• <tr id="yyy80"></tr>
  • <sup id="yyy80"></sup>
  • <tfoot id="yyy80"><noscript id="yyy80"></noscript></tfoot>
  • 99热精品在线国产_美女午夜性视频免费_国产精品国产高清国产av_av欧美777_自拍偷自拍亚洲精品老妇_亚洲熟女精品中文字幕_www日本黄色视频网_国产精品野战在线观看 ?

    Machine Learning-based Stable P2P IPTV Overlay

    2022-08-23 02:19:58MuhammadJavidIqbalIhsanUllahMuhammadAliAtiqAhmedWaheedNoorandAbdulBasit
    Computers Materials&Continua 2022年6期

    Muhammad Javid Iqbal,Ihsan Ullah,Muhammad AliAtiq AhmedWaheed Noor and Abdul Basit

    1Directorate of Information Technology,Sardar Bahadur Khan Women’s University,Quetta,Pakistan

    2Department of Computer Science and Information Technology,University of Balochistan,Quetta,Pakistan

    Abstract: Live video streaming is one of the newly emerged services over the Internet that has attracted immense interest of the service providers.Since Internet was not designed for such services during its inception,such a service poses some serious challenges including cost and scalability.Peer-to-Peer(P2P) Internet Protocol Television (IPTV) is an application-level distributed paradigm to offer live video contents.In terms of ease of deployment,it has emerged as a serious alternative to client server, Content Delivery Network(CDN)and IP multicast solutions.Nevertheless,P2P approach has struggled to provide the desired streaming quality due to a number of issues.Stability of peers in a network is one of the major issues among these.Most of the existing approaches address this issue through older-stable principle.This paper first extensively investigates the older-stable principle to observe its validity in different scenarios.It is observed that the older-stable principle does not hold in several of them.Then,it utilizes machine learning approach to predict the stability of peers.This work evaluates the accuracy of several machine learning algorithms over the prediction of stability, where the Gradient Boosting Regressor(GBR)out-performs other algorithms.Finally,this work presents a proof-of-concept simulation to compare the effectiveness of older-stable rule and machine learning-based predictions for the stabilization of the overlay.The results indicate that machine learning-based stability estimation significantly improves the system.

    Keywords: P2P IPTV; live video streaming; user behavior; overlay networks;stable peers;machine learning

    1 Introduction

    Live video streaming is a challenging service because it has stringent playback deadlines.Centralized client/server, IP multicast and peer-to-peer (P2P) are major architectures that enable video streaming service.Client/server is a classical approach which allows clients to generate video requests to servers and servers are able to respond to these requests.One major issue of these systems is scalability which requires to enhance the resources at server side to accommodate a surge in user requests,resulting in increased cost.A similar model named as CDN[1]brings some improvements to the classical client/server model to offer video streaming.In CDNs, the source pushes contents to various servers that are distributed purposefully over the Internet.Clients get their contents from nearby servers via any-casting,instead of contacting to source server,which ultimately cuts down the startup delay and reduces the traffic on the network.IP multicast works at the network layer that enables routers to maintain and manage groups.In this approach,one stream can be replicated by a router towards multiple receivers.This approach is an efficient solution but several issues[2]curtailed its deployment over the global scale.However, it is used in smaller domains that are managed by a single administration.On the other hand,P2P strategy works at the application layer[3].It organizes peers into a virtual network where they not only receive but also transmit the contents.The major advantage of P2P systems over client-server and IP multicast is their easier deployment with low cost.

    P2P approach for live streaming has its own issues.One of the major issues of these systems is their reliance on intermittently available nodes which turn the whole system unstable.Numerous approaches have been proposed to overcome this issue, resulting into several systems which can be classified into multiple categories.Such approaches either focus on resilient designs or consider the peer’s stability for improvements.The former approach focuses on the overlay network structure and content dissemination strategies which too without stability are not enough as the management overhead is increased as well as the delay.

    Concerning stability of peers in existing approaches, it is mostly determined through statistical observation which considers a node to be stable if it is present in the network since long[4–6].We term this approach as the older-stable principle.This principle is based on general statistical correlation of elapsed time and remaining time of all participating peers in the system.However,users have different choices and habits which influence their viewing patterns.One of the major issues with this approach is the measurement of correlation for sessions of all users together.As a result,strong correlation between elapsed time and remaining time has been observed that assumes long lived peers as stable ones [4].However, it has been observed that there are different behavior patterns of users during watching television.For instance, a user may leave his/her system on a channel while he/she is busy doing something else, while another user may be joining and leaving channels frequently during browsing mode[7].This work assesses extensively the older-stable principle through finding the correlation of elapsed time and remaining time in different scenarios.It separates different modes of behaviors and analyzes the correlation in each mode.The results show that this relationship remains different in different scenarios.

    Stability of peers can also be predicted through machine learning approach;however,we find very few such efforts in the literature.Therefore,this study also considers machine learning approach for the prediction of stability.To extensively evaluate the accuracy of machine learning algorithms,this work trains and tests several machine learning models.The results show that Gradient Boosting Regressor performs better than other algorithms.Furthermore, to get a comparison of older-stable principle and machine learning based predictions, this work presents a proof-of-concept simulation, and the results indicate that incorporating machine learning-based predictions can significantly improve the performance.

    Since P2P networks are networks of users,stability of peers plays a crucial role in the performance of these system.Knowing about stable peers,enables to construct a stream delivery structure in order to improve quality of streaming and provide better user experience.Several types of P2P IPTV overlays exist which can greatly benefit from an accurate estimation of stability.Therefore,this work provides a base line for machine learning based intelligent P2P IPTV overlays.

    The contribution highlights of this work are:

    ■Analysis of older-stable principle through finding correlation between elapsed time and remaining time of a session in different scenarios,

    ■Assessment of several machine learning algorithms for the prediction of stability and

    ■Simulation to analyze the effectiveness of machine learning-based stability estimation in a treebased system.

    In the rest of the article,we discuss related work in Section 2.Discussion on analysis of older-stable principle through estimation of correlation and obtained results are presented in Section 3.Section 4 evaluates the efficiency of machine learning algorithms for the prediction of stability.Simulations to compare integration of older-stable principle and machine learning based predictions into a tree-based overlay are given in Section 5.Finally,Sections 6 and 7 present conclusion and future work.

