CN106789429B - A kind of adaptive low-cost SDN network link utilization measurement method and system - Google Patents
A kind of adaptive low-cost SDN network link utilization measurement method and system Download PDFInfo
- Publication number
- CN106789429B CN106789429B CN201611214614.5A CN201611214614A CN106789429B CN 106789429 B CN106789429 B CN 106789429B CN 201611214614 A CN201611214614 A CN 201611214614A CN 106789429 B CN106789429 B CN 106789429B
- Authority
- CN
- China
- Prior art keywords
- statistics
- flow
- link utilization
- linkutil
- interface
- 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.)
- Active
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
- H04L43/0876—Network utilisation, e.g. volume of load or congestion level
Landscapes
- Engineering & Computer Science (AREA)
- Environmental & Geological Engineering (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
A kind of adaptive low-cost SDN network link utilization measurement method and system, steps of the method are: S1: defines SDN network link utilization and measure relevant parameter;S2: initiation parameter;S3: FlowRemoved message is listened to;If listened to, link utilization I_T_LinkUtil is calculated, S4 is skipped to;If do not listened to, the mode based on active measurement calculates link utilization I_T_LinkUtil, skips to rapid S5;S4: collecting the statistical information of data flow based on passive measurement mode, listens to FlowRemoved message;S5: collecting the statistical information of data flow based on active measurement method, sends FlowStatistics-Request message to interchanger;S6: data stream statistics information, record are extracted;S7: each data flow is successively calculated to the contribution rate of link utilization, record;S8: each member variable of I_T_F_LinkUtil [] is added, interface interface is calculated in the corresponding link utilization I_T_LinkUtil of periodic point, and record;S9: adjustment polling cycle T.The system is used to execute the above method.The present invention has many advantages, such as that principle is simple, applied widely, measurement accuracy is high.
Description
Technical field
Present invention relates generally to SDN network technical fields, refer in particular to a kind of adaptive low-cost SDN network link utilization
Measurement method.
Background technique
The measurement of network link utilization rate is the important content of network management, and on the one hand it helps to find link bottle in time
Neck provides data supporting for routing optimality, not only can ensure that the performance quality of application service, and can realize operator operation at
This reduction;On the other hand it facilitates making rational planning for for network link capacity, avoids the occurrence of resource and largely wastes or link bottle
The problems such as neck-shaped condition takes place frequently.Therefore, the promptly and accurately measurement of network link utilization rate has promotion network management efficiency important
Meaning.
For SDN network (such as data center network, extensive Campus Networks), due to the pipe using centralization
Reason mode, the reasonable administrative decision of controller tend to rely on to network actual motion state (including network link utilization rate view
Figure) precisely grasp to fine granularity, and also the typical case scene as SDN, the environment such as cloud data center, extensive Campus Networks are in
Reveal more complicated variation: applied business is many kinds of, the highly dynamic variation of network flow, these are all network link benefit
Huge challenge is brought with the accurate real-time measurement of rate.Under SDN network environment, how the accurate network link of acquisition in real time is sharp
With rate view, measurement cost can be effectively reduced again seem particularly significant.
Current network link utilization assessment with calculation method mainly include the following types:
(1) based on the calculation method of snmp protocol;
Port handling capacity, including value method, middle-value calculating are actively requested to target switch from controller using mib information
Method and maximum stream flow calculating method etc..For traditional IP, this method is by the way that periodically inquiry can be more preferably
Realize the true reflection to network link utilization power, but for SDN network, this method is difficult to realize measurement granularity and measurement
Balance between cost.
(2) calculation method based on elongated message VPS;
Detection host data packet of different sizes is sent into network, institute through each node on path to source node return
Message estimates bandwidth by analyzing transmission rate and the One Way Delay changing rule of probe data packet.For SDN network
Speech, it will usually a large amount of upper layer application is carried, causes Internet resources extremely limited, and this method measures expense in estimation process
It is excessive.
(3) based on message it is right/calculation method of message queue PPTD;
When adjacent two data packets pass through link, the time interval that the two increases is approximately that the size of the 2nd packet is removed
With the bandwidth of link.This method will not generate excessive load to network, but its accuracy places one's entire reliance upon on entire path
Accurate measurement to this adjacent data packet time interval, this there is two shortcomings: first is that the measurement of link utilization is not
It is enough timely, lack sensitivity;Second is that require background interference few, and SDN network flow is highly dynamic variation, environment is unstable
It is fixed.
(4) based on the calculation method of tomoscan;
Measurement data end to end is obtained according to certain measurement strategies, and by establishing linear mathematical model, realization pair
The supposition of network internal link utilization.This method is unrelated with the network protocol that internal network uses, and is also not required in measurement process
Any cooperation for wanting internal network greatly reduces and measures spent expense, but since its accuracy depends on institute
The accuracy of linear model is established, so usually measurement error is larger.
In conclusion above-mentioned network link utilization rate assessment calculation method in the prior art has the disadvantage that
1) active probe message is relied on, biggish measurement can be carried out to Netowrk tape and loaded;
2) measurement granularity is big, is not able to satisfy the performance management demand of cloud data center, extensive Campus Networks etc.;
3) measurement error is big, due to the highly dynamic variation of data flow, the method tool based on message length variation, message interval
There is limitation, and tomoscan is based on statistical method and inherently there is error.
Summary of the invention
The technical problem to be solved in the present invention is that, for technical problem of the existing technology, the present invention provides one
Simple, applied widely, measurement accuracy the is high adaptive low-cost SDN network link utilization measurement method of kind principle.
