CN105791034B - A kind of browsing class service-aware analysis method - Google Patents
A kind of browsing class service-aware analysis method Download PDFInfo
- Publication number
- CN105791034B CN105791034B CN201610320411.8A CN201610320411A CN105791034B CN 105791034 B CN105791034 B CN 105791034B CN 201610320411 A CN201610320411 A CN 201610320411A CN 105791034 B CN105791034 B CN 105791034B
- Authority
- CN
- China
- Prior art keywords
- time delay
- matter
- website
- sample
- matter difference
- 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
- 238000004458 analytical method Methods 0.000 title claims abstract description 23
- 230000004044 response Effects 0.000 claims description 31
- 230000008447 perception Effects 0.000 claims description 30
- 238000000034 method Methods 0.000 claims description 17
- 238000005070 sampling Methods 0.000 claims description 10
- 230000015572 biosynthetic process Effects 0.000 claims description 7
- 230000001934 delay Effects 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 3
- 230000001066 destructive effect Effects 0.000 claims 1
- 238000005457 optimization Methods 0.000 abstract description 13
- 238000012360 testing method Methods 0.000 abstract description 6
- 244000097202 Rathbunia alamosensis Species 0.000 abstract description 4
- 235000009776 Rathbunia alamosensis Nutrition 0.000 abstract description 4
- 238000012790 confirmation Methods 0.000 abstract description 2
- 238000010295 mobile communication Methods 0.000 abstract description 2
- 230000008901 benefit Effects 0.000 description 3
- 230000001174 ascending effect Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000006866 deterioration Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- PEDCQBHIVMGVHV-UHFFFAOYSA-N Glycerine Chemical compound OCC(O)CO PEDCQBHIVMGVHV-UHFFFAOYSA-N 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 230000003542 behavioural effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000002860 competitive effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 230000002045 lasting effect Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
-
- 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
-
- 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
- H04L43/16—Threshold monitoring
-
- 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
- H04L43/50—Testing arrangements
- H04L43/55—Testing of service level quality, e.g. simulating service usage
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Environmental & Geological Engineering (AREA)
- Quality & Reliability (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
A kind of browsing class service-aware analysis method belongs to mobile communication field, it solves how to efficiently use mass users browsing class service-aware data acquire from end side and analyzes and position demarcations to the reason of the whole network level of the browsing class service-aware difference of user, to provide reference and foundation to the bearing capacity of browsing class business and then the browsing class service-aware of promotion user to optimize network-related parameters, algorithm, agreement etc. with promotion promotion mobile network.The reason that matter difference is perceived to browse service from the terminal data acquired in practical mobile network (3G networks of several telecommunications companies of districts and cities) is analyzed, it is found that dns resolution time excessively high problem existing for certain network respectively, the problems such as Sina website's first packet time delay is bigger than normal, these problems have obtained the confirmation of the optimization department of telecommunications and have done corresponding network adjustment (optimization dns server, increase local Sina's CDN server etc.), and subsequent test, which also demonstrates problem and is resolved index, to get a promotion.
Description
Technical field
The invention belongs to mobile communication fields.
Background technology
Compared to the lower service quality (long time delay, low reliability) of internet, carrier class network service is telecom operation
The competitive advantage that quotient takes pride in for it always exerts the evaluation and optimization that put into network quality, but in this way to keep its advantage
Ossifing for thinking is resulted in, i.e., always stands and sees problem, the light user experience of weight network quality in network perspective.It is interconnected in recent years
The strong challenge of the OTT business (good business experience) in camp is netted, the market share is gradually lost, and has the danger for becoming pipeline
Danger.It changes network quality evaluation and optimization method is extremely urgent.
The traditional network evaluation of operator and optimization method mainly use drive test (DT/CQT, from network level rather than operation layer
Network KPI indexs are focused in face) network quality (i.e. KPI index systems) is evaluated, with the high speed hair of mobile Internet business
There are increasing price scissors for exhibition, network quality and customer service perception.Operator start gradually to explore in assessing network and
User is introduced in optimization perceives (QoE).DPI (depth Packet analyzing) method is paid attention in recent years, but there are still limitations, still de-
From the not natural mode of quality in terms of network.
User's perception is critically important, but the means for obtaining real user perception at present have a limitation, and one generally acknowledged most basic
Principle is more closer more accurate from user, therefore how particularly critical from user side acquisition perception, that is, is stood at the angle of user's (terminal)
Degree sees network.Therefore, part operator attempts to start to use the perception big data progress network quality acquired based on end side
It evaluates and is included in examination.The advantage of doing so is that on the basis of accurately grasping user's perception, network can be more effectively promoted
Service quality and user satisfaction improve user's viscosity, while can be further used for instructing the marketing, the client of precision
Service increases user's ARP values.
