CN110519104B - Bandwidth multiplexing method and system based on peak error degree - Google Patents

Abstract

The invention provides a bandwidth multiplexing method and system based on peak error degree. The method equally divides the time of day into n time intervals, and new time intervals formed by adding {1,2.... w } time intervals before and after the n time intervals are recorded as incremental time intervals; calculating the concentration and peak staggering degree of the bandwidth peak value mean value of each increment time period of each user in each statistical day L; screening the time interval and the increment time interval of each user according to the continuous stability of the peak staggering degree of the same increment time interval in the continuous time interval; then matching the peak error of each increment time interval of each time interval of the user; and allocating the same CDN bandwidth nodes to the increment time interval of the successfully matched user. The method of the invention can better utilize idle bandwidth resources and improve the resource utilization rate.

Description

Bandwidth multiplexing method and system based on peak error degree

Technical Field

The invention relates to the technical field of data mining, in particular to a bandwidth multiplexing method and system based on peak error degree.

Background

In the cost of the video cloud service, the bandwidth cost is a large part, but actually, most of the current bandwidths are superposed together to perform calculation based on the highest peak value, and are not intelligently allocated, so that the resource utilization rate is not high. The phenomenon that some users regularly have high bandwidth in a certain time period for a long time, but the bandwidth in other time periods is less or even idle exists, but charging and resource allocation are still matched according to the highest peak value, the phenomenon causes resource and cost waste, manual allocation needs to be monitored and calculated at any time, the situation that the user quantity is large is unrealistic, how to better utilize idle bandwidth resources is achieved automatically in a time-saving and labor-saving mode, and the problem that needs to be solved at present is solved.

Disclosure of Invention

The embodiment of the invention aims to provide a bandwidth multiplexing method based on peak error degree, and aims to solve the problems that most of the bandwidth in the prior art is superposed together and calculation is executed on the basis of the highest peak value, intelligent allocation is not performed, and the resource utilization rate is not high.

The embodiment of the invention is realized in such a way that a bandwidth multiplexing method based on the peak error degree comprises the following steps:

dividing the time of day into n time periods, and recording the time periods as a set T ═ T₁,t₂...t_i...t_n-1,t_n}；

Recording a new time interval formed by adding {1,2.... w } time intervals before and after the n time intervals as an increment time interval; w is set by the user, w is an integer and

calculating the concentration and peak staggering degree of the bandwidth peak value mean value of each increment time period of each user in each statistical day L;

screening the time interval and the increment time interval of each user according to the continuous stability of the peak staggering degree of the same increment time interval in the continuous time interval;

matching the peak error degree of each increment time interval of each time interval of the user;

and distributing the same CDN bandwidth nodes to the increment time interval of the successfully matched user.

A second objective of an embodiment of the present invention is to provide a bandwidth multiplexing system based on a peak error degree, where the system includes:

a time division module, configured to divide the time of day into n time periods, which are recorded as a set T ═ T₁,t₂...t_i...t_n-1,t_n}；

An increment time interval determining module, configured to mark, as an increment time interval, a new time interval formed by adding {1,2.... w } time intervals before and after the n time intervals; w is set by the user, w is an integer and

the concentration and peak error calculation device is used for calculating the concentration and peak error of the bandwidth peak value mean value of each increment time interval of each user in each time interval in the statistical days L;

the user time interval and increment time interval screening device is used for screening the time interval and the increment time interval of each user according to the continuous stability of the peak error of the same increment time interval in the continuous time interval;

the incremental time interval peak error matching module is used for matching the peak error of each incremental time interval of each time interval of the user;

and the CDN bandwidth node distribution module is used for distributing the same CDN bandwidth nodes to the increment time interval of the successfully matched user.

The invention has the advantages of

The invention provides a bandwidth multiplexing method and system based on peak error degree. The method equally divides the time of day into n time intervals, and new time intervals formed by adding {1,2.... w } time intervals before and after the n time intervals are recorded as incremental time intervals; calculating the concentration and peak staggering degree of the bandwidth peak value mean value of each increment time period of each user in each statistical day L; screening the time interval and the increment time interval of each user according to the continuous stability of the peak staggering degree of the same increment time interval in the continuous time interval; then matching the peak error of each increment time interval of each time interval of the user; the same CDN bandwidth nodes are distributed to the increment time interval of the successfully matched user, idle bandwidth resources can be well utilized, and the resource utilization rate is improved.

Drawings

FIG. 1 is a flow chart of a bandwidth multiplexing method based on peak error according to a preferred embodiment of the present invention;

FIG. 2 is a flowchart of the method of S2 of FIG. 1;

FIG. 3 is a flowchart of the method of S3 of FIG. 1;

FIG. 4 is a block diagram of a preferred embodiment of the present invention for a bandwidth reuse system based on peak error;

fig. 5 is a block diagram of the concentration ratio and peak error calculation apparatus of fig. 3;

fig. 6 is a block diagram of the user period and incremental period filtering apparatus of fig. 3.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to the accompanying drawings and examples, and for convenience of description, only parts related to the examples of the present invention are shown. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

The invention provides a bandwidth multiplexing method and system based on peak error degree. The method equally divides the time of day into n time intervals, and new time intervals formed by adding {1,2.... w } time intervals before and after the n time intervals are recorded as incremental time intervals; calculating the concentration and peak staggering degree of the bandwidth peak value mean value of each increment time period of each user in each statistical day L; screening the time interval and the increment time interval of each user according to the continuous stability of the peak staggering degree of the same increment time interval in the continuous time interval; then matching the peak error of each increment time interval of each time interval of the user; and allocating the same CDN bandwidth nodes to the increment time interval of the successfully matched user. The method of the invention can better utilize idle bandwidth resources and improve the resource utilization rate.