    2 Related Work

    Tang et al.[4]analyze logs of user behavior and observe a positive correlation between the elapsed duration of the current user session and the remaining duration.This motivates them to propose the construction and maintenance of a tree.The tree is built and maintained in such a way that peers with longer presence in the current session are placed near the root of the tree while nodes with shorter presence are placed as leaves.Several other approaches are based on the same older-stable principle.For instance, Wang et al.[5] propose a hybrid push-pull overlay that creates a tree-based backbone called tree-bone from stable nodes and the unstable nodes are connected to the tree-bone.Stream is pushed on the tree-bone while other nodes,out of the tree-bone,function in a mesh network to pull the content.Here, the selection of stable nodes for tree-bone is based on the older-stable principle.Similarly,Tan et al.[6]combine both stability and bandwidth to construct a stable tree where,peer’s stability is determined from the online age of the peer.

    Tian et al.[7]perform a measurement study of an existing system and observe that nodes’lifetime follows a Pareto distribution.In their parent selection strategy, they propose to choose older node as a parent node because it is more likely to be a stable one.A two-tiered overlay is presented in [8]where,tier1 consists of stable nodes in the form of a labeled tree.The identification of stable nodes for tier 1 is based on peers’ages which is like older stable principle.Budhkar et al.[9]propose an overlay management strategy to improve stability of the system.A similar strategy has been used in hybrid P2P-CDN overlay[10].Fuzzy Logic based Hybrid Overlay(FLHyO)is proposed in[11]that utilizes the fuzzy system to find out the priority of all the peers during overlay construction by considering upload bandwidth, geographical location, age and utilization of a peer.Ma et al.[12] analyze a P2P-CDN based system and propose a peer selection strategy involving machine learning instead of localization.Miguel et al.[13]propose classification of peers based on video chunk contribution.To create partnerships,they restrict each peer to choose partners from high contributing groups as well as low contributing groups to keep the system running.Nonetheless,they do not consider stability of peers in their approach.

    On the other hand, prediction of user’s current session is proposed in [14] which models the influencing parameters of user behavior through a Bayesian network to predict the length of the current session.Based on the predicted session, they propose to maintain a tree structure that minimizes the impact of abrupt departure of an upstream peer.According to this approach, a peer estimates its own session’s length at arrival to the network and schedules a maintenance just before the end of the session.During this stage,it swaps its position with its most stable child while other child nodes are notified to connect to the same or any other stable node.An improvement over the prediction accuracy of the Bayesian network through Support Vector Regressor(SVR)model is presented in[15].

    Most of the existing works utilize older-stable principle which considers a peer’s age as stability indicator.Furthermore, very few approaches use machine learning to determine the stability which too,do not explore potentially efficient and accurate algorithms.Therefore,this work first extensively evaluates the older-stable principle and then assesses several machine learning algorithms for their accuracy which have not been studied before for this problem.

    3 Analysis of Older Stable Principle

    To analyze the older stable mechanism,real traces of user activities are required.This study utilizes IPTV traces from Lancaster University [16] where an IPTV service provider (Lancaster living lab)offers video streams both for television sets and PCs installed with P2P clients.These traces contain users’logs collected over a period of 6 months duration.

    3.1 Data Preparation

    Data contain start and stop time of a channel by a user which represent join and departure time.This information can be used to prepare the dataset based on elapsed and remaining time of user where the whole session time is split in one-minute intervals.This provides the data to analyze the correlation between elapsed and remaining time of a user over each minute.An initial analysis of the traces revealed that these traces either contained a large portion of very short sessions or very long sessions spreading over multiple days.Therefore,it is crucial to also analyze the impact of these extreme sessions on overall correlation.To do so, this work calculates the correlation on the whole dataset as well as it separates such sessions from others and analyze them separately.Next,a discussion of different groups of sub-datasets is presented which are separated for analysis.

    3.1.1 Surfing Users

    Users exhibit different behaviors while watching programs.A user is considered to be in surfing mode when he/she is browsing channels and switches channels quickly.In this mode,user checks the current programs of different channels to choose one for watching.Usually,in this mode,a user may hold a channel till to two minutes[17].

    3.1.2 Viewing Users

    A user in viewing mode actually watches the channel and is interested in the content being transmitted.Sessions whose durations remain between the limits of surfing users and idle users are considered in the viewing mode[18].Additionally,this work also analyzes the correlation over different times of a day as it is evident from the user behavior measurements that behavior changes over daytime.

    3.1.3 Idle Users

    The presence of very long sessions(days)in dataset demonstrates a very strange phenomenon from which we conclude that these sessions might have resulted from the fact that the television set remained on while the user was not actually watching the channel.Therefore,this study analyzes these sessions separately to understand their correlation.In this work,sessions longer than 4 h are considered to be of an idle user.

    This study calculates the Pearson coefficient of correlation as given in Eq.(1).Here,x is the elapsed time whileyis the remaining time.andare the mean of elapsed time and remaining time respectively.

    3.2 Results

    This sub-section presents the analysis of older-stable principle through finding correlation between elapsed time and remaining time for different behavior groups, as discussed earlier.In addition to existing works,this work studies the impact of daytime on the correlation as well as the impact of each group on the rest of the traces.The Pearson coefficient of correlation, r = 0.71, at p = 0.05 that is statistically significant positive correlation between elapsed time and remaining time for all the traces,ascertaining previous results as presented in[4].Similarly,for these same combined traces,correlation values for each hour are shown in Fig.1a.

    Figure 1:Correlation of elapsed and remaining time.(a)Group-wise correlation(b)Impact of a group on overall correlation

    Here,the correlation remains strong for most part of the day however,it remains quite low during early hours of the day.This shows a little deviation from the existing work when sessions are split in different times.Concerning the analysis of traces from different user groups, traces belonging to surfing behavior show a correlation coefficient of r=0.07 at p=0.05 which means that elapsed time has no significant correlation with remaining time.Furthermore, concerning the hourly values, the figure shows again that no significant correlation is found at any hour of the day for surfing behavior.

    The value of correlation for viewing users is r=0.08 at p=0.05 which is similar to the surfing group.Furthermore, hourly results do not show any strong correlation.On the other hand, the correlation of traces for idle users is r=0.5 at p=0.05 which shows that a correlation between elapsed time and remaining time exists.However,it is evident from the hourly result given in the figure,that there is positive as well as negative correlation at different hours of the day.This indeed is an interesting observation which contradicts the older-stable principle.To dig it further,this work combines different behavior groups and presents the results in Fig.1b.This figure analyzes the effect of each group on overall result.It can be seen that excluding surfing behavior traces has no effect whatsoever on overall result.Moreover,idle behavior has the most impact on the overall result.Therefore,the lack of idle sessions in the traces reduces the value of correlation coefficient.