In order to solve the above technical problems, the invention adopts the following technical scheme:
A kind of adaptive low-cost SDN network link utilization measurement method, the steps include:
S1: it defines SDN network link utilization and measures relevant parameter;
S2: according to measurement target, successively initiation parameter;Assuming that the polling cycle point of Flow-Statistics is t;
S3: FlowRemoved message is listened in 2 μ time windows of t periodic point;If listening to FlowRemoved report
Text then calculates link utilization I_T_LinkUtil in such a way that FlowRemoved message is based on passive measurement, jumps to step
Rapid S4;If not listening to FlowRemoved message, it is based on actively measuring using FlowStatistics-Request message
Mode calculate link utilization I_T_LinkUtil, go to step S5;
S4: the statistical information Flow-Statistics of data flow is collected based on passive measurement mode, is listened to
FlowRemoved message;Assuming that being ζ at the time of listening to FlowRemoved message for the first time in 2 μ time windows, when recording ζ
Active all data flow I_T_ActiveFlow are carved, and I_T_ActiveNum is recorded in its quantity;By Flow-
Statistics is extracted and regularization module successively listens to FlowRemoved message corresponding to each alive data stream;
S5: collecting the statistical information Flow-Statistics of data flow based on active measurement method, sends to interchanger
FlowStatistics-Request message;Record period point t+ μ moment active all data flow I_T_ActiveFlow, and
I_T_ActiveNum is recorded in its quantity;FlowStatistics-Request message, switch responds are sent to interchanger
And FlowStatistics-Response message is returned to, and this report is listened to by Flow-Statistics extraction and regularization module
Text;
S6: data stream statistics information is extracted from FlowRemoved message or FlowStatistics-Response message
I_T_F_Statistics [] is recorded in FlowStatistics;
S7: it successively calculates each data flow in I_T_F_Statistics [] and I_ is recorded to the contribution rate of link utilization
T_F_LinkUtil [], and judge that final link utilization I_T_LinkUtil can be calculated;
S8: each member variable of I_T_F_LinkUtil [] is added, and it is corresponding in periodic point that interface interface is calculated
Link utilization I_T_LinkUtil, and the corresponding link utilization set I_ of the interface interface is recorded
LinkUtil[];
S9: calculating the variation delta LinkUtil of this and a preceding link utilization, and according to I_T_ActiveNum and
The polling cycle T of Δ LinkUtil adjustment Flow-Statistics.
Further improvement as the method for the present invention: the step S1 includes:
1.1 traffic flow informations are described as follows, and: I_T_ActiveFlow={ interface, time, AF_list } is
The alive data adfluxion at interface interface time moment is closed, and I_T_F_Statistics []=interface, time,
Flow, statistics } be interface interface time time data stream flow statistical information, I_T_ActiveNum=
{ interface, time, num } is the number of interface interface time moment alive data stream;
Each parameter when 1.2 data stream statistics information Flow-Statistics are collected is described as follows, T Flow-
The polling cycle of Statistics;(t- μ, t+ μ) is the time window that periodic point t listens to FlowRemoved message;When ζ is 2 μ
Between at the time of listen to FlowRemoved message in window;W is the high latency that link utilization calculates;
Each parameter when 1.3 data stream statistics information Flow-Statistics polling cycles adjusted is described as follows, and α is to increase
The big Flow-Statistics collection period multiplies parameter;β is the reduction Flow-Statistics collection period except parameter;NFmax
The threshold parameter of number of data streams can be carried for the link of setting;ΔmaxFor the upper limit value parameter of link utilization variation;
ΔminFor the lower limit value parameter of link utilization variation;
1.4 network link utilization rate parameters are described as follows, and I_LinkUtil [] is the corresponding link benefit of interface interface
With rate set;I_T_LinkUtil is interface interface time moment corresponding link utilization;I_T_F_LinkUtil
[] is contribution rate set of each data flow to link utilization I_T_LinkUtil;Δ LinkUtil is the change of link utilization
Change.
Further improvement as the method for the present invention: the process of the step S4 are as follows:
S4.1: in the 2 μ windows (t- μ, t+ μ) of periodic point, FlowRemoved message is listened to, it is assumed that specifically the moment is
ζ;
S4.2: it checks the active data flow of ζ moment Target Link, and initializes I_T_ActiveFlow and I_T_
ActiveNum;
S4.3: the corresponding FlowRemoved message of all alive data streams is successively listened to, and therefrom parses data flow
Statistical information Flow-Statistics, record to I_T_F_Statistics [].
Further improvement as the method for the present invention: the process of the step S5 are as follows:
S5.1: in the 2 μ windows (t- μ, t+ μ) of periodic point, Flow-Statistics enquiry module is not listened to
FlowRemoved message;
S5.2: it checks the active data flow of t+ μ moment Target Link, and initializes I_T_ActiveFlow and I_T_
ActiveNum;
S5.3: FlowStatistics-Request message is sent to target switch;
S5.4: target switch responds FlowStatistics-Request, sends FlowStatistics- to controller
Response message;
S5.5: listening to FlowStatistics-Response message, and therefrom parses the statistical information of data flow
Flow-Statistics, record to I_T_F_Statistics [].