But how delimit positioning to the poor problem of perception and guides network adjustment to promote service-aware, still lacks system
The method of property.Current thinking still carries out assessing network with the perception big data of end side acquisition, has not seen on how to profit
The method that the perception big data acquired with end side carries out network problem positioning and optimization.And network problem positioning and optimization are promoted
Still mainly around utilizing DT/CQT and DPI to carry out.
The analysis optimization method of service-aware worsening reason includes mainly at present:
The defect of method one (problem repetition and deep investigation are carried out based on drive test, that is, DT/CQT):
(1) it is difficult to effectively reappear the on-fixed problem, it is difficult to find the perception deterioration problem of non-wireless network factor.
Current Main be exactly the customer complaint the problem of place carry out a large amount of DT/CQT test calls, on the one hand where acquisition
The network-related parameters of position, at the same time it is wished that reproduction the complained problem of user (not getting through phone such as) by this method, but by
In the variability of network signal and network user's load, often many network quality problems are not fixations, it is difficult to reappear.
Other than attempting reproduction in addition to a large amount of testing, without good method.
(2) larger manpower and time cost are expended.
The defect of method two (being based on DPI data analyses):
(1) DPI Disposition & Operation cost is huge, can not carry out in the area for not having DPI systems;
(2) can not accurately acquire with par time delay class perception information end to end, in the case of can not acquiring no network interaction
Service-aware information.Therefore the perception caused by such factor can not also be deteriorated and effectively analyze and position.
Invention content
The present invention is to solve how to efficiently use the mass users acquired from end side to browse class service-aware data pair
Demarcation is analyzed and positioned to the reason of the whole network level of the browsing class service-aware difference of user, to be optimization network correlation ginseng
Number, algorithm, agreement etc. are to promote mobile network to the bearing capacity of browsing class business and then promote the browsing class industry of user
Business perception provides reference and foundation.
A kind of browsing class service-aware analysis method, which is characterized in that include the following steps:
Step 1:Perception data acquires
Acquire user browsing class business conduct information, including information should include:Date, the time, network formats,
Cell ID, terminal longitude and latitude are currently located province and districts and cities' title, field strength, signal quality, user identifier, terminal iidentification, terminal
Model, browser title, browse web sites title, and browse web sites URL, and browse web sites IP, first packet time delay, page open time delay, DNS
Parse time delay, TCP connection time delay, GET request time delay and reception response delay;
Step 2:Perception data cleans
Above-mentioned original sampling data is cleaned before being stored in database, browse service perception information is inserted into and adopts
Sample tables of data, data item include:On the date, the time, network type, cell ID, current home saves and districts and cities, terminal longitude and latitude,
Field strength, signal quality, user identifier, terminal iidentification, terminal models, browser title, browse web sites title, and browse web sites URL,
Browse web sites IP, first packet time delay, page open time delay, dns resolution time delay, TCP connection time delay, sends request time delay and receives and rings
Answer time delay;
Step 3:Matter difference sample judges
Take the index that most important two perception index is evaluated as matter difference in sampled data, i.e. " first packet time delay " and " page
Open time delay in face ", correspondingly, matter difference thresholding includes " first packet time delay matter difference thresholding " and " page open time delay matter difference thresholding ";
According in browse service perception information sampling data table data storing order and preset matter difference thresholding, in table
Each efficiently sampling data judge whether it is matter difference sample one by one;For the 4 i.e. DNS of delay in phases acquired in sampled data
It parses time delay, TCP connection time delay, GET request time delay and receives response delay, it is also respectively that it is poor with respective pre-set matter
Thresholding is compared, and is judged and is marked whether the sample is " dns resolution time-out sample ", " TCP connection time-out sample ", " GET is asked
Seek overtime sample " and " receiving response timeout sample ";
Step 4:Network level matter difference is analysed
This step is from 6 different dimensions, i.e. website, website IP, delay in phases, time terminal and user, to that may cause
The reason of network level of the browsing class service-aware difference of user, carries out analyzing and positioning in network;
Step 5:The matter difference origin cause of formation exports
It is exported after above-mentioned each germplasm difference origin cause of formation is summarized.