Example one

FIG. 1 is a flow chart of a bandwidth multiplexing method based on peak error according to a preferred embodiment of the present invention; the method comprises the following steps:

s0, dividing the time of day into n periods, and recording as the set T ═ T₁,t₂...t_i...t_n-1,t_n}；

S1, recording a new time interval formed by adding {1,2.... w } time intervals before and after the n time intervals as an increment time interval; w is set by the user, w is an integer and

each incremental time period of each time period is denoted as a first data set group:

incremental time period set Z (t) of 1 st time period₁)＝{t₁z₁,t₁z₂......t₁z_w}

Incremental time period set Z (t) for time period 2₂)＝{t₂z₁,t₂z₂......t₂z_w}……

Incremental time period set Z (t) of the ith time period_i)＝{t_iz₁,t_iz₂......t_iz_w}……

Incremental period set Z (t) for nth period_n)＝{t_nz₁,t_nz₂......t_nz_w}。

Wherein, t_nz_wA w-th incremental period representing an nth period, and so on; the x-th increment period in each increment period set comprises 2x +1 periods, and x is more than or equal to 1 and less than or equal to w;

incremental time period set Z (t) of 1 st time period₁)＝{t₁z₁,t₁z₂......t₁z_wIn (c) } the reaction solution is,

1 st incremental period t₁z₁Including period t_1-1,t₁,t₁₊₁}；

Incremental period of time t of 2₁z₂Including the period t_1-2,t_1-1,t₁,t₁₊₁,t₁₊₂}；……

The x-th incremental period t₁z_xIncluding period t_1-x...t_1-2,t_1-1,t₁,t₁₊₁,t₁₊₂...t_1+x}；……

W-th incremental period t₁z_wIncluding period t_1-w...t_1-2,t_1-1,t₁,t₁₊₁,t₁₊₂...t_1+w}；

Wherein, t_1-1,t_1-2,t_1-x,t_1-wRepresents the nth, n-1, n-x +1, n-w +1 time period of the day before the current day; t is t₁₊₁,t₁₊₂,t_1+x,t_1+wRepresents the 2 nd, 3 rd, 1+ x, 1+ w period;

……

incremental time period set Z (t) of the ith time period_i)＝{t_iz₁,t_iz₂......t_iz_wIn (c) } the reaction solution is,

1 st incremental period t_iz₁Including period t_i-1,t_i,t_i+1}；

Incremental period of time t of 2_iz₂Including period t_i-2,t_i-1,t_i,t_i+1，t_i+2}；……

The x-th incremental period t_iz_xIncluding period t_i-x...t_i-2,t_i-1,t_i,t_i+1，t_i+2...t_i+x}；……

W incremental period t_iz_wIncluding period t_i-w...t_i-2,t_i-1,t_i,t_i+1，t_i+2...t_i+w}；

If i-x is less than or equal to 0, ti-x represents the nth-x +1 time period of the day before the day, and otherwise, represents the ith-x time period; if the i + x is larger than or equal to n, the ti + x represents the i + x-n time period of the day after the day, otherwise, the i + x time period is represented;

……

incremental period set Z (t) for nth period_n)＝{t_nz₁,t_nz₂......t_nz_wIn (c) } the reaction solution is,

1 st incremental period t_nz₁Including period t_n-1,t_n,t_n+1}；

Incremental period of time t of 2_nz₂Including period t_n-2,t_n-1,t_n,t_n+1，t_n+2}；……

The x-th incremental period t_nz_xIncluding period t_n-x...t_n-2,t_n-1,t_n,t_n+1,t_n+2...t_n+x}；……

W-th incremental period t_nz_wIncluding period t_n-w...t_n-2,t_n-1,t_n,t_n+1,t_n+2...t_n+w}。

Wherein, t_n-1,t_n-2,t_n-x,t_n-wRepresents the n-1 th, n-2 th, n-x th and n-w th time periods; t is t_n+1,t_n+2,t_n+x,t_n+wRepresents the 1 st, 2 nd, x, w th time period of the day after the day;

s2, calculating the concentration and peak staggering degree of the bandwidth peak value mean value of each increment time period of each user in the statistical days L;

FIG. 2 is a flowchart of the method of S2 of FIG. 1; the method comprises the following steps:

step A0: calculating the network bandwidth peak value average value of each user in the n periods within the statistical number of days;

wherein the content of the first and second substances,

representing the mean value of the network bandwidth peak value of the current time period in the ith time period, and so on; p (t)_i)₁、p(t_i)₂、p(t_i)_LRespectively representing the network bandwidth peaks at day 1, day 2 and day L of the i-th period.

Step A1: counting the network bandwidth peak value average value of each time period contained in each increment time period of each user in the counting days L;

incremental time period set Z (t) of the ith time period_i)＝{t_iz₁,t_iz₂......t_iz_wIn (c) } the reaction solution is,

x incremental period t_iz_xThe network bandwidth peak average value containing time interval is set as

Wherein if i-x is less than or equal to 0, then

The peak average value of the network bandwidth of the n-x +1 th time period of the day before the current day is represented, otherwise, the peak average value of the network bandwidth of the n-x +1 th time period of the day before the current day is represented

Representing the mean value of the network bandwidth peak values of the ith-x period; if i + x is greater than or equal to n, then

The peak value average value of the network bandwidth of the i + x-n time period of the day after the day is represented, otherwise, the average value of the network bandwidth of the i + x-n time period of the day after the day is represented

Network bandwidth peak value representing i + x time periodMean value;

step A2: calculating the relative frequency of the network bandwidth peak value mean value of each increment period contained in each user in the statistic days L;

relative frequency: the ratio of the single-period bandwidth peak value mean value to the sum of the total-period bandwidth peak value mean values, wherein the xth incremental period comprises 2x +1 periods;

in the incremental period of the i-th period,

representing a period t contained in the xth incremental period_iThe relative frequency of the mean value of the network bandwidth peak value, and so on; x is more than or equal to 1 and less than or equal to w;

incremental time period set Z (t) of the ith time period_i)＝{t_iz₁,t_iz₂......t_iz_wIn (c) } the (c) is,

1 st incremental period t_iz₁The relative frequency number set of the network bandwidth peak mean value containing the time period is F_i1＝{F₁(t_i-1),F₁(t_i),F₁(t_i+1)}；

Incremental period of time t of 2_iz₂The relative frequency number set of the network bandwidth peak mean value containing the time period is F_i2＝{F₂(t_i-2),F₂(t_i-1),F₂(t_i),F₂(t_i+1),F₂(t_i+2)}；……

The x-th incremental period t_iz_xThe relative frequency number set of the network bandwidth peak mean value containing the time period is F_ix＝{F_x(t_i-x),...F_x(t_i-2),F_x(t_i-1),F_x(t_i),F_x(t_i+1),F_x(t_i+2),...F_x(t_i+x)}；……