    To understand variations in correlation coefficient during different times of the day, this work analyzes the sessions generated at different times.Fig.2 presents the histograms of frequency distributions of all the combined sessions at four different timings of the day.It is evident from Fig.2 that the number of sessions at 04:00,05:00,and 11:00 are low whereas at 23:00 this number is a lot higher.It means that the number of sessions has also an impact over the correlation coefficient value.

    Figure 2:Sessions’frequency distributions in different day times.(a)Number of sessions at 04:00(b)Number of sessions at 05:00(c)Number of sessions at 11:00(d)Number of sessions at 23:00

    Since the older-stable principle is applied to individual users while it is deduced from combined behaviors in existing works, this work also considers the individual case.To perform an analysis of individual behaviors this study extracts traces of two users with most appearances in the system.Figs.3a and 3b presents the correlations of overall sessions for each of these two users over different times of the day.These results show both very weak positive and negative correlations which means that the general observation of all users together cannot be localized for individual users.

    Figure 3:Correlations for individual users.(a)User 1(b)User 2

    To conclude the analysis of older-stable principle, this study observes through extensive investigation that this principle does not always hold.Furthermore,it is dependent on time of the day and overall population of users.The older-stable rule remains intact under the analysis of all the traces of all users together which was established in similar analyses.This study observed that the strong correlation is mainly the result of idle sessions which may be present in a dataset under investigation.

    4 Contextual Approach

    In contrast to older-stable principle, the contextual approach estimates the length of current session of a user from previous sessions as well as contextual information.This section presents an extensive study of the contextual approach to evaluate the efficiency of several machine learning models over the estimation of stability.First, it discusses the machine learning models used in this work and then it presents the methodology adapted in this approach as well as the results.

    4.1 Predictive Models for User Behavior

    Our goal is to estimate the current session of a user which is a continuous variable therefore it becomes a regression type problem.In the literature, only a Bayesian Network (BN) and Support Vector Regressor (SVR) have been used for the prediction of current session from contextual information[15].This work includes other machine learning models to get a comparison and enhance the accuracy further.Since,this is a Regression problem,this study applies several Regression models and selects those which produce comparable results.The newer models chosen for this study are KNearest Neighbors Regressor (KNNR), Random Forest Regressor (RFR) and Gradient Boosting Regressor(GBR).This work also includes the previously used Bayesian network and SVR models for comparison.First,we discuss all these models.

    4.1.1 K-Nearest Neighbor Regression

    K Nearest Neighbors Regression is an adapted form of K-Nearest Neighbor(KNN)for continuous variables.KNN is a supervised learning mechanism for the prediction of discrete variables.The core of this technique is the use of nearest neighbors for the estimation of unknown or missing variable.To determine the nearest neighbors,it uses the distance metric such as Euclidean distance from missing point.KNNR estimates the output label by taking an average of the labels of its k-nearest neighbors.

    This work uses the Python machine learning library to build the KNNR model.This library provides KNNR that can be configured with different settings according to the nature of the problem.Different parameters are set for this model to get best settings for this problem.A few major parameters and their settings as chosen for this model are described below.

    · n neighbors(K):

    Kis the number of nearest points called neighbors that the algorithm considers estimating a missing point.The value ofKhas direct impact on the query processing time of a few algorithms such as Ball tree and the KD tree.A larger value ofKincreases the query time due to two reasons: Firstly, a larger value ofKrequires to search a bigger part of the neighbors’space,leading to more time.Secondly,usingK >1 requires queuing of internal results during tree traversal.A value ofK=9 produces better results in this work.

    · algorithm:There are different algorithms available that include brute,ball tree,kd tree and auto values.One can set either of these algorithms or otherwise the auto option let the system choose the best one for the problem.This work chooses the latter option.

    · P:P specifies the method of computing distance between points.A value of 1 invokes manhattan distance,2 provides a method equivalent to Euclidean distance and any other random value sets the distance metric as minkowski distance.

    4.1.2 Random Forest Regressor

    Random Forest Regressor combines multiple decision tree classifiers by fitting them on subsamples of the dataset selected randomly.Further,it averages the results of these classifiers to control over-fitting and enhance accuracy.This work utilizes the model provided by the Python machine learning library[19]to use RFR.This model provides a number of configurable parameters that can be fine-tuned according to the problem.The selected parameters are presented next.

    · n estimators: This parameter specifies the number of trees to be used by RFR.This work evaluates the performance of the model with different values and chooses a value of 50 for this parameter as it provides better results.

    · criterion:RFR uses this parameter for the measurement of quality of a split.This study selects Mean Squared Error(MSE)as a value of this parameter.

    · max depth:This parameter specifies the maximum size of the tree.This work does not impose any such limitation and uses the default value ofNonefor this parameter.

    · bootstrap:This parameter is a boolean type which if false,the model uses the entire dataset to construct each tree.This study uses the valuetruefor this parameter.

    · max samples:In case the bootstrap value is true,this parameter can be used to limit the size of the sub-samples.This study uses the default value which isNone.

    4.1.3 Gradient Boosting Regressor

    Gradient Boosting is based on AdaBoosting which starts to train a weak learner such as a decision tree through assigning an equal weight to each observation.Once the first tree is evaluated,it changes the weights of observations in such a way that difficult to classify observations get increased weights and easy to classify observations get reduced weights.In the next iteration another tree is grown with new weighted observations.The goal is to improve the predictions of the first tree.At the end of the second iteration,the model is an ensemble of the two trees.This process is continued until a specified number of iterations is achieved.

    Gradient Boosting works in similar way by training several models gradually, additively, and sequentially.Gradient Boosting differs from AdaBoost in the identification of limitations of weak learners.In contrast to AdaBoost model,gradient boosting identifies the limitations by using gradients in the loss function such as squared error.This is actually a measure of goodness of fitting model’s coefficients to the data.Since this work attempts to predict the session of a user, the loss function would base on the error of the prediction between actual and predicted session.

    4.1.4 Support Vector Regressor

    Support Vector Regressor (SVR) is a generalization of Support Vector Machine (SVM) for regression problems.SVM models the binary classification problem as convex optimization problem through finding the maximum margins that separate the hyperplane.Support vectors represent the hyperplane.To generalize SVM to SVR, anε-insensitive region around the function is introduced.This region is termed asε-tube.This tube attempts to solve the optimization problem for continuous variable through finding a tube that best approximates the continuous variable as well as balancing the complexity of the model and prediction error.In brief,firstly,it finds the flattest tube that contains most of the training instances.Secondly,it solves the convex optimization problem,having a unique solution.The training samples lying outside the tube boundary are contained in support vectors as hyperplane.These are the most influential instances impacting the shape of the tube[20].

