TWI610581B - Smart data pricing method carried out based on sdn - Google Patents

Abstract

The present invention implements a plurality of different data fusion model methods in a software-defined network (SDN), thereby completing a smart data charging method of the present invention, which includes: an average data fusion model method (average Method data fusion model, AM model method, population weighted average method data fusion model (PWAM model method) and relational matrix data fusion model (RM model method) . The results of the simulation verification show that the smart data charging method of the present invention can accurately monitor the network traffic state of the terminal network devices at different locations (areas) and the bandwidth thereof. By applying the smart data charging method of the present invention to the network system, the Internet Provider (ISP) can appropriately adjust the network of each area according to the manner of monitoring the traffic state of the network and the bandwidth used. Road traffic and bandwidth to solve the problem of network congestion or insufficient bandwidth.

Description

Intelligent data charging method realized by software-defined network technology

The present invention belongs to the technical field of network service and billing management, and particularly relates to a network-defined network technology that can appropriately adjust network traffic and bandwidth in each area according to network traffic status and usage bandwidth. Smart data billing method.

The emergence of the Internet has completely changed the way people live and interact. Further, the emergence of cloud services and big data makes people need to use network services all the time. For this reason, effective management of network resources and provision of good network service quality by users have become very important issues. On the other hand, in view of the high cost of accessing the Internet, some people still cannot fully enjoy or use the convenience of the mobile Internet; in short, between demand and cost The contradiction has hindered the promotion and development of the mobile Internet to a certain extent. Therefore, how to control and divide users based on their network service requirements and their Internet access modes Network traffic is a very important issue.

Internet Service Providers (ISPs) usually use the Smart Data Pricing (SDP) method to set different pricing plans according to different users' network usage, so as to encourage users to adjust their The purpose of the Internet access model is to alleviate the problem of network congestion during peak hours while maintaining a certain level of network service quality. The effectiveness of the time-sharing method depends on the measurement stability and accuracy of the network bandwidth usage; however, due to the complexity of the network traffic flow, it is necessary to grasp the network usage and traffic in different regions all the time and accurately. The status is not an easy task.

In view of this, Stanford University professor Nick Mckeown proposed a new network management concept: Software Defined Network (SDN). Among them, the software-defined network technology mainly achieves the purpose of more flexible management of network behavior patterns by separating the control layer. Figure 1 shows the architecture of a software-defined network. As shown in FIG. 1, the architecture of the software definition network 1' (hereinafter referred to as SDN 1') includes an application layer 11', a control layer 12' and a data layer 13'. It is worth noting that the architecture of SDN 1' allows the user to compile a web application (application, APP) 111' in the application layer 11'; and the web application 111' can pass the northbound application interface (Northbound API) 14' takes the relevant information transmitted by the control layer 12' to perform related applications. At the same time, one or more controllers 121' in the control layer 12' will pass through the southbound application interface (Southbound The API) 15' informs the switch 131' in the data layer 13' of the actions that need to be performed.

OpenFlow is a southbound application interface that belongs to a protocol that is assigned down to the data layer 13' by the control layer 12'. This type of agreement provides a standard and open standard for communication between the controller and the switch 131'; thus, the network administrator can write or optimize various applications of the controller, thereby achieving more Functional network management mode. For example, network administrators can monitor network bandwidth usage and network traffic status through the SDN 1' architecture.

However, since the state of the Internet traffic varies according to regional differences, how to obtain global network status becomes a challenging task.

In view of this, the inventor of the case was improved and innovated, and after years of painstaking research, he finally designed and developed a software-defined network technology through the combination of the SDN 1' and data fusion. Achieved smart data billing method.

The main purpose of the present invention is to provide a smart data charging method implemented by software-defined network technology. In the present invention, a plurality of different data fusion model methods are implemented in a software-defined network (SDN), thereby completing a smart data charging method of the present invention, which includes: average data fusion Model approach (average method data fusion model, AM model method), population weighted average method data fusion model (PWAM model method) and relational matrix data fusion model (relation-matrix method data fusion model) method). Further, in order to verify the feasibility of the smart data charging method, the inventor of the present invention deploys five OpenFlow controllers in a commercial network simulator, and simultaneously corresponds to five presets of different terminal network devices. Location and deployment of 40 OpenFlow switches. The results of the simulation verification show that the intelligent data charging method of the present invention can accurately monitor the network traffic state of the terminal network devices in different locations (areas) and the bandwidth thereof. It is conceivable that an Internet provider (ISP) can appropriately adjust the network traffic status and the bandwidth according to the method of monitoring the smart data charging method of the present invention. Network traffic and bandwidth in each region to address network congestion or insufficient bandwidth usage.