Further improvement as the method for the present invention: the process of the step S9 are as follows:
S9.1: the active number of data streams I_T_ActiveNum of current ink and neighbouring link utilization twice are determined
Variation delta LinkUtil;
S9.2: judge whether I_T_ActiveNum is greater than preset link and can carry the threshold value of number of data streams
NFmax, if it is greater, then reducing Flow-Statistics polling cycle T=T/ β;If it is not greater, then jumping to S9.3;
S9.3: judge whether Δ LinkUtil is greater than the upper limit value Δ of preset link utilization variationmax, if it does,
Then reduce Flow-Statistics polling cycle T=T/ β;If it is not greater, then jumping to S9.4;
S9.4: judge whether Δ LinkUtil is less than the lower limit value Δ of preset link utilization variationmin, if it is lower,
Then increase Flow-Statistics polling cycle T=T* α;If it is not, then Flow-Statistics polling cycle T is kept
It is constant;
S9.5: Flow-Statistics polling cycle T is updated.
Further improvement as the method for the present invention: the process of the step S7 are as follows:
S7.1 judges whether the calculating time limit of link utilization has exceeded;For passive measurement mode, that is, judge be this moment
It is no to have exceeded ζ+W;For active measurement method, that is, judge whether exceeded t+ μ+W this moment;If exceeded, go to step
S7.2;If without departing from go to step S7.3;
S7.2 judges whether the contribution rate of all alive data streams has all been computed, that is, judges current I_T_F_LinkUtil
Whether the number of members of [] is equal to I_T_ActiveNum;If equal, go to step S8;If unequal, by Flow-
Statistics is extracted and regularization module continues to listen to subsequent FlowRemoved message or FlowStatistics-
Response message;
S7.3 update cycle point t is current time, restores each variable, go to step S5.
The present invention further provides a kind of adaptive low-cost SDN network link utilization measuring systems, comprising:
Flow-Statistics enquiry module, for being responsible for collecting data stream statistics letter from interchanger in each periodic point
Flow-Statistics is ceased, perhaps collects or utilizes in such a way that FlowRemoved message is based on passive measurement
FlowStatistics-Request message is collected based on the mode of active measurement;Which kind of mode is specifically taken to depend on the module
FlowRemoved message whether is listened in (t- μ, t+ μ) time window, passive measurement mode is used if listening to,
It is on the contrary then use active measurement method;
Flow-Statistics is extracted and regularization module, for being responsible for listening to the FlowRemoved report from interchanger
Text or FlowStatistics-Response message, and from the statistical information Flow- for wherein parsing data flow
Statistics, record to I_T_F_Statistics [];
Data flow is used to successively be counted according to the record result of I_T_F_Statistics [] to link contribution rate computing module
Each data flow is calculated to the contribution situation of link utilization, and is recorded to I_T_F_LinkUtil [];
Link utilization computing module is used to calculate final link benefit according to the record result of I_T_F_LinkUtil []
With rate, it can obtain whether all final link utilization depends on the contribution rate of all alive data stream I_T_ActiveFlow
It has been being computed that, i.e., I_T_F_LinkUtil [], whether number of members was equal to I_T_ActiveNum, if unequal, by
Flow-Statistics is extracted and regularization module continues to listen to FlowRemoved or FlowStatistics-Response report
Text, but if it has been more than preset upper limit that link utilization, which calculates the time, the Flow-Statistics inquiry for being zeroed original,
Measurement method of taking the initiative is inquired again;
Flow-Statistics analysis module, for being responsible for calculating the variation delta of neighbouring link utilization twice
LinkUtil, using one of the foundation as adjustment Flow-Statistics polling cycle;
Flow-Statistics polling cycle adjusts module, according to link current active stream quantity and neighbouring chain twice
The variable quantity of road utilization rate updates Flow-Statistics polling cycle to adjust.
Further improvement as present system: the Flow-Statistics enquiry module is also used to initialize I_
T_ActiveFlow and I_T_ActiveNum.
Further improvement as present system: the Flow-Statistics analysis module is used to be responsible for analysis
Flow-Statistics and link utilization information provide effective link utilization view to tension management application.
Compared with the prior art, the advantages of the present invention are as follows: adaptive low-cost SDN network link utilization of the invention
Measurement method and system are a kind of SDN network link utilization measurement methods based on actively and passively mode, to solve
SDN network (the especially high performance network such as cloud data center network, extensive Campus Networks) network link utilization rate calculates excessive
Dependent on active probe message, measurement granularity is not fine enough, measurement scheme less pertinence leads to problems such as measurement error bigger than normal;
The present invention has many advantages, such as that principle is simple, applied widely, measurement accuracy is high.
Detailed description of the invention
Fig. 1 is the flow diagram of the method for the present invention.
Fig. 2 is the flow diagram of present invention Flow-Statistics passive query in specific application example.
Fig. 3 is the flow diagram of present invention Flow-Statistics active inquiry in specific application example.
Fig. 4 is the present invention flow diagram that Flow-Statistics polling cycle adjusts in specific application example.
Specific embodiment
The present invention is described in further details below with reference to Figure of description and specific embodiment.
As shown in Figure 1, adaptive low-cost SDN network link utilization measurement method of the invention, is a kind of based on master
Dynamic and passive mode SDN network link utilization measurement method, its steps are as follows:
S1: it defines SDN network link utilization and measures relevant parameter.
1.1 traffic flow informations are described as follows, and: I_T_ActiveFlow={ interface, time, AF_list } is
The alive data adfluxion at interface interface time moment is closed, and I_T_F_Statistics []=interface, time,
Flow, statistics } be interface interface time time data stream flow statistical information, I_T_ActiveNum=
{ interface, time, num } is the number of interface interface time moment alive data stream.