Step 4 is specific as follows:
Step 4a:Matter difference web analytics
The targeted website that several websites are the analysis of network level matter difference is selected first,
To selected N number of target browsing class website, each website this month total number of samples, first packet time delay matter difference sample are counted respectively
Number, page open time delay matter difference sample number, and calculate separately first packet matter rate, page matter rate and business according to following formula
Matter rate:
First packet matter rate=first packet time delay matter difference sample number/total number of samples * 100%
(1)
Page matter rate=page open time delay matter difference sample number/total number of samples * 100%
(2)
Business matter rate=first packet matter rate * w1+ page matter rate * w2
(3)
Wherein w1 and w2 is respectively the weight of first packet matter rate and page matter rate,
If certain business matter rate to browse web sites is more than pre-determined threshold, which is " matter difference website ";
Step 4b:The IP analyses of matter difference
To single website, the business matter rate of the IP whole sample points, the sample of the IP are counted as unit of the IP of the website
This number accounts for the ratio of the website total number of samples;To business matter rate be more than pre-determined threshold IP, be identified as the website " matter is poor
IP";
Step 4c:Delay in phases is analyzed
For four composition periods of a web page browsing process page open time delay:When dns resolution time delay, TCP connection
Prolong, GET request time delay, receive response delay;The matter rate for analyzing every section of time delay finds out time delay segmentation of problems;
Piecewise analysis matter difference time delay is segmented:
(1) DNS timeout issues:If in the page open time delay matter difference sample of certain website, dns resolution time-out sample number is big
In pre-determined threshold, then positions the website and there is " DNS timeout issues ", and calculate all page open time delay matter difference samples in the website
" average dns resolution time delay ";
(2) TCP connection timeout issue:If in the page open time delay matter difference sample of certain website, TCP connection time-out sample
Number is more than pre-determined threshold, then positions the website and there is " TCP connection timeout issue ", and calculate all page open time delays in the website
" the average TCP connection time delay " of matter difference sample;
(3) GET request timeout issue:If in the page open time delay matter difference sample of certain website, GET request time-out sample
Number is more than pre-determined threshold, then positions the website and there is " GET request timeout issue ", and calculate all page open time delays in the website
" the average GET request time delay " of matter difference sample;
(4) response timeout problem is received:If in the page open time delay matter difference sample of certain website, receiving response timeout sample
This number is more than pre-determined threshold, then positions the website and there are problems that " receiving response timeout ", and calculate all page opens in the website
" the average received response delay " of time delay matter difference sample;
Step 4d:Piecewise analysis when matter difference
As unit of 24 hours one day each hours, total sample of the target complete website within each hour is counted respectively
Number, first packet time delay matter difference sample number, page open time delay matter difference sample number, and calculated separately according to the formula of front (1)~(3)
First packet matter rate, page matter rate and business matter rate;
If the business matter rate within some period is more than pre-determined threshold, which is identified as " matter difference period ";
Step 4e:Poor-quality terminal is analyzed
Reach each terminal models of certain predetermined thresholding for the total sample size of business, counts target complete website respectively every
Total number of samples, first packet time delay matter difference sample number, page open time delay matter difference sample number under a terminal models, and according to front
Formula (1)~(3) calculate separately first packet matter rate, page matter rate and business matter rate;
If the business matter rate of some terminal models is more than pre-determined threshold, which is identified as " matter difference end
End ";
Step 4f:Matter difference customer analysis
Reach the single user of certain predetermined thresholding for the total sample size of business, counts target complete website respectively each
Total number of samples, first packet time delay matter difference sample number, page open time delay matter difference sample number under user, and according to the formula of front
(1)~(3) calculate separately the first packet matter rate, page matter rate and business matter rate of each user;
If the business matter rate of some user is more than pre-determined threshold, the IMSI of the user or IMEI are identified as " matter
Poor user ";
Step 4a-4f need not be limited and be executed sequence.
Description of the drawings
Fig. 1 is the complete algorithm flow chart of the present invention.
Fig. 2 is the detail flowchart of step 4 in algorithm flow.
Specific implementation mode
The present invention provides a kind of using the browsing class business conduct data of user's intelligent terminal acquisition, from the level of the whole network
Carry out the analysis of the mobile network user business matter difference origin cause of formation.Wherein step 1~2 are lasting progress, and step 3~5 are generally the period
Property carry out (such as each calendar month carries out primary, is described by taking the moon as an example below).Specific steps are described in detail as follows:
Step 1:Perception data acquires
The interface API provided from the operating system of terminal by being deployed in the forms such as the App on mass users intelligent terminal
Acquire user browsing class business App (such as UCweb, QQ browser) behavioural information, including information should include:Day
Phase, time, network formats, cell ID, terminal longitude and latitude are currently located province and districts and cities' title, field strength (name under different standards
Claim different:Such as the RxLevel of GSM network, the RSRP etc. of LTE network), (title is different under different standards for signal quality:Such as C/
I or SINR or RSRQ etc.), user identifier (IMSI), terminal iidentification (IMEI or MEID), terminal models, browser App titles,
Browse web sites title, and browse web sites URL, and browse web sites IP, first packet time delay, page open time delay, dns resolution time delay, TCP connection
Time delay, GET request time delay receive response delay etc..