W-th incremental period t_iz_wThe relative frequency number set of the network bandwidth peak mean value containing the time period is F_iw＝{F_w(t_i-w),...F_w(t_i-2),F_w(t_i-1),F_w(t_i),F_w(t_i+1),F_w(t_i+2),...F_w(t_i+w)}；

Wherein if i-x is less than or equal to 0, F_x(t_i-x) A relative frequency representing a mean of network bandwidth peaks for an nth-x +1 th time period of the day preceding the day contained in the xth incremental time period, otherwise F_x(t_i-x) A relative frequency representing the mean of the network bandwidth peaks for time period i-x contained in the xth incremental time period, and so on;

if i + x is greater than or equal to n, F_x(t_i+x) A relative frequency representing a mean of network bandwidth peaks for an i + x-n time period of a day subsequent to the day contained in the x-th incremental time period, otherwise F_x(t_i+x) The relative frequency of the network bandwidth peak means of the time period i + x contained in the xth incremental time period is represented, and so on;

step A3: calculating the mean value and the standard deviation of data in the relative frequency set of the network bandwidth peak mean value of each increment time period containing time period;

of the incremental periods of the ith period, the xth incremental period contains a set of relative frequencies F of the mean of the bandwidth peaks of the period_ixThe mean and standard deviation of (a) are:

step A4: calculating the concentration and peak staggering degree of the bandwidth peak value mean value of each user;

of the increment periods of the i-th period, the x-th increment period t_iz_xThe concentration and the peak error of the bandwidth peak value mean value containing the time interval are respectively as follows:

concentration ratio C_ixComprises the following steps:

degree of peak error S_ixComprises the following steps:

concentration is an index describing the concentration (i.e., steep or flat) of data distribution, and a positive concentration value means a narrower pattern, larger data, and flat data. The closer to 0, the closer to the positive too distribution. The peak error degree is an index for describing symmetry, and is negative when the data inclines to the right; if skewed to the left, the degree of peak error is positive.

Optionally, step S2 may be followed by the steps of: screening and filtering the users according to the peak staggering degree and the concentration;

the method specifically comprises the following steps: selecting users with retention concentration ratio of [ -1,1] and peak error ratio >0.3, and deleting other users;

s3, screening the time interval and the increment time interval of each user according to the continuous stability of the peak-to-peak error of the same increment time interval in the continuous time interval;

FIG. 3 is a flowchart of the method of S3 of FIG. 1; the method comprises the following steps:

step B1, taking the peak shift degrees of q time periods before and after each time period (total 2q +1 time periods) from the n time periods as:

H＝{S_(i-q)x,...S_(i-2)x,S_(i-1)x,S_ix,S_(i+1)x,S_(i+2)x,...S_(i+q)x}；

step B2, calculating the continuous stability K of the peak error degree of the 2q +1 time intervals_μ；

Wherein the content of the first and second substances,

step B3, K meeting the threshold condition_μTaking the time interval i and the increment time interval x of the corresponding user as candidates;

wherein, the threshold condition is as follows:

s4: matching the peak-to-peak error of each increment time interval of each time interval of the user;

the method specifically comprises the following steps: performing pairwise comparison matching on all users according to a peak-error matching rule, and selecting the best matching user;

false kurtosis matching rule (rule 1 and rule 2 are satisfied simultaneously):

rule 1: peak error degree S of x increment time interval of i time intervals of two users_ix(u_k) And S_ix(u_j) The signs are opposite;

rule 2: selecting min | S_ix(u_k)₊S_ix(u_j) Matching users of | is carried out;

wherein S is_ix(u_k)、S_ix(u_j) Respectively represent users u_k、u_jA staggering of the bandwidth peak means for an x incremental period of an ith period; j is more than or equal to 1 and less than or equal to m;

if there are at least 2 identical min | S_ix(u_k)₊S_ix(u_j) If the concentration ratio is the same in sign (positive or negative), the concentration ratio is closest to min | C |)_ix(u_k)-C_ix(u_j) Matching is performed by the users of l. Wherein C is_ix(u_k)、C_ix(u_j) Representing user u_k、u_jThe concentration of the bandwidth peak means at the xth incremental period of the ith period.

Further, step S4 may be followed by: matching the magnitude of the bandwidth peak value average value of the user;

the method comprises the following specific steps: performing pairwise comparison matching on all users according to the magnitude matching rule of the bandwidth peak value mean value, and selecting the best matching user;

magnitude matching rule of bandwidth peak value mean value:

and is

Wherein G is_ix(u_k)、G_ix(u_j) Respectively represent the u-th user_k、u_jJ is more than or equal to 1 and less than or equal to m in a data set of a time interval network bandwidth peak value mean value contained in the xth increment time interval of the ith time interval in the statistical days L; max (variable) represents taking the maximum value of the variable; average (variable) represents averaging the variables.

S5: allocating the same CDN bandwidth nodes to the increment time interval of the successfully matched user;

the method specifically comprises the following steps:

suppose user u_k、u_jIs matched with the x increment period of the ith period, then the user u is presented_k、u_jThe x-th increment period t_iz_xIncluding a period t_i-x...t_i-2,t_i-1,t_i,t_i+1，t_i+2...t_i+xAllocate the same CDN bandwidth node.

The method for allocating bandwidth nodes adopts a conventional method known in the art, and is not described herein again.

Example two

FIG. 4 is a block diagram of a preferred embodiment of the present invention for a bandwidth reuse system based on peak error; the system comprises:

a time division module, configured to divide the time of day into n time periods, which are recorded as a set T ═ T₁,t₂...t_i...t_n-1,t_n}；

An increment time interval determining module, configured to mark, as an increment time interval, a new time interval formed by adding {1,2.... w } time intervals to the n time intervals; w is used bySet at home, w is an integer and

each incremental time period of each time period is denoted as a first data set group:

incremental time period set Z (t) of 1 st time period₁)＝{t₁z₁,t₁z₂......t₁z_w}

Incremental time period set Z (t) for time period 2₂)＝{t₂z₁,t₂z₂......t₂z_w}……

Incremental time period set Z (t) of the ith time period_i)＝{t_iz₁,t_iz₂......t_iz_w}……