In order to achieve the above-mentioned primary object of the present invention, the inventor of the present invention proposes an embodiment of the smart data charging method implemented by the software-defined network technology, which includes: a network traffic monitoring unit, which is A plurality of locations within the area monitor a plurality of network traffic flows using at least one network service to obtain one or more bytes and a total network traffic; wherein each location includes one or more switches, And the monitoring of the plurality of network traffic flows is based on The network traffic monitoring unit and the switch are implemented by an OpenFlow protocol; a data fusion unit is configured to monitor the data monitored by the network traffic monitoring unit through the switches based on at least one data fusion model. The data is executed by a data fusion process to obtain a degree of convergence of the converged network; and a pricing unit calculates a denomination rate based on the degree of congestion of the converged network obtained by the data fusion unit, and further calculates the denomination rate according to the denomination rate Calculate a payable network usage fee for each user in the area.

1'‧‧‧Software Definition Network

11'‧‧‧Application layer

111'‧‧‧Web application

12'‧‧‧Control layer

121'‧‧‧ Controller

13'‧‧‧data layer

131'‧‧‧Switch

14'‧‧‧Northbound Application Interface

15'‧‧‧Southward application interface

1‧‧‧Smart data billing method

11‧‧‧Network Traffic Monitoring Unit

111‧‧‧Open flow controller

112‧‧‧ counter

12‧‧‧Data Fusion Unit

13‧‧‧Price unit

2‧‧‧ Area

21‧‧‧ position

22‧‧‧Switch

The technical contents of the present invention and the effects of the objects of the present invention will be further understood by referring to the detailed description of the present invention and the accompanying drawings. FIG. 1 is a structural diagram showing a software-defined network; FIG. 2 is a schematic diagram of the present invention. An operational architecture diagram of a smart data billing method implemented by software-defined network technology; Figure 3 is a graph showing time versus network traffic flow; and Figure 4 is a graph showing time versus network traffic flow .

In order to more clearly describe the smart data charging method implemented by the software-defined network technology proposed by the present invention, the following will be described in detail with reference to the drawings.

Referring to FIG. 2, an operational architecture diagram of a smart data charging method implemented by a software-defined network technology according to the present invention is shown. Please refer to FIG. 1 together, and the smart data charging method 1 (hereinafter referred to as smart data charging method 1) implemented by the software-defined network technology of the present invention can be compiled into a web application as shown in FIG. 1 ( The application, APP) 111' is implemented in the infrastructure of the Software Defined Network (SDN). As shown in FIG. 2, the smart data charging method 1 of the present invention mainly includes a network traffic monitoring unit 11, a data fusion unit 12, and a pricing unit 13.

In the operational architecture of the present invention, when a user (client) in a plurality of locations 21 in a zone 2 uses at least one network service (ie, accesses Internet behavior), The network traffic monitoring unit 11 monitors a plurality of Internet traffic flows generated by the access network behavior, and then obtains one or more bytes from the network traffic flow (bytes) ) with a total network traffic (Y _{k, l, m} , total bytes). It should be noted that each location 21 will have one or more switches 22, and the monitoring of the plurality of network traffic flows is based on the open flow between the network traffic monitoring unit 11 and the switch 22 (OpenFlow). Realized by agreement. In order to explain the function of the network traffic monitoring unit 11 more clearly, the parameters that will appear in the following description are first organized in the following table (1).

As shown in FIG. 2, the network traffic monitoring unit 11 includes one or more OpenFlow controllers 111 and a counter 112. The OpenFlow controller 111 controls the OpenFlow controllers based on the OpenFlow protocol. The switch 22 obtains the one or more bytes and the total network traffic ( Y _{k , l , m} ). On the other hand, the counter 112 is used to count the total number of bytes ( a _{i , j} ) of the bytes. It should be particularly noted that the network traffic monitoring unit 11 divides a plurality of time slots ( t _{k , l , m} ) into 24 hours a day, and obtains the total network based on the following mathematical formulas (1) and (2), respectively. The total number of users ( N _{k , l , m} ) using network services in traffic ( Y _{k , l , m} ) and region 2 (or location 21):

The statistics provided by Chunghwa Telecom show that the average network traffic used by each user to access the Internet is 5.5GB; therefore, based on the setting of one month to 30 days, we can calculate each user in a single The network traffic used in the time slot is 7.641x10 ⁶ . Briefly, the present invention estimates the total number of users using network services ( N _{k , l ,} by dividing the total network traffic ( Y _{k , l , m} ) monitored by 7.641× ^{10 6} _{. m} ).