Wherein:Cell ID is the combination for the identification parameter for uniquely determining a cell, is generally compiled by great Qu numbers+cell
Number composition.It is different to parameter name used in heterogeneous networks, as GSM, WCDMA, TD-SCDMA network LAC+CI, LTE are
TAC+PCI+CI。
Wherein:Being currently located province and districts and cities' title generally can be by auxiliary positioning (such as Baidu positions) API of calling terminal
Interface obtains.
Wherein:" first packet time delay " be defined as from user initiate web page browsing request to receive destination server respond first
The duration undergone between a HTTP 200OK message packets.First packet time delay=dns resolution time delay+TCP connection response delay+
GET request response delay.
Wherein:" page open time delay " is defined as finishing (only to entire HTTP page-downloadings from user's initiation browse request
Page text content, do not include resource second load) duration.Page open time delay=first packet time delay+reception response delay.
Wherein:" dns resolution time delay " refers to since terminal initiates dns resolution request to the time delay for completing dns resolution;"TCP
Connection time delay ", which refers to, to be terminated to establish the time delay completed to TCP connection (three-way handshake) from dns resolution;" GET request time delay " refers to from hair
Go out GET request to the time delay for receiving first TCP data packet (200OK containing HTTP);" receive response delay " refers to from receiving first
A response data packet starts to terminal Fa Chu [FIN, ACK]The time delay of (finishing receiving).
Step 2:Perception data cleans
Above-mentioned original sampling data is cleaned before being stored in database, main includes rejecting invalid data (key
Data item lacks, and numerical value overflows etc.), being inserted into " browse service perception information sampling data table ", (data item includes:Date,
Time, network type, cell ID, current home saves and districts and cities, terminal longitude and latitude, field strength, signal quality, user identifier, eventually
End mark, terminal models, browser App titles, browse web sites title, and browse web sites URL, and browse web sites IP, first packet time delay, page
Time delay, dns resolution time delay are opened in face, and TCP connection time delay sends request time delay, receives response delay etc.).
Table can be built as unit of a districts and cities, i.e., the sampled data of one prefecture-level local network is placed on (each net on a table
The data of network standard are placed on individual table), each list is stayed alone reason.It, can be with if there is no districts and cities' information in sampled data
It is determined in the way of (if so, such as LAC and districts and cities' mapping table) by LAC Deng great area's number informations tabling look-up.
Data in table are ranked up:Main sequence is arranged by the areas cell ID Zhong great number ascending order, and secondary sequence presses cell
Number ascending order arranges.
Step 3:Matter difference sample judges
Take the index that most important two perception index is evaluated as matter difference in sampled data, i.e. " first packet time delay " and " page
Open time delay in face ", correspondingly, matter difference thresholding includes " first packet time delay matter difference thresholding " and " page open time delay matter difference thresholding ".
According in " browse service perception information sampling data table " data storing order and preset matter difference thresholding, to table
In each efficiently sampling data, judge whether it is matter difference sample one by one.Because in a browse service perception sample generally simultaneously
The sampled value for including " first packet time delay " and " page open time delay ", therefore, it is necessary to judge and mark whether the sample is " first respectively
Packet delay matter difference sample " and " page open time delay matter difference sample ".
For 4 delay in phases i.e. dns resolution time delay, TCP connection time delay, the GET request time delay acquired in sampled data
With receive response delay, it is compared respectively with respective pre-set matter difference thresholding also, judges and marks the sample to be
It is no for " dns resolution time-out sample ", " TCP connection time-out sample ", " GET request time-out sample " and " reception response timeout sample
This ".
It should be noted that different network formats should define different thresholding values respectively.With telecommunications 3G network and
For 4G networks, " first packet time delay matter difference thresholding " can be respectively defined as 800ms according to user's subjective experience and network measured value
And 500ms, and " page open time delay matter difference thresholding " is then set to 3000ms and 1500ms respectively.
Step 4:Network level matter difference is analysed
This step is from 6 different dimensions (website, website IP, delay in phases, time, terminal, user) to that may cause net
The reason of network level of the browsing class service-aware difference of user, carries out analyzing and positioning in network.6 sub-steps are specifically included, that is, are walked
Rapid 4a~4f, as follows:
Step 4a:Matter difference web analytics
The targeted website that several websites are the analysis of network level matter difference is selected first (generally takes the most preceding N of family visit capacity
A website, N are predeterminable), the analysis of following steps is carried out only for selected targeted website.