Each parameter when 1.2 data stream statistics information Flow-Statistics are collected is described as follows, T Flow-
The polling cycle of Statistics;(t- μ, t+ μ) is the time window that periodic point t listens to FlowRemoved message;When ζ is 2 μ
Between at the time of listen to FlowRemoved message in window;W is the high latency that link utilization calculates.
Each parameter when 1.3 data stream statistics information Flow-Statistics polling cycles adjusted is described as follows, and α is to increase
The big Flow-Statistics collection period multiplies parameter;β is the reduction Flow-Statistics collection period except parameter;NFmax
The threshold parameter of number of data streams can be carried for the link of setting;ΔmaxFor the upper limit value parameter of link utilization variation;
ΔminFor the lower limit value parameter of link utilization variation.
1.4 network link utilization rate parameters are described as follows, and I_LinkUtil [] is the corresponding link benefit of interface interface
With rate set;I_T_LinkUtil is interface interface time moment corresponding link utilization;I_T_F_LinkUtil
[] is contribution rate set of each data flow to link utilization I_T_LinkUtil;Δ LinkUtil is the change of link utilization
Change.
S2: according to measurement target, successively initialize I_T_ActiveFlow, I_T_FlowStatistics, T, μ, α, β,
Δmax、Δmin、NFmaxParameter;Assuming that the polling cycle point of Flow-Statistics is t.
S3:Flow-statistics enquiry module listens to FlowRemoved message in 2 μ time windows of t periodic point.
If listening to FlowRemoved message, link utilization is calculated in such a way that FlowRemoved message is based on passive measurement
Rate I_T_LinkUtil, go to step S4;If not listening to FlowRemoved message, FlowStatistics- is utilized
Request message calculates link utilization I_T_LinkUtil based on the mode of active measurement, and go to step S5.
S4: the statistical information Flow- of data flow is collected based on passive measurement mode (listening to FlowRemoved message)
Statistics.Assuming that being ζ at the time of listening to FlowRemoved message for the first time in 2 μ time windows, the record ζ moment is active
All data flow I_T_ActiveFlow, and I_T_ActiveNum is recorded in its quantity;It is mentioned by Flow-Statistics
It takes and regularization module successively listens to FlowRemoved message corresponding to each alive data stream.
S5: it is collected based on active measurement method (controller sends FlowStatistics-Request message to interchanger)
The statistical information Flow-Statistics of data flow.Record period point t+ μ moment active all data flow I_T_
ActiveFlow, and I_T_ActiveNum is recorded in its quantity.Flow-statistics enquiry module is sent to interchanger
FlowStatistics-Request message, switch responds simultaneously return to FlowStatistics-Response message, and by
Flow-Statistics is extracted and regularization module listens to this message.Network is made in order to reduce acquisition Flow-Statistics
At huge load, interchanger merges the FlowStatistics-Response of a plurality of data flow, then returns to control
Device processed.
S6:Flow-Statistics is extracted and regularization module is from FlowRemoved message or FlowStatistics-
Data stream statistics information FlowStatistics is extracted in Response message, and regularization is carried out according to unified format, then
I_T_F_Statistics [] is recorded.
S7: data flow successively calculates in I_T_F_Statistics [] each data flow to chain to link contribution rate computing module
Can the contribution rate of road utilization rate be recorded I_T_F_LinkUtil [], and judge calculate final link utilization I_T_
LinkUtil。
S8: each member variable of I_T_F_LinkUtil [] is added, and it is corresponding in periodic point that interface interface is calculated
Link utilization I_T_LinkUtil, and the corresponding link utilization set I_ of the interface interface is recorded
LinkUtil[]。
S9: calculating the variation delta LinkUtil of this and a preceding link utilization, and according to I_T_ActiveNum and
The polling cycle T of Δ LinkUtil adjustment Flow-Statistics.The detailed process of step S9 are as follows:
S9.1 judges whether I_T_ActiveNum is greater than preset link and can carry the threshold value of number of data streams
NFmax.If it is greater, then reducing Flow-Statistics polling cycle T=T/ β;If it is not greater, then going to step
S9.2。
S9.2 judges whether Δ LinkUtil is greater than the upper limit value Δ of preset link utilization variationmax.If it does,
Then reduce Flow-Statistics polling cycle T=T/ β;If it is not greater, then the S9.3 that gos to step.
S9.3 judges whether Δ LinkUtil is less than the lower limit value Δ of preset link utilization variationmin.If it is lower,
Then increase Flow-Statistics polling cycle T=T* α;If it is not, then Flow-Statistics polling cycle T is kept
It is constant.
As shown in Fig. 2, be Flow-Statistics passive query flow chart, i.e., it is of the invention in specific application example
The detailed process of above-mentioned steps S4 are as follows:
S4.1: in the 2 μ windows (t- μ, t+ μ) of periodic point, Flow-Statistics enquiry module is listened to
FlowRemoved message, it is assumed that the specific moment is ζ;
S4.2: it checks the active data flow of ζ moment Target Link, and initializes I_T_ActiveFlow and I_T_
ActiveNum;
S4.3: it is corresponding that all alive data streams are successively listened to by Flow-Statistics extraction and regularization module
FlowRemoved message, and the statistical information Flow-Statistics of data flow is therefrom parsed, record to I_T_F_
Statistics[]。
As shown in figure 3, be Flow-Statistics active inquiry flow chart, i.e., it is of the invention in specific application example
The detailed process of above-mentioned steps S5 are as follows:
S5.1: in the 2 μ windows (t- μ, t+ μ) of periodic point, Flow-Statistics enquiry module is not listened to
FlowRemoved message;
S5.2: it checks the active data flow of t+ μ moment Target Link, and initializes I_T_ActiveFlow and I_T_
ActiveNum;
S5.3: FlowStatistics-Request is sent from Flow-Statistics enquiry module to target switch
Message;
S5.4: target switch responds FlowStatistics-Request, sends FlowStatistics- to controller
Response message, in order to reduce the load for bringing network, after multiple FlowStatistics-Response messages being merged
It is then forwarded to controller;
S5.5: being extracted by Flow-Statistics and regularization module listens to FlowStatistics-Response report
Text, and the statistical information Flow-Statistics of data flow is therefrom parsed, record to I_T_F_Statistics [].