。值得說明的是，該方法可以是平均資料融合模型方法(average method data fusion nodel)；並且，透過該方法計算獲得之融合網路壅塞程度

係為一平均融合網路壅塞程度

。其中，該平均資料融合模型方法由以下數學式(3)與(4)所表示：

Continue to refer to Figure 2. In the present invention, the data fusion unit 12 is designed to perform a data fusion process on the data data monitored by the network traffic monitoring unit 11 through the switches 22 based on at least one data fusion model method, thereby obtaining a fusion network. Road congestion level

. It should be noted that the method may be an average method data fusion node (algorge method data fusion node); and the degree of convergence of the network obtained by the method is calculated.

Is an average convergence network congestion level

. The average data fusion model method is represented by the following mathematical formulas (3) and (4):

1≦ m ≦M................................................ .......(4)

The x _{k , l , m} in equation (3) is expressed as the degree of network congestion without data fusion, which can be calculated by the following mathematical formula (3-1):

；這表示監測自多個位置21的數據資料已經透過特定的資料融合模型方法完成資料融合之數據處理。 Moreover, by observing equations (3) and (3-1), it can be found that the degree of network congestion before data fusion is represented by the symbolic expression x _{k , l , m} , and the degree of network congestion after data fusion is represented by a symbol table. Formula

This means that data from multiple locations 21 has been monitored for data processing through data fusion through a specific data fusion model approach.

係為一人口加權平均融合網路壅塞程度

。其中，該人口加權平均資料融合模型方法由以下數學式(5)與(6)所表示：

In addition, the method may also be a population weighted average method data fusion model; and the degree of convergence of the network obtained by the method is calculated.

a population-weighted average converged network congestion

. Among them, the population weighted average data fusion model method is represented by the following mathematical formulas (5) and (6):

γ _k,l,m in equation (5) is expressed as population weight and can be calculated by the following mathematical formula (5-1):

之時便需要將不同位置21彼此間的網路交通狀態的關聯性一併考慮進來。為了更清楚地解釋這部分，在此先將後續說明之中會出現的參數整理於下表(2)之中。 It is worth noting that two things must be considered at the same time when designing the data fusion model method: (1) there must be a peak period of network traffic flow during the day; and (2) during the non-peak period, within the same area 2 The different locations 21 will show different network traffic conditions. Therefore, in calculating the degree of congested network congestion

At that time, it is necessary to take into account the correlation of the network traffic states of the different locations 21 with each other. In order to explain this part more clearly, the parameters that will appear in the following description are first organized in the following table (2).

As shown in FIG. 2, since a region 2 contains a plurality of locations 21, one can further obtain a distance matrix (D _k,l ) of the locations 21 based on the various parameters listed in Table (2). The distance matrix (D _k,l ) is represented by the following mathematical formula (I):

Moreover, we can further obtain a relation matrix similar to the distance matrix (D _{k, l} ). The correlation matrix (R _k,l ) is represented by the following mathematical formula (II):

即用以表示同一地區2內的任兩個位置21之間的網路交通狀態的關聯性。並且，該關聯性矩陣元素

與該距離矩 陣元素

可由以下數學式(III)所表示：

It is conceivable that the correlation matrix elements in the correlation matrix

That is, it is used to indicate the association of the network traffic state between any two locations 21 in the same area 2. And the associated matrix element

Matrix element with the distance

It can be expressed by the following formula (III):

係為一關係矩陣融合網路壅塞程度

。其中，該關係矩陣資料融合模型方法由以下數學式(7)與(8)所表示：

Through the establishment of the correlation matrix (R _k,l ), the strength of the mutual influence of the network traffic state between the two locations 21 can be estimated. Thus, based on the correlation matrix, a relational matrix data fusion model method can be established. It is worth noting that the degree of congestion of the converged network is calculated by this method.

Is a relationship matrix fusion network congestion degree

. The relation matrix data fusion model method is represented by the following mathematical formulas (7) and (8):

表示為M個維度(即，區域2中的m 個位置21)的支援比率向量。 In equations (7)-(8)

A support ratio vector expressed as M dimensions (ie, m positions 21 in region 2).

而計算出一計價率(α _k,l)，並進一步地根據該計價率計算出區域2內的每個用戶(client)的一應付網路使用費(Z^*)。為了更加清楚地解釋計價單元13之功能，在此先將後續說明之中會出現的參數整理於下表(3)之中。 Continue to refer to Figure 2. In the present invention, the pricing unit 13 is based on the degree of congestion of the converged network obtained by the data fusion unit 12.

