International journal of future generation communication and networking, Dec 31, 2016
Against the background of a surge in software-defined network (SDN) as well as urgent demands for... more Against the background of a surge in software-defined network (SDN) as well as urgent demands for LTE-WIFI internetworking and IPv6 network integration, the paper integrates LTE mobile telecommunication technology, wireless local area network and 5G (the fifth generation mobile communication system) into one, in an attempt to complete routing, distribution, and transmission of different radio frequency signals (i.e. LTE, WIFI) on the SDN basis. As a result, large-area coverage of LTE-WIFI network signals is realized in a way that meeting the demands of reliability, safety, flexibility and extendibility for mobile internet. What is more, different complementary wireless telecommunication systems are closed combined with each other. Through the use of SDN integration technology, the paper designs network integration architecture of IP-LTE mobile internet and WIFI, and also studies on the application of SDN-based IP optical integration network, which has broad prospects.
International journal of future generation communication and networking, Dec 31, 2016
Against the backdrop of Internet of Things on campus, this paper makes Internet of things cover t... more Against the backdrop of Internet of Things on campus, this paper makes Internet of things cover the entire campus through wired and wireless accesses. As a result, campus sees the fulfillment of information transfer via computers, mobile phones, devices and other articles, which is the embodiment of intelligence. Besides, OPNET is used for modeling and simulation of wireless network structure in SDN intelligent campus, with a view to analyzing the feasibility and advantages of various campus network topologies. Then, a reasonable scheme is presented responding to different types of topologies, which is conducive to better building an intelligent campus.
2021 IEEE 23rd Int Conf on High Performance Computing & Communications; 7th Int Conf on Data Science & Systems; 19th Int Conf on Smart City; 7th Int Conf on Dependability in Sensor, Cloud & Big Data Systems & Application (HPCC/DSS/SmartCity/DependSys)
To keep pace with the exploding data volume raised from geographical distribution edge networks, ... more To keep pace with the exploding data volume raised from geographical distribution edge networks, more and more edge servers have been built in recent years. As the computing power and storage capacity are different on each server, requests have to be transferred from one server to another before finally being responded and returned back to users. Such server-toserver transmission naturally introduce non negligible latency, which inevitably affects the quality of service (QoS). To eliminate this transmission latency, Internet Service Providers (ISPs) are building or renting more edge servers (both computing servers and storage servers) to reduce transmission distance and enhance the server configuration, which brings great costs. Fortunately, through a large number of real trace analysis, we found that it is possible to reduce server number while keeping the QoS! In this paper, we first disclose three key characteristics from Kuaishou Company: (1) Unbalanced request frequencies on different servers; (2) Imprecise latency measure on server-to-server transmission; (3) Nonlinear latency reduction to server number increment. Based on these findings, we propose a frequencyaware edge storage server deployment strategy Frend that is an improved Genetic Algorithm to optimize the number of edge storage servers by the internal diffusion capability that is a new latency measure called S2SL. Through a series of experiments using real application data, we demonstrate that while achieving the same S2SL, Frend can reduce the number of required edge storage servers by up to 56% compared with the state-of-the-art Anveshak method.
2021 IEEE International Conference on Big Data (Big Data), 2021
As mobile networks proliferate, we are experiencing a strong diversification of services, which r... more As mobile networks proliferate, we are experiencing a strong diversification of services, which requires greater flexibility from the existing network. Network slicing is proposed as a promising solution for resource utilization in 5G and future networks to address this dire need. In network slicing, dynamic resource orchestration and network slice management are crucial for maximizing resource utilization. Unfortunately, this process is too complex for traditional approaches to be effective due to a lack of accurate models and dynamic hidden structures. We formulate the problem as a Constrained Markov Decision Process (CMDP) without knowing models and hidden structures. Additionally, we propose to solve the problem using CLARA, a Constrained reinforcement LeArning based Resource Allocation algorithm. In particular, we analyze cumulative and instantaneous constraints using adaptive interior-point policy optimization and projection layer, respectively. Evaluations show that CLARA clearly outperforms baselines in resource allocation with service demand guarantees.
2018 IFIP Networking Conference (IFIP Networking) and Workshops, 2018
Much attention of the research community has focused on performance analysis of cache networks un... more Much attention of the research community has focused on performance analysis of cache networks under various caching policies. However, the issue of how to evaluate and compare caching policies for cache networks has not been adequately addressed. In this paper, we propose a novel and general framework for evaluating caching policies in a hierarchical network of caches. We introduce the notion of a hit probability/rate matrix, and employ a generalized notion of majorization as the basic tool for evaluating caching policies for various performance metrics. We discuss how the framework can be applied to existing caching policies, and conduct extensive simulation-based evaluation to demonstrate the utility and accuracy of our framework.