Incremental period set Z (t) for nth period_n)＝{t_nz₁,t_nz₂......t_nz_w}。

Wherein, t_nz_wA w-th incremental period representing an nth period, and so on; the x-th increment period in each increment period set comprises 2x +1 periods, and x is more than or equal to 1 and less than or equal to w;

incremental time period set Z (t) of 1 st time period₁)＝{t₁z₁,t₁z₂......t₁z_wIn (c) } the reaction solution is,

1 st incremental period t₁z₁Including period t_1-1,t₁,t₁₊₁}；

Incremental period of time t of 2₁z₂Including period t_1-2,t_1-1,t₁,t₁₊₁,t₁₊₂}；……

The x-th incremental period t₁z_xIncluding period t_1-x...t_1-2,t_1-1,t₁,t₁₊₁,t₁₊₂...t_1+x}；……

W-th incremental period t₁z_wIncluding period t_1-w...t_1-2,t_1-1,t₁,t₁₊₁,t₁₊₂...t_1+w}；

Wherein, t_1-1,t_1-2,t_1-x,t_1-wRepresents the nth, n-1, n-x +1, n-w +1 time period of the day before the current day; t is t₁₊₁,t₁₊₂,t_1+x,t_1+wRepresents the 2 nd, 3 rd, 1+ x, 1+ w period of the day after the day;

……

incremental time period set Z (t) of the ith time period_i)＝{t_iz₁,t_iz₂......t_iz_wIn (c) } the reaction solution is,

1 st incremental period t_iz₁Including period t_i-1,t_i,t_i+1}；

Incremental period of time t of 2_iz₂Including period t_i-2,t_i-1,t_i,t_i+1，t_i+2}；……

The x-th incremental period t_iz_xIncluding period t_i-x...t_i-2,t_i-1,t_i,t_i+1，t_i+2...t_i+x}；……

W-th incremental period t_iz_wIncluding period t_i-w...t_i-2,t_i-1,t_i,t_i+1，t_i+2...t_i+w}；

Wherein if i-x is less than or equal to 0, t is_i-xRepresents the nth-x +1 time period of the day before the current day, and otherwise represents the ith-x time period; if i + x is greater than or equal to n, t_i+xRepresents the i + x-n time period of the day after the day, otherwise represents the i + x time period;

……

incremental period set Z (t) for nth period_n)＝{t_nz₁,t_nz₂......t_nz_wIn (c) } the reaction solution is,

1 st incremental period t_nz₁Including period t_n-1,t_n,t_n+1}；

Incremental period of time t of 2_nz₂Including period t_n-2,t_i-1,t_n,t_n+1，t_n+2}；……

The x-th incremental period t_nz_xIncluding period t_n-x...t_n-2,t_n-1,t_n,t_n+1,t_n+2...t_n+x}；……

W incremental period t_nz_wIncluding period t_n-w...t_n-2,t_n-1,t_n,t_n+1,t_n+2...t_n+w}。

Wherein, t_n-1,t_n-2,t_n-x,t_n-wRepresents the n-1 th, n-2 th, n-x th and n-w th time periods; t is t_n+1,t_n+2,t_n+x,t_n+wRepresents the 1 st, 2 nd, x, w th time period of the day after the day;

the concentration and peak error calculation device is used for calculating the concentration and peak error of the bandwidth peak value mean value of each increment time interval of each user in each time interval in the statistical days L;

the user time interval and increment time interval screening device is used for screening the time interval and the increment time interval of each user according to the continuous stability of the peak error of the same increment time interval in the continuous time interval;

the incremental time interval peak error matching module is used for matching the peak error of each incremental time interval of each time interval of the user;

and the CDN bandwidth node distribution module is used for distributing the same CDN bandwidth nodes to the increment time interval of the successfully matched user.

Further, the bandwidth multiplexing system based on peak error degree further comprises a concentration and peak error degree screening module (not shown in the figure), which is connected with the concentration and peak error degree calculating device and is used for selecting users with retention concentration degree of [ -1,1] and peak error degree >0.3 and deleting other users;

further, the bandwidth multiplexing system based on the peak error degree further includes a bandwidth peak mean value magnitude matching module (not shown in the figure), connected to the incremental time interval peak error degree matching module, and configured to match the magnitude of the bandwidth peak mean value of the user;

further, fig. 5 is a structural diagram of the concentration ratio and peak error calculation apparatus in fig. 3; the concentration and peak error calculation device includes:

the time interval bandwidth peak value average value calculation module is used for calculating the network bandwidth peak value average value of each user in the n time intervals within the statistical days;

wherein the content of the first and second substances,

representing the mean value of the network bandwidth peak value of the current time period in the ith time period, and so on; p (t)_i)₁、p(t_i)₂、p(t_i)_LRespectively representing the network bandwidth peaks at day 1, day 2 and day L of the i-th period.

The incremental period bandwidth peak value average value counting module is used for counting the network bandwidth peak value average value of each period contained in each incremental period within counting days L of each user;

incremental time period set Z (t) of the ith time period_i)＝{t_iz₁,t_iz₂......t_iz_wIn (c) } the reaction solution is,

the x-th incremental period t_iz_xThe network bandwidth peak average value containing time interval is set as

Wherein if i-x is less than or equal to 0, then

The peak average value of the network bandwidth of the n-x +1 th time period of the day before the current day is represented, otherwise, the peak average value of the network bandwidth of the n-x +1 th time period of the day before the current day is represented

Representing the mean value of the network bandwidth peak values of the ith-x period; if i + x is greater than or equal to n, then

The peak value average value of the network bandwidth of the i + x-n time period of the day after the day is represented, otherwise, the average value of the network bandwidth of the i + x-n time period of the day after the day is represented

Representing the mean value of the network bandwidth peak values in the i + x time period;

the relative frequency calculation module is used for calculating the relative frequency of the network bandwidth peak value mean value of each increment time period contained in each user in the counting days L;

relative frequency: the ratio of the single-period bandwidth peak value mean value to the sum of the total-period bandwidth peak value mean values, wherein the xth incremental period comprises 2x +1 periods;

in the incremental period of the i-th period,

representing a period t contained in the xth incremental period_iThe relative frequency of the mean value of the network bandwidth peak value, and so on; x is more than or equal to 1 and less than or equal to w;

incremental time period set Z (t) of ith time period_i)＝{t_iz₁,t_iz₂......t_iz_wIn (c) } the reaction solution is,