To selected N number of target browsing class website, each website this month total number of samples, first packet time delay matter difference sample are counted respectively
Number, page open time delay matter difference sample number, and calculate separately first packet matter rate, page matter rate and business according to following formula
Matter rate:
First packet matter rate=first packet time delay matter difference sample number/total number of samples * 100%
(1)
Page matter rate=page open time delay matter difference sample number/total number of samples * 100%
(2)
Business matter rate=first packet matter rate * w1+ page matter rate * w2
(3)
Wherein w1 and w2 is respectively the weight of first packet matter rate and page matter rate, and the sum of w1 and w2 should be equal to 1, generally may be used
Take w1=0.6, w2=0.4.The value size of the two weight depends on the importance of index, and first packet time delay characterization is from terminal
Whether whether the transmission channel to targeted website connected and respond sufficiently fast, and page open time delay is then on the basis of first packet time delay
Index, importance is low compared with first packet time delay, in general rule of thumb can be by first packet matter rate and page matter rate
Weight is taken as 0.6 and 0.4 respectively.
If certain business matter rate to browse web sites is more than pre-determined threshold (default value 20%), which is
" matter difference website ".
Step 4b:The IP analyses of matter difference
To single website, the business matter rate of the IP whole sample points is counted as unit of the IP of the website (according to above
Formula (1)~(3) same mode), the sample number of the IP account for the ratio of the website total number of samples.It is more than pre- to business matter rate
Gating limits the IP of (default value 20%), is identified as " the matter difference IP " of the website.
Step 4c:Delay in phases is analyzed
For four composition periods of a web page browsing process page open time delay:When dns resolution time delay, TCP connection
Prolong, GET request time delay, receive response delay.The matter rate for analyzing every section of time delay finds out time delay segmentation of problems.
Piecewise analysis matter difference time delay is segmented:
(4) DNS timeout issues:If in the page open time delay matter difference sample of certain website, dns resolution time-out sample number is big
In pre-determined threshold, such as 50% (i.e. dns resolution matter Cha Shuai >50%) it, then positions the website and there is " DNS timeout issues ".And it calculates
" the average dns resolution time delay " of all page open time delay matter difference samples in the website.
(5) TCP connection timeout issue:If in the page open time delay matter difference sample of certain website, TCP connection time-out sample
Number is more than pre-determined threshold, such as 50% (i.e. TCP connection matter Cha Shuai >50%) website, is then positioned in the presence of " TCP connection time-out is asked
Topic ".And calculate " the average TCP connection time delay " of all page open time delay matter difference samples in the website.
(6) GET request timeout issue:If in the page open time delay matter difference sample of certain website, GET request time-out sample
Number is more than pre-determined threshold, such as 50% (i.e. GET request matter Cha Shuai >50%) website, is then positioned in the presence of " GET request time-out is asked
Topic ".And calculate " the average GET request time delay " of all page open time delay matter difference samples in the website.
(7) response timeout problem is received:If in the page open time delay matter difference sample of certain website, receiving response timeout sample
This number is more than pre-determined threshold, and such as 50% (receives response matter Cha Shuai >50%) it, then positions the website and there is " reception response timeout
Problem ".And calculate " the average received response delay " of all page open time delay matter difference samples in the website.
Step 4d:Piecewise analysis when matter difference
As unit of 24 hours one day each hours, total sample of the target complete website within each hour is counted respectively
Number, first packet time delay matter difference sample number, page open time delay matter difference sample number, and calculated separately according to the formula of front (1)~(3)
First packet matter rate, page matter rate and business matter rate.
If the business matter rate within some period is more than pre-determined threshold (default value 20%), which is identified as
" matter difference period ".
Step 4e:Poor-quality terminal is analyzed
Each terminal models for reaching certain predetermined thresholding (such as 1000) for the total sample size of business, count target complete respectively
Total number of samples, first packet time delay matter difference sample number, page open time delay matter difference sample number of the website under each terminal models, and press
First packet matter rate, page matter rate and business matter rate are calculated separately according to formula (1)~(3) of front.
If the business matter rate of some terminal models is more than pre-determined threshold (default value 20%), the terminal models mark
Know is " poor-quality terminal ".
Step 4f:Matter difference customer analysis
The single user of certain predetermined thresholding (such as 1000) is reached (with the IMSI or IMEI of user for the total sample size of business
Distinguish), total number of samples, first packet time delay matter difference sample number, the page of the target complete website under each user are counted respectively
Time delay matter difference sample number is opened, and calculates separately first packet matter rate, the page matter of each user according to the formula of front (1)~(3)
Rate and business matter rate.
If the business matter rate of some user is more than pre-determined threshold (default value 20%), by the IMSI of the user or
IMEI is identified as " matter difference user ".
Step 5:The matter difference origin cause of formation exports
It is exported after above-mentioned each germplasm difference origin cause of formation is summarized.It can be exported by the form of following each table:
We use the above method and the acquisition from practical mobile network (3G networks of several telecommunications companies of districts and cities)
The reason that terminal data perceives browse service matter difference is analyzed, and is found that the dns resolution time existing for certain network respectively
The problems such as excessively high problem, Sina website's first packet time delay bigger than normal, these problems have obtained the confirmation of the optimization department of telecommunications and have done phase
Network adjustment (optimization dns server, the local Sina's CDN server of increase etc.) is answered, subsequent test also demonstrates problem and obtains
Solution index gets a promotion.It is demonstrated experimentally that this method can effectively find the main reason for network service perception deterioration, for electricity
The network aware optimization of letter operator promotes work and provides important valuable reference information.