A rate of valuation ( α _k,l ) is calculated, and a payable network usage fee (Z ^* ) for each client in the area 2 is further calculated based on the rate. In order to explain the function of the pricing unit 13 more clearly, the parameters that will appear in the subsequent description are first organized in the following table (3).

所計算而得的計價率(α _k,l)。該計價率(α _k,l)係透過以下數學式(9)與(10)而獲得：

As shown in Fig. 2, since there are a total of K time slots divided from 24 hours a day, we can further obtain the network based on data fusion in a time slot based on the parameters listed in Table (3). Road congestion level

The calculated pricing rate ( α _k,l ). The pricing rate ( α _k,l ) is obtained by the following mathematical formulas (9) and (10):

It must be added that the rate is 1 on the first day of the month. On the other hand, on other days in the month, the pricing rate ( α _k,l ) must be calculated by the above equations (9) and (10). Next, the pricing unit 13 can calculate the payable network usage fee (Z ^* ) of each user (client) based on the following mathematical formula (11): Z ^* = Σ _{k , l , m} α _{k , l} Z _{k , l , m} ............................................... (11).

It is worth noting that the payable network usage fee is calculated based on the different network traffic used by the user in different locations 21 and different time slots; therefore, the user may change its usage in order to reduce the network usage fee. The habit of Internet services, such as changing the time period for accessing the Internet. To confirm this section, the inventors using SDN-based simulator EstiNet ^TM conducted experiments. In the experimental design, referring to the architecture diagram of FIG. 2, the inventors deployed 40 switches 22 corresponding to five locations 21 of different terminal network devices. At the same time, one month is set to 30 days and the day is divided into 24 time slots (ie, L=30, K=24).

In view of the above, in order to reduce the network usage fee, the user may change the behavior mode of accessing the Internet in different locations 21 or different time slots. So in order to simulate this part, we have designed an incentive function, and the excitation function is represented by the following mathematical formulas (a) and (b):

Moreover, some parameters related to the excitation function are organized in the following table (4).

See Figure 3 for a graph showing time versus network traffic. Among them, the information about the four data curves in Figure 3 is organized in the following table (5).

Comparing curve A, curve B, curve C, and curve D, it can be found that the peak time of users accessing the Internet in Taipei City has shifted from 12-16 to 14-18; this result shows that most of the Taipei city is located. The user system is indeed motivated to change the time period during which it accesses the Internet.

Please continue to refer to Figure 4, which shows a graph of time versus network traffic. Among them, the information about the four data curves in Figure 4 is organized in the following table (6).

Comparing the curve A' with the curve B', it can be found that in the early stage of implementation of the smart data charging method of the present invention, the Keelung user first changed the behavior mode of accessing the Internet to reduce the network usage traffic. Next, comparing the curve B', the curve C' and the curve D', it can be found that the peak period of the user accessing the Internet in Keelung is shifted from 14-16 to 15-18. Therefore, the simulation experimental data of FIG. 3 and FIG. 4 confirm that the implementation of the intelligent data charging method of the present invention can accurately monitor different locations (regions). The network traffic status of the terminal network device and its usage bandwidth.

Features and effects

Based on the above detailed description of the intelligent data charging method implemented by the software-defined network technology proposed by the present invention, it is believed that engineers and telecommunication operators familiar with network communication technologies can easily find that the present invention is applied to practical applications. The following characteristics and effects are manifested: (1) Unlike the traditional smart data pricing (SDP), which is susceptible to the complexity of the network traffic flow, it is not possible to accurately grasp the different locations in time (regions). The network traffic state of the terminal network device, the smart data charging method of the present invention completes the appropriate region based on the monitored network traffic state and usage bandwidth by combining the SDN and data fusion model methods. A high-performance intelligent data charging method for adjusting network traffic and bandwidth; (2) by implementing the present invention, an Internet Provider (ISP) can not only alleviate the problem of network congestion or insufficient bandwidth usage, At the same time, it is also possible to adaptively establish a suitable pricing plan according to the network usage habits of different users.

It is to be understood that the foregoing detailed description of the preferred embodiments of the present invention , should be included in the scope of the patent in this case.

In summary, this case is not only innovative in terms of technical thinking, And with the above-mentioned multiple functions that are not in the traditional methods of the past, it has fully complied with the statutory invention patents of novelty and progressiveness. If you apply in accordance with the law, you are requested to approve the application for the invention patent, so as to invent the invention. .