1 st incremental period t_iz₁The relative frequency number set of the network bandwidth peak mean value containing the time period is F_i1＝{F₁(t_i-1),F₁(t_i),F₁(t_i+1)}；

2 nd increment period t_iz₂The relative frequency number set of the network bandwidth peak mean value containing the time period is F_i2＝{F₂(t_i-2),F₂(t_i-1),F₂(t_i),F₂(t_i+1),F₂(t_i+2)}；……

The x-th incremental period t_iz_xThe relative frequency number set of the network bandwidth peak mean value containing the time period is F_ix＝{F_x(t_i-x),...F_x(t_i-2),F_x(t_i-1),F_x(t_i),F_x(t_i+1),F_x(t_i+2),...F_x(t_i+x)}；……

W-th incremental period t_iz_wThe relative frequency number set of the network bandwidth peak mean value containing the time period is F_iw＝{F_w(t_i-w),...F_w(t_i-2),F_w(t_i-1),F_w(t_i),F_w(t_i+1),F_w(t_i+2),...F_w(t_i+w)}；

Wherein if i-x is less than or equal to 0, F_x(t_i-x) A relative frequency representing a mean of network bandwidth peaks for an nth-x +1 th time period of the day preceding the day contained in the xth incremental time period, otherwise F_x(t_i-x) The relative frequency of the network bandwidth peak means of the time period i-x contained in the xth incremental time period is represented, and so on;

if i + x is greater than or equal to n, F_x(t_i+x) A relative frequency representing a mean of network bandwidth peaks for an i + x-n time period of a day subsequent to the day contained in the x-th incremental time period, otherwise F_x(t_i+x) The relative frequency of the network bandwidth peak means of the time period i + x contained in the xth incremental time period is represented, and so on;

the relative frequency mean and standard deviation calculation module is used for calculating the mean and standard deviation of data in the relative frequency set of the network bandwidth peak mean value of each increment period containing period;

of the incremental periods of the ith period, the xth incremental period contains a set of relative frequencies F of the mean of the bandwidth peaks of the period_ixThe mean and standard deviation of (a) are:

the concentration and peak error calculation module is used for calculating the concentration and peak error of the bandwidth peak value mean value of each user;

of the increment periods of the i-th period, the x-th increment period t_iz_xThe concentration and the peak error of the bandwidth peak value mean value containing the time interval are respectively as follows:

concentration ratio C_ixComprises the following steps:

degree of peak error S_ixComprises the following steps:

further, fig. 6 is a block diagram of the user time period and incremental time period filtering apparatus in fig. 3. The user time interval and increment time interval screening device comprises:

a continuous time interval extraction module, configured to, on the premise that an incremental time interval number x (an xth incremental time interval) is the same, extract, from the n time intervals, peak error degrees of q time intervals before and after each time interval (total 2q +1 time intervals) as:

S＝{S_(i-q)x,...S_(i-2)x,S_(i-1)x,S_ix,S_(i+1)x,S_(i+2)x,...S_(i+q)x}；

a peak error degree continuous stability calculation module for calculating the continuous stability K of the peak error degree of the 2q +1 time periods_μ；

Wherein the content of the first and second substances,

a user time interval and increment time interval screening module for screening the K meeting the threshold condition_μTaking the time interval i and the increment time interval x of the corresponding user as candidates;

wherein, the threshold condition is as follows:

further, in the incremental period peak-offset matching module, matching the peak-offset of each incremental period of each period of the user specifically comprises:

performing pairwise comparison matching on all users according to a peak-error matching rule, and selecting the best matching user;

false kurtosis matching rule (rule 1 and rule 2 are satisfied simultaneously):

rule 1: peak error S of x increment time interval of i time intervals of two users_ix(u_k) And S_ix(u_j) The signs are opposite;

rule 2: selecting min | S_ix(u_k)+S_ix(u_j) Matching users of | is carried out;

wherein S is_ix(u_k)、S_ix(u_j) Respectively represent users u_k、u_jThe peak-to-peak error of the bandwidth peak average of the xth incremental period of the ith period; j is more than or equal to 1 and less than or equal to m;

if there are at least 2 identical min | S_ix(u_k)+S_ix(u_j) If the concentration ratio is the same in sign (positive or negative), the concentration ratio is closest to min | C_ix(u_k)-C_ix(u_j) Matching is performed by the users of l. Wherein C is_ix(u_k)、C_ix(u_j) Representing user u_k、u_jThe concentration of the bandwidth peak means at the xth incremental period of the ith period.

Further, in the concentration and peak shift screening module, the screening and filtering of the user according to the peak shift and the concentration specifically includes: selecting users with retention concentration ratio of [ -1,1] and peak error ratio >0.3, and deleting other users;

further, in the bandwidth peak average magnitude matching module, matching the magnitude of the bandwidth peak average of the user specifically includes: performing pairwise comparison matching on all users according to the magnitude matching rule of the bandwidth peak value mean value, and selecting the best matching user;

magnitude matching rule of bandwidth peak value mean value:

and is

Wherein G_ix(u_k)、G_ix(u_j) Respectively represent the u-th user_k、u_jJ is more than or equal to 1 and less than or equal to m in a data set of a time interval network bandwidth peak value mean value contained in the xth increment time interval of the ith time interval in the statistical days L; max (variable) represents taking the maximum value of the variable; average (variable) represents averaging the variables.

Further, in the CDN bandwidth node allocation module, allocating the same CDN bandwidth node to an increment period of the period of time when the user is successfully matched;

the method specifically comprises the following steps:

suppose user u_k、u_jIs matched with the xth incremental time period of the ith time period, then the user u is presented_k、u_jThe x-th increment period t_iz_xIncluding a period t_i-x...t_i-2,t_i-1,t_i,t_i+1，t_i+2...t_i+xAllocate the same CDN bandwidth node.

It will be understood by those skilled in the art that all or part of the steps in the method according to the above embodiments may be implemented by hardware related to program instructions, and the program may be stored in a computer readable storage medium, such as ROM, RAM, magnetic disk, optical disk, etc.

The above description is intended to be illustrative of the preferred embodiment of the present invention and should not be taken as limiting the invention, but rather, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the invention.