As shown in figure 4, being Flow-Statistics polling cycle of the present invention to adjust flow chart, i.e., in specific application example
In, the detailed process of above-mentioned steps S9 of the invention are as follows:
S9.1: the active number of data streams I_T_ of current ink is determined by Flow-Statistics analysis module
The variation delta LinkUtil of ActiveNum and neighbouring link utilization twice;
S9.2: judge whether I_T_ActiveNum is greater than preset link and can carry the threshold value of number of data streams
NFmax, if it is greater, then reducing Flow-Statistics polling cycle T=T/ β;If it is not greater, then jumping to S9.3;
S9.3: judge whether Δ LinkUtil is greater than the upper limit value Δ of preset link utilization variationmax, if it does,
Then reduce Flow-Statistics polling cycle T=T/ β;If it is not greater, then jumping to S9.4;
S9.4: judge whether Δ LinkUtil is less than the lower limit value Δ of preset link utilization variationmin, if it is lower,
Then increase Flow-Statistics polling cycle T=T* α;If it is not, then Flow-Statistics polling cycle T is kept
It is constant.
S9.5: module is adjusted by Flow-Statistics polling cycle and updates Flow-Statistics polling cycle T.
In specific application example, the detailed process of above-mentioned steps S7 of the invention are as follows:
S7.1 judges whether the calculating time limit of link utilization has exceeded.For passive measurement mode, that is, judge be this moment
It is no to have exceeded ζ+W;For active measurement method, that is, judge whether exceeded t+ μ+W this moment.If exceeded, go to step
S7.2;If without departing from go to step S7.3.
S7.2 judges whether the contribution rate of all alive data streams has all been computed, that is, judges current I_T_F_LinkUtil
Whether the number of members of [] is equal to I_T_ActiveNum.If equal, step S8 of the invention is jumped to;If unequal,
Then extracted by Flow-Statistics and regularization module continue to listen to subsequent FlowRemoved message or
FlowStatistics-Response message.
S7.3 update cycle point t is current time, restores each variable, jumps to step S5 of the invention.
In conjunction with Fig. 1, adaptive low-cost SDN network link utilization measuring system of the invention is come in a manner of module
Complete processing to each step, mainly include Flow-Statistics enquiry module, Flow-Statistics extract and
Regularization module, data flow analyze link contribution rate computing module, link utilization computing module, Flow-Statistics
Module, Flow-Statistics polling cycle adjust module.Wherein:
Flow-Statistics enquiry module is responsible for collecting data stream statistics information from interchanger in each periodic point
Flow-Statistics is perhaps collected or is utilized in such a way that FlowRemoved message is based on passive measurement
FlowStatistics-Request message is collected based on the mode of active measurement, which kind of mode is specifically taken to depend on the module
FlowRemoved message whether is listened in (t- μ, t+ μ) time window, passive measurement mode is used if listening to,
On the contrary then use active measurement method, which can initialize I_T_ActiveFlow and I_T_ActiveNum simultaneously.
Flow-Statistics is extracted and regularization module be responsible for listening to FlowRemoved message from interchanger or
FlowStatistics-Response message, and from the statistical information Flow-Statistics for wherein parsing data flow, remember
It records to I_T_F_Statistics [].Data flow is to link contribution rate computing module according to the note of I_T_F_Statistics []
Record result successively calculates each data flow to the contribution situation of link utilization, and records to I_T_F_LinkUtil [].
Link utilization computing module calculates final link utilization according to the record result of I_T_F_LinkUtil [],
It can obtain whether all final link utilization depends on the contribution rate of all alive data stream I_T_ActiveFlow
It calculates, i.e., I_T_F_LinkUtil [], whether number of members was equal to I_T_ActiveNum, if unequal, by Flow-
Statistics is extracted and regularization module continues to listen to FlowRemoved or FlowStatistics-Response message, but
It is if it has been more than preset upper limit that link utilization, which calculates the time, master is taken in the Flow-Statistics inquiry for being zeroed original
Dynamic measurement method is inquired again.
Flow-Statistics analysis module is responsible for calculating the variation delta LinkUtil of neighbouring link utilization twice,
Using one of the foundation as adjustment Flow-Statistics polling cycle, in addition also it is responsible for analysis Flow-Statistics and chain
Road utilization rate information, to provide effective link utilization view to tension management application.
Flow-Statistics polling cycle adjusts module according to link current active stream quantity and neighbouring link twice
The variable quantity of utilization rate updates Flow-Statistics polling cycle to adjust.
The above is only the preferred embodiment of the present invention, protection scope of the present invention is not limited merely to above-described embodiment,
All technical solutions belonged under thinking of the present invention all belong to the scope of protection of the present invention.It should be pointed out that for the art
For those of ordinary skill, several improvements and modifications without departing from the principles of the present invention should be regarded as protection of the invention
Range.