    This study utilizes the SVR model provided by the Python machine learning library.This model provides several parameters that can be set to fine-tune the model for customized solutions.To describe the configuration of the model used in this work,its parameters are given below.

    · Kernel:The Python machine learning library provides several kernel functions such as Linear,Polynomial,Sigmoid,and Radial Basis.Since the Kernel function performs the optimization,its selection is not straightforward.Therefore,this work analyzed the performance of each kernel empirically on the dataset and chose the Radial Basis Function(RBF)kernel provided by[19].

    · Gamma:The gamma parameter determines the influence of the spread of training sample.This influence is far for small values while close for large values.This work experiments with different values of this parameter and chooses 0.1 as its value in the developed model.

    · Epsilon(ε):Parameterεis used to control the width of the epsilon-tube.This value influences the number of support vectors which are used to construct the regression function.A greater value ofεmeans fewer support vectors.Nonetheless,greater values mean more flat estimates.This work experiments with several values and chooses 0.1 as a suitable one.

    ·C: This is called the regularization parameter which manages the tradeoff between achieving a low training error and a low testing error.It generalizes a model to unseen data.Again,this work chooses through experimentation the value of this parameter which is 100.

    4.1.5 Bayesian Network Model

    The joint distribution of a set of variables having possible mutual causal relationships is represented by a Bayesian network through a directed acyclic graph and conditional probability distributions.The acyclic graph consists of nodes and edges.Nodes represent variables that contain conditional probability distributions while edges represent their causal relationships.The goal of a Beyesian network is to estimate a posterior conditional probability distribution of a target variable after observing the evidence.Each variable in a Bayesian network is conditionally independent from all its child nodes given the states of its parents.This property sometimes significantly reduces the number of parameters that are required to characterize the Joint Probability Distribution (JPD) of variables.For a given Bayesian network,the joint probability distribution overUcan be computed as shown in Eq.(2).

    Building a Bayesian network model has two parts,namely structure and parameters.The structure of the model places variables in relationships while parameters are set as prior probabilities of each node.Both of them can be built manually or learned from data[21,22].This work builds the structure manually as well as learns it from data;however,no significant difference in overall results of the two models is observed.Concerning the parameters,they are learned from data.

    4.2 Methodology

    The methodology this work adapts for contextual approach is depicted in Fig.4.It starts with the generation of the dataset and ends by the prediction of session duration.The main stages are data generation and preparation into dataset,model building and validation.We discuss these major parts one by one.

    Figure 4:Contextual approach for stability estimation

    4.2.1 Data Generation and Preparation

    The dataset used earlier in correlation analysis is limited in size therefore,this work uses a synthetic dataset.This dataset is generated through a semi-Morkovian model presented in [23].The total number of instances generated from this model are 462,825.The dataset contains the following major features.

    · Join Time Tj:This field shows the time-of-the-day at which a user joins a channel.The time is represented in minutes and its value is from 1 to 1440.

    · Departure Time Td:This feature shows again time-of-the-day at which a user quits a channel.The value of departure time has the same range as of join time.

    · Session Duration: The time duration in minutes between the join time and departure time is session duration in our dataset.This variable shows the stability of a user and models are tested for accuracy over the estimation of this variable.

    · Arrival Rate: Arrival rate or joining rate represents the number of users joining during a particular minute of the day.This field contains value of arrival rates for each minute of the day.

    · Departure Rate: Departure rate represents the number of users leaving/quitting a channel during a specific minute of a day.Departure rates for each minute is recorded in this field.

    · Population: Just like arrival and departure rates, population is also a collective variable that contains the count of online users at a specific minute of the day.

    · Content Type:The model contains three generic kinds of contents which are fiction,sports and real.The dataset records the type of the content being broadcast at each minute.

    In feature extraction phase, this work selects variables of interest in light of the work presented in [18].It calculates the session duration variable from the join time and departure time and selects the variables shown in Fig.5.Here,session duration is the target variable while other variables have influencing relationships to it.

    Figure 5:Session duration and related variables

    The dataset is split into training set and test set(validation set)using holdout method through the train test split function of Python library[19].The random state parameter is set to 5,which ensures the same random number sequence generation at each run.The parameter test size is set as 0.3,which randomly splits 30%instances of the actual dataset into test set and remaining in the training set.

    4.2.2 Model Building and Validation

    This work configures each machine learning model according to the specifications given in the description of models above.Then,it builds every model using the training subset.To test(validate)the model,this work uses the testing set instances.Here,session duration(target variable)is dropped and is let to be predicted by the model for its accuracy assessment during the validation step.

    4.3 Experiments and Results

    As discussed above each of these models is assessed over the prediction of the session duration.To evaluate the accuracy of each model,this work uses three error calculation techniques namely,Mean Absolute Error(MAE),Mean Squared Error(MSE)and Root Mean Squared Error(RMSE).

    ·MAE:It is an average of absolute differences between the actual session and the predicted session.The MAE is a linear score which means that all the individual differences are weighted equally in the average.It can be computed through Eq.(3).

    where,ndenotes the number of records in the data set,yis the actual session,andis the predicted session by the model.Furthermore,yiandare predicted,and actual values respectively forithsession.

    ·MSE:It calculates the square difference between the model predicted session and the actual session for each point,and then averages these values.MSE can be computed through Eq.(4).

    ·RMSE:RMSE is calculated in two stages.First,normal mean error is taken.Then,the square root is added to allow the errors ratio to be the same as the targets number.The RMSE can be computed as in Eq.(5).

    To compare the results of newly tested models(GBR,KNNR and RFR)to previously used models(BN and SVR),the MAE,MSE and RMSE of all these models are depicted in Fig.6.It can be easily noted from the figure that GBR, KNNR and RFR produce lower errors than SVR and BN, hence showing more accuracy.

    Concerning the MAE,the RFR model produces the lowest error which is 0.83.It means that RFR performs better for outliers.However,being an absolute error,this metric is not considered as a good choice to compare models.Furthermore,the MAE does not assign weights to errors such as,a higher weight to larger errors.Therefore,to compare the accuracy of the models,this study focusses on MSE and RMSE.It can be noted that GBR produces the least MSE and RMSE among all the tested models.Therefore,based on these two metrics,GBR emerges as the best model for this problem.