Finally it should be noted that above example only to technical scheme of the present invention is described rather than to this technology method
It is limited, the present invention can above extend to other modifications, variation, application and embodiment, and therefore, it is considered that institute in application
There are such modification, variation, application, embodiment all within the scope of the spirit or teaching of the present invention.
Claims (2)
1. a kind of browsing class service-aware analysis method, which is characterized in that include the following steps:
Step 1:Perception data acquires
Acquire user browsing class business conduct information, as " original sampling data ", including information should include:Day
Phase, Hour Minute Second, network formats, cell ID, terminal longitude and latitude are currently located province and districts and cities' title, field strength, and signal quality is used
Family identifies, and terminal iidentification, terminal models, browser title, browse web sites title, and browse web sites URL, and browse web sites IP, first packet
Time delay, page open time delay, dns resolution time delay, TCP connection time delay, GET request time delay and reception response delay;
Wherein " field strength " refers to the watt level of the currently received useful electromagnetic wave signal of user terminal;" signal quality " refers to using
The power ratio of useful signal and other destructive signals in the terminal current Received Signal of family;
Step 2:Perception data cleans
Above-mentioned original sampling data is cleaned before being stored in database, is inserted into browse service perception information hits
According to table, data item includes:On the date, Hour Minute Second, network type, cell ID, current home saves and districts and cities, terminal longitude and latitude, field
By force, signal quality, user identifier, terminal iidentification, terminal models, browser title, browse web sites title, and browse web sites URL, clear
Look at website IP, first packet time delay, page open time delay, dns resolution time delay, TCP connection time delay sends request time delay and receive and respond
Time delay;
Step 3:Matter difference sample judges
Take the most important index evaluated as matter difference of two perception indexs in sampled data, i.e. " first packet time delay " and " page is beaten
Open time delay ", correspondingly, matter difference thresholding includes " first packet time delay matter difference thresholding " and " page open time delay matter difference thresholding ";
According in browse service perception information sampling data table data storing order and preset matter difference thresholding, to respectively having in table
Sampled data is imitated, judges whether it is matter difference sample one by one;For the 4 i.e. dns resolutions of delay in phases acquired in sampled data
Time delay, TCP connection time delay, GET request time delay and reception response delay, also respectively by itself and respective pre-set matter difference thresholding
It is compared, judges and mark the sample whether be " dns resolution time-out sample ", " TCP connection time-out sample ", " GET request is super
When sample " and " receive response timeout sample ";
Step 4:Network level matter difference is analysed
This step is from 6 different dimensions, i.e. website, website IP, delay in phases, Hour Minute Second, terminal and user, to that may cause
The reason of network level of the browsing class service-aware difference of user, carries out analyzing and positioning in network;
Wherein, " website " refers to " browse web sites title ", i.e., the title for the specific website that active user is accessed;" website IP " refers to and works as
Before provide to the user specific " website " net page browse information service particular server IP address;
Step 5:The matter difference origin cause of formation exports
It is exported after above-mentioned each germplasm difference origin cause of formation is summarized.
