandwidth estimation tools: Evaluation and analysis
Ribeiro et al. Spatio-temporal available bandwidth estimation with STAB
CN100396020C (en) Network tightening link circuit positioning method based on double rate periodic flow technolgy
Mnisi et al. Active throughput estimation using RTT of differing ICMP packet sizes
Fei et al. Accurate and effective inference of network link loss from unicast end-to-end measurements
Igai et al. A simple estimation scheme for upper bound of link utilization based on RTT measurement
JP2004312725A5 (en)
Zangrilli et al. Applying principles of active available bandwidth algorithms to passive tcp traces
Machiraju et al. A measurement-friendly network (MFN) architecture
Seddik-Ghaleb et al. Emulating end-to-end losses and delays for ad hoc networks
Fei et al. Unicast-based inference of network link delay statistics
Popescu et al. Measurement of one-way transit time in IP routers
Marttinen et al. Estimating KPIs in deployed heterogeneous 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: 20081224

Termination date: 20110310