Claims (9)
1. a kind of adaptive low-cost SDN network link utilization measurement method, which is characterized in that step are as follows:
S1: it defines SDN network link utilization and measures relevant parameter;
S2: according to measurement target, successively initiation parameter;Assuming that the polling cycle point of Flow-Statistics is t;
S3: FlowRemoved message is listened in 2 μ time windows of t periodic point;If listening to FlowRemoved message,
Link utilization I_T_LinkUtil then is calculated in such a way that FlowRemoved message is based on passive measurement, is gone to step
S4;If not listening to FlowRemoved message, using FlowStatistics-Request message based on active measurement
Mode calculates link utilization I_T_LinkUtil, and go to step S5;
S4: collecting the statistical information Flow-Statistics of data flow based on passive measurement mode, listens to FlowRemoved report
Text;Assuming that being ζ, record ζ moment active all numbers at the time of listening to FlowRemoved message for the first time in 2 μ time windows
I_T_ActiveNum is recorded according to stream I_T_ActiveFlow, and by its quantity, I_T_ActiveFlow=interface,
Time, AF_list } it is closed for the alive data adfluxion at interface interface time moment, I_T_ActiveNum=
{ interface, time, num } is the number of interface interface time moment alive data stream;By Flow-Statistics
It extracts and regularization module successively listens to FlowRemoved message corresponding to each alive data stream;
S5: collecting the statistical information Flow-Statistics of data flow based on active measurement method, sends to interchanger
FlowStatistics-Request message;Record period point t+ μ moment active all data flow I_T_ActiveFlow, and
I_T_ActiveNum is recorded in its quantity;FlowStatistics-Request message, switch responds are sent to interchanger
And FlowStatistics-Response message is returned to, and this report is listened to by Flow-Statistics extraction and regularization module
Text;
S6: data stream statistics information is extracted from FlowRemoved message or FlowStatistics-Response message
FlowStatistics is recorded I_T_F_Statistics [], and I_T_F_Statistics []=interface, time,
Flow, statistics } be interface interface time time data stream flow statistical information;
S7: it successively calculates each data flow in I_T_F_Statistics [] and I_T_F_ is recorded to the contribution rate of link utilization
LinkUtil [], I_T_F_LinkUtil [] they are contribution rate set of each data flow to link utilization I_T_LinkUtil, and
Can judgement calculate final link utilization I_T_LinkUtil;
S8: each member variable of I_T_F_LinkUtil [] is added, interface interface is calculated in the corresponding chain of periodic point
Road utilization rate I_T_LinkUtil, and the corresponding link utilization set I_LinkUtil of the interface interface is recorded
[];
S9: the variation delta LinkUtil of this and preceding a link utilization is calculated, and according to I_T_ActiveNum and Δ
The polling cycle T of LinkUtil adjustment Flow-Statistics.
2. adaptive low-cost SDN network link utilization measurement method according to claim 1, which is characterized in that institute
Stating step S1 includes:
1.1 traffic flow informations are described as follows: I_T_ActiveFlow={ interface, time, AF_list } is interface
The alive data adfluxion at interface time moment is closed, and I_T_F_Statistics []=interface, time, flow,
Statistics } be interface interface time time data stream flow statistical information, I_T_ActiveNum=
{ interface, time, num } is the number of interface interface time moment alive data stream;
Each parameter when 1.2 data stream statistics information Flow-Statistics are collected is described as follows, T Flow-Statistics
Polling cycle;(t- μ, t+ μ) is the time window that periodic point t listens to FlowRemoved message;ζ is to detect in 2 μ time windows
At the time of hearing FlowRemoved message;W is the high latency that link utilization calculates;
Each parameter when 1.3 data stream statistics information Flow-Statistics polling cycles adjusted is described as follows, and α is to increase
The Flow-Statistics collection period multiplies parameter;β is the reduction Flow-Statistics collection period except parameter;NFmaxFor
The link of setting can carry the threshold parameter of number of data streams;ΔmaxFor the upper limit value parameter of link utilization variation;
ΔminFor the lower limit value parameter of link utilization variation;
1.4 network link utilization rate parameters are described as follows, and I_LinkUtil [] is the corresponding link utilization of interface interface
Set;I_T_LinkUtil is interface interface time moment corresponding link utilization;I_T_F_LinkUtil [] is
Contribution rate set of each data flow to link utilization I_T_LinkUtil;Δ LinkUtil is the variation of link utilization.
3. adaptive low-cost SDN network link utilization measurement method according to claim 1, which is characterized in that institute
State the process of step S4 are as follows:
S4.1: in the 2 μ windows (t- μ, t+ μ) of periodic point, FlowRemoved message is listened to, it is assumed that the specific moment is ζ;
S4.2: it checks the active data flow of ζ moment Target Link, and initializes I_T_ActiveFlow and I_T_ActiveNum;
S4.3: the corresponding FlowRemoved message of all alive data streams is successively listened to, and therefrom parses the statistics of data flow
Information Flow-Statistics, record to I_T_F_Statistics [].
4. adaptive low-cost SDN network link utilization measurement method according to claim 1, which is characterized in that institute
State the process of step S5 are as follows:
S5.1: in the 2 μ windows (t- μ, t+ μ) of periodic point, Flow-Statistics enquiry module is not listened to
FlowRemoved message;
S5.2: it checks the active data flow of t+ μ moment Target Link, and initializes I_T_ActiveFlow and I_T_
ActiveNum;
S5.3: FlowStatistics-Request message is sent to target switch;
S5.4: target switch responds FlowStatistics-Request, sends FlowStatistics- to controller
Response message;
S5.5: listening to FlowStatistics-Response message, and therefrom parses the statistical information Flow- of data flow
Statistics, record to I_T_F_Statistics [].