    Figure 6:Models’accuracy

    To have a look at individual predictions,a scatter plot of actualvs.predicted values for each of GBR,KNNR,RFR,SVR and BN is shown in Fig.7.In each plot,the actual sessions are shown on the x-axis and the predicted sessions on the y-axis.In case of an ideally perfect model,the predicted session should be exactly the actual session, therefore the plot will make the slope.Hence, all the over-estimations shall appear above the slope and the underestimations shall appear below the slope.Similarly,more points near the slope with low spread show overall low error.

    Figure 7: Continued

    Figure 7: Actual vs. predicted session durations (in minutes).(a) GBR (b) KNR (c) RFR (d) SVR(e)BN

    Comparing these plots,affirms the overall performance of each model discussed earlier.Here,all the three newly proposed models show better accuracy then the previous models.Among the three,GBR produces the least spread of points from the slope and therefore shows better accuracy than all the compared models.

    5 Stabilization of P2P Overlays

    Predictions of machine learning models discussed above can be used to stabilize the P2P overlay.Since there are several stream distribution strategies in P2P IPTV networks,which drives the overlay structure therefore stabilization strategies can be in multiple ways.For instance, a tree can be constructed in such a way that the stable peers are placed near the root while the unstable ones are joined at the leaves.Furthermore, algorithms to maintain a stable tree during the frequent arrivals and departures shall be required.Similarly,in hybrid system,a stable peer may be chosen for the push operation in order to minimize the pull operations.

    This work presents a proof-of-concept simulation which considers a single tree system.The data are taken from the logs of one day of the same 10,000 users,used as a dataset in evaluation of machine learning models in three cases of tree construction.In the first case,a tree is built randomly in which peers arrive and leave the system according to their actual join time and departure time.It does not involve any stabilization mechanism.This is termedrandomin this work.The second case involves the older-stable principle which constructs and maintains the tree based on the elapsed time of the user.It means that the peer present since longer is considered the stable one and is placed higher in the tree.Finally,we construct the tree based on the predicted sessions of the machine learning model.The results of GBR are utilized in this case as GBR was the most accurate model among all tested models.

    The simulation starts at minute 1 and simulates the on and off time of each user as given in the dataset.On each minute it checks the join time of each user and if it matches with the current time,it lets the user join an existing peer as a child.Each peer has an outdegree between 1 and 5 which is randomly assigned.Similarly, the departure time of each user is checked on each minute and if it matches the current time, the peer exits.At the same time the number of descendant peers of the departing peer is recorded which shows the impact of departure.The tree is periodically stabilized after every 3 min time according to applied stabilization strategy.

    As discussed earlier that intermittent presence of peers is a major issue in P2P IPTV systems,this experiment analyzes the impact of an abrupt departure.The sudden departure of an upstream peer discontinues the stream to all its descendant peers.Therefore,on each departure,this number is counted to show the resilience of a particular tree against such departures.

    The simulation is run for each case and the number of impacted peers at every minute is record.Fig.8a depicts the number of impacted peers in each case.To note the variations in number of online users, the population is also shown in Fig.8b.It is obvious from Fig.8a that stabilization based on machine learning predictions minimizes the impact of departures significantly to a negligible level.However, at the other hand, the stabilization based on older-stable rule shows poor performance as compared to other approaches.It might be due to the fact that this dataset does not contain idle sessions therefore correlation-based prediction produces poor results.It also emphasizes the use of a more robust mechanism such as machine learning.

    Figure 8:Stabilization comparison.(a)Disruptions(b)Population

    6 Conclusion

    This work focusses on the stability of peers in a P2P IPTV architecture.It first extensively analyzed the older-stable principle,the widely used approach for determination of stable nodes.For this purpose,traces of IPTV systems were analyzed for the correlations of elapsed time and remaining time in several scenarios.This analysis revealed that the older stable principle stands true for user sessions when correlation is calculated together.However, in case different behavior groups are separated,this principle does not remain true as very weak, or no correlation is observed.Moreover, in case of individual users,again no correlation was found.

    Secondly,machine learning algorithms were evaluated for their accuracy to predict the stability.This study considered GBR, KNNR and RFR and compared their performance with previously used models namely, SVR and Bayesian network.Over the same training and testing data, all these algorithms performed better than SVR and Bayesian network.Furthermore, GBR produced better accuracy among the newly tested models.

    Finally,a proof-of-concept simulation was conducted which considered the construction of a tree structure based on three methods.In one method, the tree was randomly constructed without any stability mechanism while in other two methods tree was built and stabilized based on older stable principle and machine learning based predictions.Simulation results revealed that machine learning based stabilization can bring significant improvement to the system.

    7 Future Work

    This work presents a small-scale simulation which only considers a single tree.Therefore,future work needs to consider machine learning-based hybrid topology.Furthermore,concrete experiments on real network are needed to validate the system for implementation in deployed systems.Machine learning can be utilized to include other parameters such as available bandwidth and delay.The same approach can be extended to other user-oriented networks.

    Funding Statement:The authors received no specific funding for this study.

    Conflicts of Interest:The authors declare that they have no conflicts of interest to report regarding the present study.