Claims (18)

1. A bandwidth multiplexing method based on peak error degree is characterized by comprising the following steps:

dividing the time of day into n time periods, and recording the time periods as a set T ═ T₁,t₂...t_i...t_n-1,t_n}；

Recording a new time interval formed by adding {1,2.... w } time intervals before and after the n time intervals as an increment time interval; w is set by the user, w is an integer and

calculating the concentration and peak error of the bandwidth peak value mean value of each increment time interval of each user in each time interval in the statistical number of days L, wherein the method comprises the following steps:

step A0: calculating the network bandwidth peak value average value of each user in the n periods within the statistical number of days;

wherein the content of the first and second substances,

representing the mean value of the network bandwidth peak value of the current time period in the ith time period, and so on; p (t)_i)₁、p(t_i)₂、p(t_i)_LRespectively representing the network bandwidth peak values of the 1 st day, the 2 nd day and the L th day in the ith period;

step A1: counting the network bandwidth peak value average value of each time period contained in each increment time period of each user in the counting days L;

incremental time period set Z (t) of the ith time period_i)＝{t_iz₁,t_iz₂......t_iz_wIn (f), the xth incremental period t_iz_xThe peak average value set of the network bandwidth containing time period is

Wherein if i-x is less than or equal to 0, then

The network bandwidth peak value average value of the n-x +1 time period of the previous day of the day is represented, otherwise

Representing the mean value of the network bandwidth peak values of the ith-x period; if i + x is greater than or equal to n, then

The peak value average value of the network bandwidth of the i + x-n time period of the day after the day is represented, otherwise, the average value of the network bandwidth of the i + x-n time period of the day after the day is represented

Representing the mean value of the network bandwidth peak values in the i + x time period;

step A2: calculating the relative frequency of the network bandwidth peak value mean value of each increment period contained in each user in the statistic days L;

relative frequency: the ratio of the single-period bandwidth peak value mean value to the sum of the total-period bandwidth peak value mean values, wherein the xth incremental period comprises 2x +1 periods;

in the incremental period of the i-th period,

representing a period t contained in the xth incremental period_iThe relative frequency of the mean value of the network bandwidth peak value is more than or equal to 1 and less than or equal to w;

incremental time period set Z (t) of the ith time period_i)＝{t_iz₁,t_iz₂......t_iz_wIn (c) } the reaction solution is,

1 st incremental period t_iz₁The relative frequency number set of the network bandwidth peak mean value containing the time period is F_i1＝{F₁(t_i-1),F₁(t_i),F₁(t_i+1)}；

Incremental period of time t of 2_iz₂The relative frequency number set of the network bandwidth peak mean value containing the time period is F_i2＝{F₂(t_i-2),F₂(t_i-1),F₂(t_i),F₂(t_i+1),F₂(t_i+2)}；

The x-th incremental period t_iz_xThe relative frequency set of the network bandwidth peak value mean value containing the time period is

F_ix＝{F_x(t_i-x),...F_x(t_i-2),F_x(t_i-1),F_x(t_i),F_x(t_i+1),F_x(t_i+2),...F_x(t_i+x)}；

Wherein if i-x is less than or equal to 0, F_x(t_i-x) A relative frequency representing a mean of network bandwidth peaks for an nth-x +1 th time period of the day preceding the day contained in the xth incremental time period, otherwise F_x(t_i-x) A relative frequency representing a mean of network bandwidth peaks for time periods i-x contained in the xth incremental time period;

if i + x is greater than or equal to n, F_x(t_i+x) A relative frequency representing a mean of network bandwidth peaks for an i + x-n time period of a day subsequent to the day contained in the x-th incremental time period, otherwise F_x(t_i+x) A relative frequency representing a mean of network bandwidth peaks for a period i + x contained in the xth incremental period;

step A3: calculating the mean value and the standard deviation of data in the relative frequency set of the network bandwidth peak mean value of each increment time period containing time period;

of the incremental periods of the ith period, the xth incremental period contains a band of periodsRelative frequency set F of wide peak mean_ixThe mean and standard deviation of (a) are:

；

step A4: calculating the concentration and peak staggering degree of the bandwidth peak value mean value of each user;

of the increment periods of the i-th period, the x-th increment period t_iz_xThe concentration and the peak error of the bandwidth peak value mean value containing the time interval are respectively as follows:

concentration ratio C_ixComprises the following steps:

degree of peak error S_ixComprises the following steps:

；

screening the time interval and the increment time interval of each user according to the continuous stability of the peak staggering degree of the same increment time interval in the continuous time interval;

matching the peak error degree of each increment time interval of each time interval of the user;

and distributing the same CDN bandwidth nodes to the increment time interval of the successfully matched user.

2. The method for band width multiplexing based on the degree of peaking of claim 1,

the method for calculating the concentration and the peak error of the bandwidth peak value mean value of each increment time interval of each user in the statistical days L further comprises the following steps: and screening and filtering the users according to the peak staggering degree and the concentration.

3. The method for bandwidth multiplexing based on degree of peak error of claim 1,

after the matching of the peak-to-peak error of each incremental time interval of each time interval of the user, the method further comprises the following steps: and matching the magnitude of the bandwidth peak value average of the user.

4. The method for bandwidth multiplexing based on degree of peak error of claim 1,

incremental time period set Z (t) of the ith time period_i)＝{t_iz₁,t_iz₂......t_iz_wIn (c) } the reaction solution is,

the x-th incremental period t_iz_xIncluding period t_i-x...t_i-2,t_i-1,t_i,t_i+1，t_i+2...t_i+x}；

Wherein if i-x is less than or equal to 0, then t_i-xRepresents the nth-x +1 time period of the day before the current day, and otherwise represents the ith-x time period; if i + x is larger than or equal to n, then t_i+xRepresents the i + x-n time period of the day after the day, otherwise represents the i + x time period; x is more than or equal to 1 and less than or equal to w.

5. The method for bandwidth multiplexing based on degree of peak error of claim 2,

the screening and filtering of the user according to the peak load and the peak load are specifically as follows:

and selecting users with retention concentration ratio of (-1, 1) and peak error ratio >0.3, and deleting other users.

6. The method of claim 3, wherein the bandwidth multiplexing based on the degree of peak error,

the step of matching the magnitude of the bandwidth peak value mean value of the user specifically comprises: performing pairwise comparison matching on all users according to the magnitude matching rule of the bandwidth peak value mean value, and selecting the best matching user;

magnitude matching rule of bandwidth peak value mean value:

and is

Wherein G_ix(u_k)、G_ix(u_j) Respectively represent the u-th user_k、u_jJ is more than or equal to 1 and less than or equal to m in a data set of a time interval network bandwidth peak value mean value contained in the xth increment time interval of the ith time interval in the statistical days L; max (variable) represents taking the maximum value of the variable; average (variable) represents averaging the variables.