5. adaptive low-cost SDN network link utilization measurement method according to claim 1, which is characterized in that institute
State the process of step S9 are as follows:
S9.1: the change of the active number of data streams I_T_ActiveNum of current ink and neighbouring link utilization twice are determined
Change amount Δ LinkUtil;
S9.2: judge whether I_T_ActiveNum is greater than preset link and can carry the threshold value NF of number of data streamsmax, such as
Fruit is greater than, then reduces Flow-Statistics polling cycle T=T/ β, and β is to reduce Flow-Statistics to collect removing for period
Parameter;If it is not greater, then jumping to S9.3;
S9.3: judge whether Δ LinkUtil is greater than the upper limit value Δ of preset link utilization variationmax, if it is greater, then contracting
Small Flow-Statistics polling cycle T=T/ β;If it is not greater, then jumping to S9.4;
S9.4: judge whether Δ LinkUtil is less than the lower limit value Δ of preset link utilization variationmin, if it is less, increasing
Big Flow-Statistics polling cycle T=T* α, α be increase the Flow-Statistics collection period multiply parameter;If no
It is less than, then Flow-Statistics polling cycle T is remained unchanged;
S9.5: Flow-Statistics polling cycle T is updated.
6. adaptive low-cost SDN network link utilization measurement method described according to claim 1~any one of 5,
It is characterized in that, the process of the step S7 are as follows:
S7.1 judges whether the calculating time limit of link utilization has exceeded;For passive measurement mode, that is, whether to judge this moment
It is the high latency that link utilization calculates beyond ζ+W, W;For active measurement method, that is, judge whether exceeded this moment
t+μ+W;If exceeded, go to step S7.2;If without departing from go to step S7.3;
S7.2 judges whether the contribution rate of all alive data streams has all been computed, that is, judges current I_T_F_LinkUtil []
Whether number of members is equal to I_T_ActiveNum;If equal, go to step S8;If unequal, by Flow-
Statistics is extracted and regularization module continues to listen to subsequent FlowRemoved message or FlowStatistics-
Response message;
S7.3 update cycle point t is current time, restores each variable, go to step S5.
7. a kind of adaptive low-cost SDN network link utilization measuring system characterized by comprising
Flow-Statistics enquiry module, for being responsible for collecting data stream statistics information from interchanger in each periodic point
Flow-Statistics is perhaps collected or is utilized in such a way that FlowRemoved message is based on passive measurement
FlowStatistics-Request message is collected based on the mode of active measurement;Which kind of mode is specifically taken to depend on the module
FlowRemoved message whether is listened in (t- μ, t+ μ) time window, (t- μ, t+ μ) is that periodic point t is listened to
The time window of FlowRemoved message uses passive measurement mode, on the contrary then use active measurement side if listening to
Formula;
Flow-Statistics is extracted and regularization module, for be responsible for listening to FlowRemoved message from interchanger or
FlowStatistics-Response message, and from the statistical information Flow-Statistics for wherein parsing data flow, remember
It records to I_T_F_Statistics [], I_T_F_Statistics []={ interface, time, flow, statistics }
For the statistical information of interface interface time time data stream flow;
Data flow is used to successively be calculated respectively according to the record result of I_T_F_Statistics [] to link contribution rate computing module
Data flow records to I_T_F_LinkUtil [] the contribution situation of link utilization, and I_T_F_LinkUtil [] is each number
According to stream to the contribution rate set of link utilization I_T_LinkUtil;
Link utilization computing module is used to calculate final link utilization according to the record result of I_T_F_LinkUtil []
Can rate, obtain whether all final link utilization depends on the contribution rate of all alive data stream I_T_ActiveFlow
It is computed, I_T_ActiveFlow={ interface, time, AF_list } is enlivening for interface interface time moment
Whether set of data flows, i.e. I_T_F_LinkUtil [] number of members are equal to I_T_ActiveNum, and I_T_ActiveNum is to live
The quantity of all data flows of jump is extracted by Flow-Statistics if unequal and regularization module continues to listen to
FlowRemoved or FlowStatistics-Response message, but if it has been more than default that link utilization, which calculates the time,
The upper limit, then the Flow-Statistics inquiry for being zeroed original, measurement method of taking the initiative are inquired again;
Flow-Statistics analysis module, for being responsible for calculating the variation delta LinkUtil of neighbouring link utilization twice,
Using one of the foundation as adjustment Flow-Statistics polling cycle;
Flow-Statistics polling cycle adjusts module, according to link current active stream quantity and the neighbouring benefit of link twice
It is adjusted with the variable quantity of rate and updates Flow-Statistics polling cycle.
8. adaptive low-cost SDN network link utilization measuring system according to claim 7, which is characterized in that institute
Flow-Statistics enquiry module is stated also to be used to initialize I_T_ActiveFlow and I_T_ActiveNum.