2. according to the method described in claim 1, it is characterized in that, step 4 is specific as follows:
Step 4a:Matter difference web analytics
The targeted website that several websites are the analysis of network level matter difference is selected first,
To selected N number of target browsing class website, count respectively each website this month total number of samples, first packet time delay matter difference sample number,
Page open time delay matter difference sample number, and calculate separately first packet matter rate, page matter rate and business matter according to following formula
Rate:
First packet matter rate=first packet time delay matter difference sample number/total number of samples * 100% (1)
Page matter rate=page open time delay matter difference sample number/total number of samples * 100% (2)
Business matter rate=first packet matter rate * w1+ page matter rate * w2 (3)
Wherein w1 and w2 is respectively the weight of first packet matter rate and page matter rate,
If certain business matter rate to browse web sites is more than pre-determined threshold, which is " matter difference website ";
Step 4b:The IP analyses of matter difference
To single website, the business matter rate of the IP whole sample points, the sample number of the IP are counted as unit of the IP of the website
Account for the ratio of the website total number of samples;It is more than the IP of pre-determined threshold to business matter rate, is identified as " the matter difference IP " of the website;
Step 4c:Delay in phases is analyzed
For four composition periods of a web page browsing process page open time delay:Dns resolution time delay, TCP connection time delay,
GET request time delay receives response delay;The matter rate for analyzing every section of time delay finds out time delay segmentation of problems;
Piecewise analysis matter difference time delay is segmented:
(1) DNS timeout issues:If in the page open time delay matter difference sample of certain website, dns resolution time-out sample number is more than pre-
Gating limits, then positions the website and there is " DNS timeout issues ", and calculate all page open time delay matter difference samples in the website
" average dns resolution time delay ";
(2) TCP connection timeout issue:If in the page open time delay matter difference sample of certain website, TCP connection time-out sample number is big
In pre-determined threshold, then positions the website and there is " TCP connection timeout issue ", and it is poor to calculate all page open time delay matter in the website
" the average TCP connection time delay " of sample;
(3) GET request timeout issue:If in the page open time delay matter difference sample of certain website, GET request time-out sample number is big
In pre-determined threshold, then positions the website and there is " GET request timeout issue ", and it is poor to calculate all page open time delay matter in the website
" the average GET request time delay " of sample;
(4) response timeout problem is received:If in the page open time delay matter difference sample of certain website, receiving response timeout sample number
More than pre-determined threshold, then positions the website and there are problems that " receiving response timeout ", and calculate all page open time delays in the website
" the average received response delay " of matter difference sample;
Step 4d:Piecewise analysis when matter difference
As unit of 24 hours one day each hours, count respectively total number of samples of the target complete website within each hour,
First packet time delay matter difference sample number, page open time delay matter difference sample number, and calculate separately head according to the formula of front (1)~(3)
Packet matter rate, page matter rate and business matter rate;
If the business matter rate within some period is more than pre-determined threshold, which is identified as " matter difference period ";
Step 4e:Poor-quality terminal is analyzed
Reach each terminal models of certain predetermined thresholding for the total sample size of business, counts target complete website respectively at each end
Total number of samples, first packet time delay matter difference sample number, the page open time delay matter difference sample number under model are held, and according to the formula of front
(1)~(3) calculate separately first packet matter rate, page matter rate and business matter rate;
If the business matter rate of some terminal models is more than pre-determined threshold, which is identified as " poor-quality terminal ";
Step 4f:Matter difference customer analysis
Reach the single user of certain predetermined thresholding for the total sample size of business, counts target complete website respectively in each user
Under total number of samples, first packet time delay matter difference sample number, page open time delay matter difference sample number, and according to the formula of front (1)~
(3) the first packet matter rate, page matter rate and business matter rate of each user are calculated separately;
If the business matter rate of some user is more than pre-determined threshold, the IMSI of the user or IMEI are identified as " matter difference use
Family ";
Step 4a-4f need not be limited and be executed sequence.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610320411.8A CN105791034B (en) | 2016-05-15 | 2016-05-15 | A kind of browsing class service-aware analysis method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610320411.8A CN105791034B (en) | 2016-05-15 | 2016-05-15 | A kind of browsing class service-aware analysis method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105791034A CN105791034A (en) | 2016-07-20 |
CN105791034B true CN105791034B (en) | 2018-10-26 |
Family
ID=56378658
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610320411.8A Expired - Fee Related CN105791034B (en) | 2016-05-15 | 2016-05-15 | A kind of browsing class service-aware analysis method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105791034B (en) |
Families Citing this family (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108271176B (en) * | 2016-12-30 | 2021-05-04 | 中国移动通信集团福建有限公司 | Method and system for determining base station cell quality difference root cause |