7. The method as claimed in claim 6, wherein the screening the time interval and the incremental time interval of each user according to the continuous stability of the peak-to-peak error of the same incremental time interval in the continuous time interval comprises:

step B1, under the premise that the incremental time period number x is the same, taking out the peak error degrees of q time periods before and after each time period from the n time periods as:

H＝{S_(i-q)x,...S_{(i-2)x,S(i-1)x,Six,S(i+1)x,S(i+2)x},...S_(i+q)x}；

step B2, calculating the continuous stability K of the peak error degree of the 2q +1 time intervals_μ；

Wherein the content of the first and second substances,

step B3, K meeting the threshold condition_μTaking the time interval i and the increment time interval x of the corresponding user as candidates;

the threshold conditions are:

8. the method for multiplexing bandwidths based on the degree of peak error as claimed in claim 7 wherein the degree of peak error of each incremental time interval of each time interval of the user is matched;

the method specifically comprises the following steps: performing pairwise comparison matching on all users according to a peak-error matching rule, and selecting the best matching user;

matching the rule with the degree of peak error, and simultaneously satisfying the rule 1 and the rule 2:

rule 1: peak error degree S of x increment time interval of i time intervals of two users_ix(u_k) And S_ix(u_j) The signs are opposite;

rule 2: selecting min | S_ix(u_k)+S_ix(u_j) Matching users of | is carried out;

wherein S is_ix(u_k)、S_ix(u_j) Respectively represent users u_k、u_jA staggering of the bandwidth peak means for an x incremental period of an ith period; j is more than or equal to 1 and less than or equal to m;

if there are at least 2 identical min | S_ix(u_k)+S_ix(u_j) If the concentration ratio is the same in sign (positive or negative), the concentration ratio is closest to min | C_ix(u_k)-C_ix(u_j) Matching users of | is carried out; wherein C is_ix(u_k)、C_ix(u_j) Representing user u_k、u_jThe concentration of the bandwidth peak means at the xth incremental period of the ith period.

9. The bandwidth multiplexing method based on the degree of peak error as claimed in claim 8, wherein the allocating the same CDN bandwidth node to the incremental time period of the time period matching the successful user is specifically:

suppose user u_k、u_jIs matched with the x increment period of the ith period, then the user u is presented_k、u_jThe x-th incremental period t_iz_xIncluding a period t_i-x...t_i-2,t_i-1,t_i,t_i+1，t_i+2...t_i+xAllocate the same CDN bandwidth node.

10. A system for bandwidth multiplexing based on degrees of peak error, the system comprising:

a time division module for dividing the time of day into n time periods, and recording as a set T ═ T₁,t₂...t_i...t_n-1,t_n}；

An increment time interval determining module, configured to mark, as an increment time interval, a new time interval formed by adding {1,2.... w } time intervals before and after the n time intervals; w is set by the user, w is an integer and

concentration and peak error calculation device for calculating the concentration and peak error of the bandwidth peak value mean value of each increment time interval of each user in each time interval in the statistical days L, comprising:

the time interval bandwidth peak value average value calculating module is used for calculating the network bandwidth peak value average value of each user in the n time intervals within the statistical days;

wherein, the first and the second end of the pipe are connected with each other,

representing the mean value of the network bandwidth peak value of the current time period in the ith time period, and so on; p (t)_i)₁、p(t_i)₂、p(t_i)_LRespectively representing the network bandwidth peak values of the 1 st day, the 2 nd day and the L th day in the ith period;

the incremental time interval bandwidth peak value average value counting module is used for counting the network bandwidth peak value average value of each time interval contained in each incremental time interval within the counting days L of each user;

incremental time period set Z (t) of ith time period_i)＝{t_iz₁,t_iz₂......t_iz_wIn (f), the xth incremental period t_iz_xThe network bandwidth peak average value containing time interval is set as

Wherein if i-x is less than or equal to 0, then

The peak average value of the network bandwidth of the n-x +1 th time period of the day before the current day is represented, otherwise, the peak average value of the network bandwidth of the n-x +1 th time period of the day before the current day is represented

Representing the mean value of the network bandwidth peak values of the ith-x period; if i + x is greater than or equal to n, then

The network bandwidth peak value average value of the i + x-n time period of the day after the day is represented, otherwise, the network bandwidth peak value average value of the i + x-n time period of the day after the day is represented

Representing the mean value of the network bandwidth peak values in the i + x time period;

the relative frequency calculation module is used for calculating the relative frequency of the network bandwidth peak value mean value of each increment period contained in each user in the counting days L;

relative frequency: the ratio of the single-period bandwidth peak value mean value to the sum of the total-period bandwidth peak value mean values, wherein the xth incremental period comprises 2x +1 periods;

in the incremental period of the i-th period,

representing a period t contained in the xth incremental period_iThe relative frequency of the mean value of the network bandwidth peak value is more than or equal to 1 and less than or equal to w;

incremental time period set Z (t) of the ith time period_i)＝{t_iz₁,t_iz₂......t_iz_wIn (c) } the (c) is,

1 st incremental period t_iz₁The relative frequency number set of the network bandwidth peak mean value containing the time period is F_i1＝{F₁(t_i-1),F₁(t_i),F₁(t_i+1)}；

Incremental period of time t of 2_iz₂The relative frequency number set of the network bandwidth peak mean value containing the time period is F_i2＝{F₂(t_i-2),F₂(t_i-1),F₂(t_i),F₂(t_i+1),F₂(t_i+2)}；

The x-th incremental period t_iz_xThe relative frequency number set of the network bandwidth peak mean value containing the time period is

F_ix＝{F_x(t_i-x),...F_x(t_i-2),F_x(t_i-1),F_x(t_i),F_x(t_i+1),F_x(t_i+2),...F_x(t_i+x)}；