    精品久久久久久久人妻蜜臀av| 精品乱码久久久久久99久播| 国产精品国产高清国产av| 一本精品99久久精品77| 亚洲一区高清亚洲精品| 午夜精品在线福利| 男人的好看免费观看在线视频| 丝袜美腿在线中文| 亚洲性夜色夜夜综合| 国产一区二区激情短视频| 国产欧美日韩精品亚洲av| 久久精品国产清高在天天线| 一进一出好大好爽视频| 国产不卡一卡二| 午夜影院日韩av| 国产一区二区三区av在线 | 又黄又爽又刺激的免费视频.| 成熟少妇高潮喷水视频| 中出人妻视频一区二区| 国产视频内射| 欧美xxxx性猛交bbbb| 成人性生交大片免费视频hd| 老司机影院成人| 日本欧美国产在线视频| 99在线视频只有这里精品首页| 日韩亚洲欧美综合| 午夜精品在线福利| 亚洲欧美中文字幕日韩二区| 国产精华一区二区三区| 亚洲最大成人手机在线| 国产片特级美女逼逼视频| 国产精华一区二区三区| 99久久无色码亚洲精品果冻| 久久韩国三级中文字幕| 欧美激情在线99| 69人妻影院| 搡老妇女老女人老熟妇| 精品免费久久久久久久清纯| 床上黄色一级片| 国产亚洲91精品色在线| 久久久久久国产a免费观看| 亚洲av成人av| 国产欧美日韩精品亚洲av| 国产成年人精品一区二区| 菩萨蛮人人尽说江南好唐韦庄 | 成年女人永久免费观看视频| 亚州av有码| 国产高清激情床上av| 男人舔女人下体高潮全视频| 一级a爱片免费观看的视频| 国产三级中文精品| 在线天堂最新版资源| 国产精品1区2区在线观看.| 伦理电影大哥的女人| 一个人看的www免费观看视频| 伦精品一区二区三区| av专区在线播放| 亚洲成av人片在线播放无| 一个人看视频在线观看www免费| 伊人久久精品亚洲午夜| 看免费成人av毛片| 又黄又爽又免费观看的视频| 三级经典国产精品| 国产精品一区二区三区四区免费观看 | 中文字幕人妻熟人妻熟丝袜美| 校园春色视频在线观看| 特大巨黑吊av在线直播| 国产精品一二三区在线看| 性欧美人与动物交配| 夜夜夜夜夜久久久久| 三级毛片av免费| 日本色播在线视频| 老女人水多毛片| 亚洲美女搞黄在线观看 | 美女大奶头视频| 91av网一区二区| 一a级毛片在线观看| 国产高清激情床上av| 久久精品国产清高在天天线| 97超视频在线观看视频| www.色视频.com| 国产精品人妻久久久影院| 国产精品一及| 我的女老师完整版在线观看| 亚洲av.av天堂| 少妇的逼好多水| 综合色av麻豆| 成人特级av手机在线观看| 中文字幕熟女人妻在线| 岛国在线免费视频观看| 性插视频无遮挡在线免费观看| 成人无遮挡网站| 国产成人aa在线观看| 春色校园在线视频观看| 久99久视频精品免费| 女人被狂操c到高潮| 欧美丝袜亚洲另类| 精品无人区乱码1区二区| 久久精品国产亚洲网站| 晚上一个人看的免费电影| 直男gayav资源| 又黄又爽又刺激的免费视频.| 国产激情偷乱视频一区二区| 免费看光身美女| 中文亚洲av片在线观看爽| 欧美区成人在线视频| 亚洲中文字幕日韩| 精品一区二区三区视频在线观看免费| 人人妻人人澡欧美一区二区| eeuss影院久久| 精品午夜福利在线看| 亚洲无线在线观看| 亚洲aⅴ乱码一区二区在线播放| 国产精品嫩草影院av在线观看| 久久99热6这里只有精品| 天堂影院成人在线观看| 午夜精品一区二区三区免费看| 午夜日韩欧美国产| 国产白丝娇喘喷水9色精品| 麻豆久久精品国产亚洲av| 久久久午夜欧美精品| 成人国产麻豆网| 国产一区二区三区在线臀色熟女| 亚洲五月天丁香| 中文字幕精品亚洲无线码一区| 精品99又大又爽又粗少妇毛片| 大又大粗又爽又黄少妇毛片口| 天天一区二区日本电影三级| 国产极品精品免费视频能看的| 男女啪啪激烈高潮av片| 热99在线观看视频| 99在线人妻在线中文字幕| 亚洲av熟女| 久久久国产成人免费| 亚洲美女搞黄在线观看 | 日韩欧美三级三区| 99在线人妻在线中文字幕| 少妇猛男粗大的猛烈进出视频 | 精品国产三级普通话版| 久久久精品94久久精品| 免费搜索国产男女视频| 别揉我奶头~嗯~啊~动态视频| 中文字幕免费在线视频6| 亚洲av中文字字幕乱码综合| 三级经典国产精品| 日韩精品青青久久久久久| 高清日韩中文字幕在线| 俄罗斯特黄特色一大片| 久久九九热精品免费| 色av中文字幕| 成人一区二区视频在线观看| 老熟妇乱子伦视频在线观看| 1000部很黄的大片| 色综合亚洲欧美另类图片| 免费大片18禁| 国产精品嫩草影院av在线观看| 欧美成人免费av一区二区三区| 看片在线看免费视频| 中文亚洲av片在线观看爽| 日韩中字成人| 长腿黑丝高跟| 美女内射精品一级片tv| 日本黄色片子视频| 男女那种视频在线观看| 国产伦在线观看视频一区| 日本黄大片高清| av国产免费在线观看| 欧美一区二区国产精品久久精品| 性插视频无遮挡在线免费观看| 成人av在线播放网站| 亚洲第一电影网av| 欧美日韩国产亚洲二区| 亚洲中文字幕日韩| 午夜a级毛片| 久久久精品94久久精品| 久久天躁狠狠躁夜夜2o2o| 少妇熟女欧美另类| 午夜久久久久精精品| 人人妻人人澡欧美一区二区| 日本精品一区二区三区蜜桃| 色播亚洲综合网| 秋霞在线观看毛片| 真实男女啪啪啪动态图| av福利片在线观看| 美女大奶头视频| 女生性感内裤真人,穿戴方法视频| 欧美zozozo另类| 精品无人区乱码1区二区| 色尼玛亚洲综合影院| 色5月婷婷丁香| 国产视频一区二区在线看| 成人三级黄色视频| 亚洲精品粉嫩美女一区| 毛片一级片免费看久久久久| 久久99热6这里只有精品| 亚洲三级黄色毛片| 老熟妇乱子伦视频在线观看| 2021天堂中文幕一二区在线观| 亚洲国产欧洲综合997久久,| 卡戴珊不雅视频在线播放| 成年免费大片在线观看| 亚洲三级黄色毛片| 变态另类成人亚洲欧美熟女| 日韩欧美精品v在线| 久久久久久久午夜电影| 俺也久久电影网| 亚洲美女搞黄在线观看 | 精品久久久久久久久av| 在线观看午夜福利视频| 亚洲一区高清亚洲精品| 亚洲中文日韩欧美视频| 亚洲欧美精品自产自拍| avwww免费| 听说在线观看完整版免费高清| 老司机影院成人| 乱码一卡2卡4卡精品| 人妻制服诱惑在线中文字幕| 精品人妻偷拍中文字幕| 久久中文看片网| 国产精品久久久久久av不卡| 亚洲精品456在线播放app| 少妇的逼水好多| 一进一出抽搐gif免费好疼| 精品乱码久久久久久99久播| 亚洲五月天丁香| 中出人妻视频一区二区| 亚洲精华国产精华液的使用体验 | 人人妻人人澡人人爽人人夜夜 | 中文亚洲av片在线观看爽| 亚洲欧美日韩无卡精品| 中国国产av一级| 国产美女午夜福利| 老师上课跳d突然被开到最大视频| 亚洲av五月六月丁香网| 日本色播在线视频| 偷拍熟女少妇极品色| 99热全是精品| 日韩一区二区视频免费看| 国产免费一级a男人的天堂| 99热只有精品国产| 久久久久九九精品影院| 久久久久精品国产欧美久久久| 成人高潮视频无遮挡免费网站| 又粗又爽又猛毛片免费看| 国产成年人精品一区二区| 国产成人影院久久av| 99久久成人亚洲精品观看| 黄色日韩在线| 亚洲精品日韩av片在线观看| 亚洲精品国产成人久久av| 三级毛片av免费| 国产精品野战在线观看| 身体一侧抽搐| 国产精品,欧美在线| 亚洲精品国产av成人精品 | 搡女人真爽免费视频火全软件 | 欧美色欧美亚洲另类二区| 成人欧美大片| 三级国产精品欧美在线观看| 亚洲av成人av| 国产成人aa在线观看| 看片在线看免费视频| av天堂在线播放| 偷拍熟女少妇极品色| 老司机午夜福利在线观看视频| 大香蕉久久网| 久久99热这里只有精品18| 国产精品久久视频播放| 国产 一区 欧美 日韩| 高清午夜精品一区二区三区 | 国产精品免费一区二区三区在线| 亚洲精品影视一区二区三区av| 天堂影院成人在线观看| 欧美区成人在线视频| 亚洲一级一片aⅴ在线观看| 午夜老司机福利剧场| 国产在视频线在精品| 成人特级黄色片久久久久久久| 国产亚洲91精品色在线| 麻豆成人午夜福利视频| 日日摸夜夜添夜夜添小说| 国产成人精品久久久久久| 亚洲精品色激情综合| 成人av在线播放网站| av专区在线播放| 国产免费一级a男人的天堂| 99在线人妻在线中文字幕| 久久精品国产亚洲网站| 天天一区二区日本电影三级| 在线观看一区二区三区| 免费搜索国产男女视频| 久久久久久久久久久丰满| 免费观看在线日韩| 国产精品三级大全| 日本撒尿小便嘘嘘汇集6| 免费看光身美女| 色综合亚洲欧美另类图片| 男插女下体视频免费在线播放| 欧美潮喷喷水| 99久久精品国产国产毛片| 插阴视频在线观看视频| 又黄又爽又刺激的免费视频.| 日本成人三级电影网站| 亚州av有码| 中文字幕精品亚洲无线码一区| 搞女人的毛片| 九九久久精品国产亚洲av麻豆| 亚洲va在线va天堂va国产| 成人美女网站在线观看视频| 身体一侧抽搐| 91久久精品国产一区二区三区| 色视频www国产| 白带黄色成豆腐渣| 国产精品一及| 国产淫片久久久久久久久| 午夜免费男女啪啪视频观看 | 国产精品一区www在线观看| 尾随美女入室| 久久精品国产亚洲av香蕉五月| 久久国内精品自在自线图片| 午夜视频国产福利| 日本与韩国留学比较| 亚洲aⅴ乱码一区二区在线播放| 在线播放国产精品三级| 成人特级黄色片久久久久久久| АⅤ资源中文在线天堂| 国产日本99.免费观看| 色综合色国产| 日本精品一区二区三区蜜桃| 亚洲欧美中文字幕日韩二区| 精品久久久久久久人妻蜜臀av| 亚洲真实伦在线观看| 最近视频中文字幕2019在线8| 成年女人毛片免费观看观看9| 亚洲国产精品合色在线| avwww免费| 欧美xxxx黑人xx丫x性爽| 亚洲国产精品久久男人天堂| 在线观看66精品国产| 一区福利在线观看| 久久久久国内视频| 真实男女啪啪啪动态图| 日本成人三级电影网站| 亚洲综合色惰| 国模一区二区三区四区视频| 久久久久精品国产欧美久久久| 亚洲人成网站在线播放欧美日韩| 男女做爰动态图高潮gif福利片| 女人十人毛片免费观看3o分钟| 亚洲av电影不卡..在线观看| 国产毛片a区久久久久| 在线观看一区二区三区| 97热精品久久久久久| 在线播放国产精品三级| 国产精品亚洲一级av第二区| 哪里可以看免费的av片| 熟女人妻精品中文字幕| 看片在线看免费视频| 国产男靠女视频免费网站| 亚洲七黄色美女视频| 精品久久久久久久久久久久久| 国产极品精品免费视频能看的| 国产一区二区亚洲精品在线观看| 免费在线观看成人毛片| 波多野结衣巨乳人妻| 亚洲真实伦在线观看| 亚洲av熟女| 九九爱精品视频在线观看| 少妇高潮的动态图| 国产欧美日韩精品亚洲av| 99热这里只有是精品在线观看| 国语自产精品视频在线第100页| 亚洲18禁久久av| 国产精品一二三区在线看| 变态另类成人亚洲欧美熟女| 97在线视频观看| 一卡2卡三卡四卡精品乱码亚洲| 亚洲欧美精品自产自拍| 国产精品女同一区二区软件| 国产精品久久电影中文字幕| 国产高清视频在线播放一区| 亚洲七黄色美女视频| 午夜精品一区二区三区免费看| 国产一区二区三区av在线 | 男插女下体视频免费在线播放| 中文字幕av在线有码专区| 一区福利在线观看| 亚洲成人久久爱视频| h日本视频在线播放| 精品乱码久久久久久99久播| 国产视频内射| 亚洲专区国产一区二区| 淫秽高清视频在线观看| 免费看av在线观看网站| 国产精品国产高清国产av| 国产国拍精品亚洲av在线观看| 特大巨黑吊av在线直播| 精品熟女少妇av免费看| 午夜亚洲福利在线播放| 校园春色视频在线观看| 国产视频一区二区在线看| 欧美性感艳星| 人妻丰满熟妇av一区二区三区| 欧美日韩在线观看h| 色播亚洲综合网| 亚洲久久久久久中文字幕| 男女啪啪激烈高潮av片| 精品久久久久久久末码| 97超碰精品成人国产| 我的女老师完整版在线观看| 国产视频内射| 色哟哟哟哟哟哟| 久久韩国三级中文字幕| 亚洲人成网站在线观看播放| 一进一出好大好爽视频| 最新在线观看一区二区三区| 成年女人看的毛片在线观看| 精品免费久久久久久久清纯| 久久久精品大字幕| 少妇被粗大猛烈的视频| 乱系列少妇在线播放| 亚洲色图av天堂| 最近视频中文字幕2019在线8| 又黄又爽又刺激的免费视频.