CN107292519B (en) * | 2017-06-26 | 2020-11-03 | 北京联合大学 | Browsing service perception index prediction method based on multi-label learning |
CN107392020A (en) * | 2017-06-30 | 2017-11-24 | 北京奇虎科技有限公司 | Database manipulation analysis method, device, computing device and computer-readable storage medium |
CN109728920B (en) * | 2017-10-27 | 2020-08-21 | 贵州白山云科技股份有限公司 | Method and device for improving service quality of network service product |
CN110401551B (en) * | 2018-04-24 | 2022-05-13 | 中国移动通信集团广东有限公司 | Internet health degree evaluation method and system based on S1 interface |
CN111212440B (en) * | 2018-11-21 | 2021-09-03 | 华为技术有限公司 | Method and network equipment for realizing quality difference root cause analysis |
CN109728945B (en) * | 2018-12-25 | 2021-12-28 | 中电福富信息科技有限公司 | Abnormal threshold judgment method based on broadband key quality index |
CN111404713B (en) * | 2019-01-02 | 2022-12-13 | 中国移动通信有限公司研究院 | Network resource adjusting method, device and storage medium |
CN110138608B (en) * | 2019-05-09 | 2022-08-30 | 网宿科技股份有限公司 | Method and server for managing network service quality |
CN110535684A (en) * | 2019-07-24 | 2019-12-03 | 武汉绿色网络信息服务有限责任公司 | A kind of method and apparatus that web-browsing service perception assessment is realized based on DPI |
CN110417770B (en) * | 2019-07-25 | 2021-08-31 | 东北大学 | Wireless remote data real-time transmission method based on deterministic resource scheduling |
CN113676926B (en) * | 2020-05-15 | 2024-03-19 | 中国移动通信集团设计院有限公司 | User network sensing portrait method and device |
CN111953563A (en) * | 2020-07-31 | 2020-11-17 | 中国移动通信集团江苏有限公司 | User identification method, device, equipment and computer storage medium |
CN111988201B (en) * | 2020-09-01 | 2021-10-22 | 杭州东方通信软件技术有限公司 | Automatic identification method and device for quality difference router |
CN112202643B (en) * | 2020-10-12 | 2022-03-29 | 成都知道创宇信息技术有限公司 | Network detection method, device, test host and storage medium |
CN114650573B (en) * | 2020-12-21 | 2023-10-27 | 深圳市万普拉斯科技有限公司 | Network delay processing method and device and mobile terminal |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103607309A (en) * | 2013-11-29 | 2014-02-26 | 中国移动通信集团广东有限公司江门分公司 | Mapping method for service KQI and QOE |
CN105491167A (en) * | 2016-01-25 | 2016-04-13 | 国网重庆市电力公司客户服务中心 | Realizing method of sensing operation experience of browser terminal user in real time |
CN105550310A (en) * | 2015-12-14 | 2016-05-04 | 中国联合网络通信集团有限公司 | Method and apparatus for determining webpage resource distribution |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9742858B2 (en) * | 2011-12-23 | 2017-08-22 | Akamai Technologies Inc. | Assessment of content delivery services using performance measurements from within an end user client application |
-
2016
- 2016-05-15 CN CN201610320411.8A patent/CN105791034B/en not_active Expired - Fee Related
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103607309A (en) * | 2013-11-29 | 2014-02-26 | 中国移动通信集团广东有限公司江门分公司 | Mapping method for service KQI and QOE |
CN105550310A (en) * | 2015-12-14 | 2016-05-04 | 中国联合网络通信集团有限公司 | Method and apparatus for determining webpage resource distribution |
CN105491167A (en) * | 2016-01-25 | 2016-04-13 | 国网重庆市电力公司客户服务中心 | Realizing method of sensing operation experience of browser terminal user in real time |
Also Published As
Publication number | Publication date |
---|---|
CN105791034A (en) | 2016-07-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105791034B (en) | A kind of browsing class service-aware analysis method | |
CN103107958B (en) | A kind of method and system obtaining user's perception | |
JP5873188B2 (en) | Method, device and system for monitoring quality of internet access service of mobile terminal | |
CN105357699B (en) | Quality of wireless network monitors system and method | |
CN109391513B (en) | Network perception intelligent early warning and improving method based on big data | |
CN104052832B (en) | Method and device for setting IP address of domain name resolution server and server | |
US20120317151A1 (en) | Model-Based Method for Managing Information Derived From Network Traffic | |
CN105828376B (en) | Quality of service monitoring method and equipment based on user's perception | |
CN104994133A (en) | Mobile Web webpage access user experience perception evaluating method based on network KPI | |
CN105050125A (en) | Method and device for evaluating mobile data service quality oriented to user experience | |
US8818927B2 (en) | Method for generating rules and parameters for assessing relevance of information derived from internet traffic | |
CN110535684A (en) | A kind of method and apparatus that web-browsing service perception assessment is realized based on DPI | |
CN104113870B (en) | Network test method and mobile terminal | |
CN105306246B (en) | A kind of method, apparatus and server of the complaint of automatic-answering back device network class | |
CN103096356A (en) | Wireless network performance analysis method | |
WO2007091934A1 (en) | Performance monitoring of location-based service in a mobile telecommunications network | |
CN105072151A (en) | Content collaborative scheduling method and system for CDN | |
CN105050133A (en) | Method and device for detecting internet speed under Wi-Fi hotspot | |
CN105657738A (en) | Method, device and system for positioning problem of poor mobile phone service aware quality | |
CN106535240A (en) | Mobile APP centralized performance analysis method based on cloud platform | |
CN108965011A (en) | One kind being based on intelligent gateway deep packet inspection system and analysis method | |
CN104363130B (en) | The system and method for network quality case study location control is realized in communication network | |
CN113676926B (en) | User network sensing portrait method and device | |
Daengsi et al. | QoS Measurement: A Comparative Study of Speeds and Latency for 5G Network Using Different Speed Test Applications for Mobile Phones | |
CN114285760A (en) | Webpage access quality evaluation method and device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20181026 |
|
CF01 | Termination of patent right due to non-payment of annual fee |