Wherein if i-x is less than or equal to 0, F_x(t_i-x) A relative frequency representing a mean of network bandwidth peaks for an nth-x +1 th time period of the day preceding the day contained in the xth incremental time period, otherwise F_x(t_i-x) A relative frequency representing a mean of network bandwidth peaks for time periods i-x contained in the xth incremental time period;

if i + x is greater than or equal to n, F_x(t_i+x) A relative frequency representing a network bandwidth peak-to-average value of i + x-n periods of the day following the day contained in the x-th incremental period, otherwise F_x(t_i+x) Network bandwidth peak mean value representing a period i + x contained in the xth incremental periodRelative frequency of (d);

the relative frequency mean and standard deviation calculation module is used for calculating the mean and standard deviation of data in the relative frequency set of the network bandwidth peak mean value of each increment period containing period;

of the incremental periods of the ith period, the xth incremental period contains a set of relative frequencies F of the mean of the bandwidth peaks of the period_ixThe mean and standard deviation of (a) are:

；

the concentration and peak error calculation module is used for calculating the concentration and peak error of the bandwidth peak value mean value of each user;

of the increment periods of the i-th period, the x-th increment period t_iz_xThe concentration and the peak error of the bandwidth peak value mean value containing the time interval are respectively as follows:

concentration ratio C_ixComprises the following steps:

degree of peak error S_ixComprises the following steps:

；

the user time interval and increment time interval screening device is used for screening the time interval and the increment time interval of each user according to the continuous stability of the peak error of the same increment time interval in the continuous time interval;

the incremental time interval peak error matching module is used for matching the peak error of each incremental time interval of each time interval of the user;

and the CDN bandwidth node distribution module is used for distributing the same CDN bandwidth nodes to the increment time interval of the successfully matched user.

11. The kurtosis-based bandwidth multiplexing system of claim 10,

the bandwidth multiplexing system based on the peak error degree also comprises a concentration degree and peak error degree screening module which is connected with the concentration degree and peak error degree calculating device and is used for selecting users with the concentration degree of [ -1,1] and the peak error degree of >0.3 to be reserved and deleting other users.

12. The system of claim 10, wherein the system further comprises a bandwidth peak-to-average magnitude matching module, coupled to the incremental period peak-to-average matching module, for matching magnitudes of the bandwidth peak-to-average magnitudes of the users.

13. The kurtosis-based bandwidth multiplexing system of claim 10,

incremental time period set Z (t) of the ith time period_i)＝{t_iz₁,t_iz₂......t_iz_wIn (c) } the reaction solution is,

the x-th incremental period t_iz_xIncluding period t_i-x...t_i-2,t_i-1,t_i,t_i+1，t_i+2...t_i+x}；

Wherein if i-x is less than or equal to 0, t is_i-xRepresents the nth-x +1 time period of the day before the current day, and otherwise represents the ith-x time period; if i + x is greater than or equal to n, t_i+xThe time interval of the (i + x-n) th day after the current day is represented, otherwise, the time interval of the (i + x) th day is represented; x is more than or equal to 1 and less than or equal to w.

14. The kurtosis-based bandwidth multiplexing system of claim 10,

the user time interval and increment time interval screening device comprises:

a continuous time interval extraction module, configured to extract peak error degrees of q time intervals before and after each time interval from the n time intervals on the premise that the incremental time interval sequence number x is the same as:

H＝{S_(i-q)x,...S_{(i-2)x,S(i-1)x,Six,S(i+1)x,S(i+2)x},...S_(i+q)x}；

a peak error degree continuous stability calculation module for calculating the continuous stability K of the peak error degree of the 2q +1 time periods_μ；

Wherein the content of the first and second substances,

a user time interval and increment time interval screening module for screening K meeting the threshold condition_μTaking the time interval i and the increment time interval x of the corresponding user as candidates;

the threshold conditions are:

15. the kurtosis-based bandwidth multiplexing system of claim 14,

in the incremental period peak error matching module, the step of matching the peak error of each incremental period of each period of the user specifically comprises:

performing pairwise comparison matching on all users according to a peak-error matching rule, and selecting the best matching user; matching the rule with the degree of peak error, and simultaneously satisfying the rule 1 and the rule 2:

rule 1: peak error degree S of x increment time interval of i time intervals of two users_ix(u_k) And S_ix(u_j) The signs are opposite;

rule 2: selecting min | S_ix(u_k)+S_ix(u_j) Matching users of | is carried out;

wherein S is_ix(u_k)、S_ix(u_j) Respectively represent users u_k、u_jA staggering of the bandwidth peak means for an x incremental period of an ith period; j is more than or equal to 1 and less than or equal to m;

if there are at least 2 identical min | S_ix(u_k)+S_ix(u_j) If the concentration ratio is the same in sign (positive or negative), the concentration ratio is closest to min | C_ix(u_k)-C_ix(u_j) Matching users of | is carried out; wherein C is_ix(u_k)、C_ix(u_j) Representing user u_k、u_jThe concentration of the bandwidth peak means at the xth incremental period of the ith period.

16. The kurtosis-based bandwidth multiplexing system of claim 11,

in the concentration and peak shift screening module, the screening and filtering of the user according to the peak shift and the concentration specifically comprises the following steps: and selecting users with the reserved concentration ratio of [ -1,1] and the error peak ratio of >0.3, and deleting other users.

17. The kurtosis-based bandwidth multiplexing system of claim 12,

in the bandwidth peak average magnitude matching module, matching the magnitude of the bandwidth peak average of the user specifically comprises: performing pairwise comparison matching on all users according to the magnitude matching rule of the bandwidth peak value mean value, and selecting the best matching user;

magnitude matching rule of bandwidth peak value mean value:

and is provided with

Wherein G is_ix(u_k)、G_ix(u_j) Respectively representUser u_k、u_jJ is more than or equal to 1 and less than or equal to m in a data set of a time interval network bandwidth peak value mean value contained in the xth increment time interval of the ith time interval in the statistical days L; max (variable) represents taking the maximum value of the variable; average (variable) represents averaging the variables.

18. The skewness-based bandwidth multiplexing system of claim 17, wherein,

in the CDN bandwidth node allocation module, allocating the same CDN bandwidth node to the increment period of the successfully matched user specifically is:

suppose user u_k、u_jIs matched with the x increment period of the ith period, then the user u is presented_k、u_jThe x-th incremental period t_iz_xIncluding a period t_i-x...t_i-2,t_i-1,t_i,t_i+1，t_i+2...t_i+xAllocate the same CDN bandwidth node.

Images