1‧‧‧Smart data billing method

11‧‧‧Network Traffic Monitoring Unit

111‧‧‧Open flow controller

112‧‧‧ counter

12‧‧‧Data Fusion Unit

13‧‧‧Price unit

2‧‧‧ Area

21‧‧‧ position

22‧‧‧Switch

Claims (6)

A smart data charging method implemented by a software-defined network technology is implemented in a software-defined network architecture; the smart data charging method includes: a network traffic monitoring unit, Multiple locations within a zone monitor multiple Internet traffic flows while using at least one network service to obtain one or more bytes and a total network traffic ( Y _{k , l , m} , total bytes); wherein each location includes one or more switches, and the monitoring of the plurality of network traffic flows is based on an open flow between the network traffic monitoring unit and the switch (OpenFlow) protocol implementation; a data fusion unit, based on at least one data fusion model, performs a data fusion process on the data data monitored by the network traffic monitoring unit through the switches, thereby obtaining a converged network congestion degree And a pricing unit based on the degree of congestion of the converged network obtained by the data fusion unit Calculating a pricing rate ( α _{k , l} ), and further calculating a payable network usage fee ( Z ^* ) for each user (client) in the zone according to the pricing rate; the network The traffic monitoring unit further includes: one or more OpenFlow controllers, configured to control the switches based on the OpenFlow protocol, and obtain the one or more bytes and the total network And a counter for counting a total number of bytes ( a _{i , j} ) of the bytes; wherein the network traffic monitoring unit divides the time into a plurality of times 24 hours a day The slots ( t _{k , l , m} ) are obtained based on the following mathematical formulas (1) and (2) respectively for the total network traffic ( Y _{k , l , m} ) and a number of clients: Where: Y _{k , l , m is} expressed as the total network traffic; N _{k , l , m is} expressed as the total number of users; a _{i . j is} expressed as the total number of the bytes; V is expressed as being in any position a set of the switches controlled by the open flow controller; Z represents a set of a plurality of network traffic flows controlled by the open flow controller; i is an integer; j is an integer; k is expressed as the plurality of times Any one of the slots; l is represented as any day of the month; m is represented as any of the locations. For example, the smart data charging method implemented by the software-defined network technology described in claim 1 is an average data fusion model method; and the fusion network congestion obtained by the method is calculated. degree It is an average fusion network congestion degree; wherein the average data fusion model method is represented by the following mathematical formulas (3) and (4): 1≦ m ≦M................................................ ...........(4); Expressed as the average convergence network congestion degree; x _{k , l , m is} expressed as the degree of network congestion obtained in the kth time slot on the lth day of the month in the mth position; M is an integer . For example, the smart data charging method implemented by the software-defined network technology described in claim 1 is a population-weighted average data fusion model method; and the fusion network obtained by the method is calculated. Road congestion level It is a population-weighted average fusion network congestion degree; the population weighted average data fusion model method is represented by the following mathematical formulas (5) and (6): among them: Expressed as the population-weighted average converged network congestion degree; γ _{k , l , m is} expressed as a population weight obtained in the kth time slot on the lth day of the month in the mth position. For example, the smart data charging method implemented by the software-defined network technology described in the first application of the patent scope, wherein the method is a relational matrix data fusion model method; and the fusion network obtained by the method is calculated. Degree of congestion It is a relationship matrix fusion network congestion degree; the relation matrix data fusion model method is represented by the following mathematical formulas (7) and (8): among them: Expressed as the degree of convergence of the relationship matrix fusion network; Expressed as a support ratio vector among m positions; Expressed as a network congestion correlation among m locations; Represented as an affinity matrix for the plurality of locations within the zone. The smart data charging method implemented by the software-defined network technology described in claim 1 of the patent application, wherein the pricing rate ( α _{k , 1} ) is 1 on the first day of the month; On other days of the month, the valuation rate ( α _{k , l} ) is obtained by the following mathematical formulas (9) and (10): Wherein: L _q ratio is expressed as the minimum pricing; L _Q represents the maximum rate pricing; δ _q represents a minimum threshold; δ _Q represents a maximum threshold value; σ _{l -1} l expressed as to the first day of the month in the fusion An average of the degree of network congestion; u = ( L _q - L _Q ) / ( δ _q - δ _Q ). The smart data charging method implemented by the software-defined network technology described in claim 5, wherein the pricing unit is capable of being based on the pricing rate ( α _{k , l} ) based on the following mathematical formula (11) And calculate the payable network usage fee ( Z *): Z *=Σ _{k , l , m} α _{k , l} Z _{k , l , m} ................. .............................(11); where: Z ^{* is} expressed as the payable network usage fee; α _{k , l} Expressed as the valuation rate; Z _{k , l , m are} expressed as the network traffic flows obtained in the kth time slot on the lth day of the month in the mth position.