| 亚洲国产精品sss在线观看| 亚洲国产色片| 亚洲专区国产一区二区| 欧美性感艳星| 中文字幕久久专区| 日韩制服骚丝袜av| 麻豆乱淫一区二区| 亚洲中文字幕一区二区三区有码在线看| 免费大片18禁| 99精品在免费线老司机午夜| 国产高清激情床上av| 夜夜夜夜夜久久久久| 午夜福利在线观看免费完整高清在 | 亚洲av中文av极速乱| 久久久久国内视频| 日韩国内少妇激情av| 男女边吃奶边做爰视频| 最近2019中文字幕mv第一页| 在线a可以看的网站| 国产高清有码在线观看视频| 午夜免费男女啪啪视频观看 | 成人亚洲精品av一区二区| 国内少妇人妻偷人精品xxx网站| 亚洲高清免费不卡视频| 欧美日本视频| 国产精品一区二区三区四区久久| 欧美极品一区二区三区四区| 99热只有精品国产| 看非洲黑人一级黄片| 国产黄色视频一区二区在线观看 | 最近在线观看免费完整版| 黄色一级大片看看| 亚洲内射少妇av| 色尼玛亚洲综合影院| 午夜影院日韩av| 人妻少妇偷人精品九色| 国产精品亚洲一级av第二区| 99国产精品一区二区蜜桃av| 午夜免费男女啪啪视频观看 | 日本一本二区三区精品| 国产精品av视频在线免费观看| 国产69精品久久久久777片| 一区二区三区高清视频在线| 午夜福利在线观看吧| 久久久久国产网址| 亚洲国产欧美人成| 国产在线精品亚洲第一网站| 一a级毛片在线观看| 干丝袜人妻中文字幕| 神马国产精品三级电影在线观看| 在现免费观看毛片| av天堂中文字幕网| av在线播放精品| 午夜免费男女啪啪视频观看 | 国产成人影院久久av| 久久鲁丝午夜福利片| 亚洲精品色激情综合| 国产人妻一区二区三区在| 欧美日韩一区二区视频在线观看视频在线 | 97超级碰碰碰精品色视频在线观看| 中文亚洲av片在线观看爽| 性插视频无遮挡在线免费观看| 欧美人与善性xxx| 亚洲精品国产成人久久av| 可以在线观看毛片的网站| 精品乱码久久久久久99久播| 在线播放国产精品三级| 看非洲黑人一级黄片| 一级av片app| 日本与韩国留学比较| 非洲黑人性xxxx精品又粗又长| 国产免费一级a男人的天堂| 男插女下体视频免费在线播放| av天堂中文字幕网| 91久久精品国产一区二区成人| 亚洲,欧美,日韩| 欧美性感艳星| 在线国产一区二区在线| 精品久久国产蜜桃| 美女被艹到高潮喷水动态| 国产男人的电影天堂91| 国产高清三级在线| 我的女老师完整版在线观看| 日本在线视频免费播放| 一a级毛片在线观看| 午夜精品国产一区二区电影 | 国产91av在线免费观看| 精华霜和精华液先用哪个| 国产亚洲欧美98| 午夜福利在线在线| 三级毛片av免费| 22中文网久久字幕| 国产成人aa在线观看| 大型黄色视频在线免费观看| 能在线免费观看的黄片| 精品一区二区三区人妻视频| 国内精品宾馆在线| 你懂的网址亚洲精品在线观看 | 蜜臀久久99精品久久宅男| 狂野欧美白嫩少妇大欣赏| 老熟妇仑乱视频hdxx| 能在线免费观看的黄片| 国产高潮美女av| 麻豆国产av国片精品| 22中文网久久字幕| 国产淫片久久久久久久久| 精品一区二区三区视频在线观看免费| 日韩欧美精品免费久久| 国产亚洲精品久久久久久毛片| 99riav亚洲国产免费| 亚洲电影在线观看av| 给我免费播放毛片高清在线观看| 亚洲人成网站在线播放欧美日韩| 插阴视频在线观看视频| 日本黄色视频三级网站网址| 禁无遮挡网站| 国产午夜福利久久久久久| 日本欧美国产在线视频| 亚洲电影在线观看av| 国产精品嫩草影院av在线观看| 我的老师免费观看完整版| 日韩亚洲欧美综合| av在线蜜桃| 在线免费十八禁| 男女啪啪激烈高潮av片| 免费在线观看成人毛片| 亚洲乱码一区二区免费版| 干丝袜人妻中文字幕| av国产免费在线观看| 中出人妻视频一区二区| 3wmmmm亚洲av在线观看| 蜜桃久久精品国产亚洲av| 久久久久久伊人网av| 精品久久久久久久久久免费视频| 亚洲精品色激情综合| 免费人成在线观看视频色| 亚洲精品粉嫩美女一区| 欧美潮喷喷水| 亚洲最大成人av| 久久久久久久久中文| 成人漫画全彩无遮挡| 成人永久免费在线观看视频| 亚洲av第一区精品v没综合| 天天一区二区日本电影三级| 免费人成视频x8x8入口观看| 少妇高潮的动态图| 久久6这里有精品| 欧美最新免费一区二区三区| 一边摸一边抽搐一进一小说| 国产老妇女一区| 毛片女人毛片| 身体一侧抽搐| 国产人妻一区二区三区在| 国产av在哪里看| 欧美成人精品欧美一级黄| 久久亚洲精品不卡| 真实男女啪啪啪动态图| 五月玫瑰六月丁香| 日韩一区二区视频免费看| a级毛片a级免费在线| 91狼人影院| 日韩成人av中文字幕在线观看 | 综合色丁香网| 日韩,欧美,国产一区二区三区 | 精品一区二区三区人妻视频| 亚洲成人久久爱视频| 嫩草影视91久久| 国产69精品久久久久777片| 精品久久久久久久久亚洲| АⅤ资源中文在线天堂| 精品久久久久久久久久免费视频| 日韩三级伦理在线观看| 麻豆一二三区av精品| 久久久久久大精品| 日日摸夜夜添夜夜添av毛片| 日日干狠狠操夜夜爽| 日本成人三级电影网站| 久久久国产成人免费| 午夜福利在线在线| 亚洲av不卡在线观看| www.色视频.com| 国内精品久久久久精免费| 淫秽高清视频在线观看| 亚洲最大成人中文| 最后的刺客免费高清国语| 两个人的视频大全免费| 日本黄色片子视频| 亚洲美女视频黄频| 亚洲国产精品国产精品| 色吧在线观看| 99精品在免费线老司机午夜|