9. adaptive low-cost SDN network link utilization measuring system according to claim 7, which is characterized in that institute
It states Flow-Statistics analysis module to be used to be responsible for analysis Flow-Statistics and link utilization information, be managed to upper layer
The effective link utilization view of offer ought to be provided.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611214614.5A CN106789429B (en) | 2016-12-26 | 2016-12-26 | A kind of adaptive low-cost SDN network link utilization measurement method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611214614.5A CN106789429B (en) | 2016-12-26 | 2016-12-26 | A kind of adaptive low-cost SDN network link utilization measurement method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106789429A CN106789429A (en) | 2017-05-31 |
CN106789429B true CN106789429B (en) | 2019-11-22 |
Family
ID=58920835
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611214614.5A Active CN106789429B (en) | 2016-12-26 | 2016-12-26 | A kind of adaptive low-cost SDN network link utilization measurement method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106789429B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108650148B (en) * | 2018-04-16 | 2021-08-13 | 湖南省星岳天璇科技有限公司 | SDN data flow-statistics low-overhead acquisition method |
CN110677311A (en) * | 2019-04-24 | 2020-01-10 | 广州西麦科技股份有限公司 | Port flow statistic period adjusting method and related device |
CN111817910B (en) * | 2020-06-22 | 2021-08-13 | 电子科技大学 | Network level measuring method for network flow |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7116639B1 (en) * | 2000-12-21 | 2006-10-03 | International Business Machines Corporation | System and method for determining network discrete utilization |
CN1881909A (en) * | 2006-05-15 | 2006-12-20 | 西安西电捷通无线网络通信有限公司 | Method for co-collecting IP network performance by active type measure and passive type measure |
CN101192951A (en) * | 2006-11-29 | 2008-06-04 | 华为技术有限公司 | Measuring method and device for utilization rate of IPv6 network link and IPv6 network router |
CN104662841A (en) * | 2012-07-13 | 2015-05-27 | 适应性频谱和信号校正股份有限公司 | Method and system for performance measurement of a communication link |
US9438478B1 (en) * | 2015-11-13 | 2016-09-06 | International Business Machines Corporation | Using an SDN controller to automatically test cloud performance |
-
2016
- 2016-12-26 CN CN201611214614.5A patent/CN106789429B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7116639B1 (en) * | 2000-12-21 | 2006-10-03 | International Business Machines Corporation | System and method for determining network discrete utilization |
CN1881909A (en) * | 2006-05-15 | 2006-12-20 | 西安西电捷通无线网络通信有限公司 | Method for co-collecting IP network performance by active type measure and passive type measure |
CN101192951A (en) * | 2006-11-29 | 2008-06-04 | 华为技术有限公司 | Measuring method and device for utilization rate of IPv6 network link and IPv6 network router |
CN104662841A (en) * | 2012-07-13 | 2015-05-27 | 适应性频谱和信号校正股份有限公司 | Method and system for performance measurement of a communication link |
US9438478B1 (en) * | 2015-11-13 | 2016-09-06 | International Business Machines Corporation | Using an SDN controller to automatically test cloud performance |
Also Published As
Publication number | Publication date |
---|---|
CN106789429A (en) | 2017-05-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Li et al. | LLMP: exploiting LLDP for latency measurement in software-defined data center networks | |
Tahaei et al. | A multi-objective software defined network traffic measurement | |
US20150074258A1 (en) | Scalable performance monitoring using dynamic flow sampling | |
US20050232227A1 (en) | Method and apparatus for characterizing an end-to-end path of a packet-based network | |
CN106789429B (en) | A kind of adaptive low-cost SDN network link utilization measurement method and system | |
US8520547B2 (en) | System and method for measuring interface utilization using policers | |
US11171869B2 (en) | Microburst detection and management | |
EP3639475A1 (en) | Qoe-based catv network capacity planning and upgrade system | |
CN108600038A (en) | Adaptive low-cost SDN Business Streams based on ARIMA are handled up measuring method and system | |
CN110324260B (en) | Network function virtualization intelligent scheduling method based on flow identification | |
CN110149239B (en) | Network flow monitoring method based on sFlow | |
Bi et al. | On precision and scalability of elephant flow detection in data center with SDN | |
CN113132180B (en) | Cooperative type large flow detection method facing programmable network | |
CN110213530A (en) | Method for early warning, device and readable storage medium storing program for executing | |
CN111711545A (en) | Intelligent encrypted flow identification method based on deep packet inspection technology in software defined network | |
CN109714795A (en) | A kind of method for managing resource, resource management system and device based on SDN network slice | |
CN109952743B (en) | System and method for low memory and low flow overhead high flow object detection | |
CN100566265C (en) | Network route band width measuring method based on steady sampling sequence | |
CN102664773A (en) | Method and device for detecting network flow | |
CN102355376B (en) | Traffic monitoring method and device | |
Wang et al. | Modeling residual-geometric flow sampling | |
Alaslani et al. | Intelligent edge: An instantaneous detection of iot traffic load | |
JP2014116840A (en) | Communication quality estimation device | |
CN107820141B (en) | Method and system for sensing service flow of passive optical network | |
CN108183892B (en) | Message processing method and device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
TA01 | Transfer of patent application right | ||
TA01 | Transfer of patent application right |
Effective date of registration: 20180529 Address after: 410205 412-A, room 9, 4 building, phase 1, Changsha middle power software park, No. 39, Jian Shan Road, Changsha hi tech Development Zone, Hunan. Applicant after: Hunan star Yuen Tian Sheng Technology Co., Ltd. Address before: 410205 A10 901, Luyang District, 408 Luoxi Road, Tongzi, Changsha high tech Development Zone, Hunan Applicant before: Hunan Jiatai sail Network Technology Co. Ltd. |
|
GR01 | Patent grant | ||
GR01 | Patent grant |