Proceedings of the 24th International Conference on World Wide Web, 2015
Reachability is crucial to many network operations in various complex networks. More often than n... more Reachability is crucial to many network operations in various complex networks. More often than not, however, it is not sufficient simply to know whether a source node s can reach a target node t in the network. Additional information associated with reachability such as how long or how many possible ways node s may take to reach node t. In this paper we analyze another piece of important information associated with reachability -which we call pivotality. Pivotality captures how pivotal a role that a node k or a subset of nodes S may play in the reachability from node s to node t in a given network. We propose two important metrics, the avoidance and transit hitting times, which extend and generalize the classical notion of hitting times. We show these metrics can be computed from the fundamental matrices associated with the appropriately defined random walk transition probability matrices and prove that the classical hitting time from a source to a target can be decomposed into the avoidance and transit hitting times with respect to any third node. Through simulated and real-world network examples, we demonstrate that these metrics provide a powerful ranking tool for the nodes based on their pivotality in the reachability.
We establish asymptotic upper and lower bounds on the asymptotic decay rate of persession queue l... more We establish asymptotic upper and lower bounds on the asymptotic decay rate of persession queue length tail distributions for a multiple-queue system where a single constant rate server services the queues using the generalized processor sharing (GPS) scheduling discipline. In the special case where there are only two queues, the upper and lower bounds match, yielding the optimal bound proved in [15]. The dynamics of bandwidth sharing of a multiple-queue GPS system is captured using the notion of partial feasible sets, and the bounds are obtained using the sample-path large deviation principle. The results have implications in call admission control for high-speed communication networks.
Proceedings of the 3rd annual ACM workshop on Mining network data, 2007
With the widespread adoption of SIP-based VoIP, understanding the characteristics of SIP traffic ... more With the widespread adoption of SIP-based VoIP, understanding the characteristics of SIP traffic behavior is critical to problem diagnosis and security protection of IP Telephony. In this paper, we propose a general methodology for profiling SIP-based VoIP traffic behavior at multiple levels: SIP server host, server entity and individual user levels. Using SIP traffic traces captured in a production VoIP service, we illustrate the characteristics of SIP-based VoIP traffic behavior in an operational network and demonstrate the effectiveness of our general profiling methodology. In particular, we show how our profiling methodology can help identify performance anomalies through a case study.
Proceedings of the 8th ACM Conference on Embedded Networked Sensor Systems, 2010
The ability to move energy around makes it feasible to build distributed energy storage systems t... more The ability to move energy around makes it feasible to build distributed energy storage systems that can robustly extend the lifetime of networked sensor systems. eShare supports the concept of energy sharing among multiple embedded sensor devices by providing designs for energy routers (i.e., energy storage and routing devices) and related energy access and network protocols. In a nutshell, energy routers exchange energy sharing control information using their data network while sharing energy freely among connected embedded sensor devices using their energy network. To improve sharing efficiency subject to energy leakage, we develop an effective energy charging and discharging mechanism using an array of ultra-capacitors as the main component of an energy router. We extensively evaluate our system under six real-world settings. Results indicate our charging and discharging control can effectively minimize the energy leaked away. Moreover, the energy sharing protocol can quantitatively share 113J energy with 96.82% accuracy in less than 2 seconds.
Netflix is the leading provider of on-demand Internet video streaming in the US and Canada, accou... more Netflix is the leading provider of on-demand Internet video streaming in the US and Canada, accounting for 29.7% of the peak downstream traffic in US. Understanding the Netflix architecture and its performance can shed light on how to best optimize its design as well as on the design of similar on-demand streaming services. In this paper, we perform a measurement study of Netflix to uncover its architecture and service strategy. We find that Netflix employs a blend of data centers and Content Delivery Networks (CDNs) for content distribution. We also perform active measurements of the three CDNs employed by Netflix to quantify the video delivery bandwidth available to users across the US. Finally, as improvements to Netflix's current CDN assignment strategy, we propose a measurement-based adaptive CDN selection strategy and a multiple-CDN-based video delivery strategy, and demonstrate their potentials in significantly increasing user's average bandwidth.
Proceedings of the ACM SIGCOMM 2013 conference on SIGCOMM, 2013
With the proliferation of online social networking (OSN) and mobile devices, preserving user priv... more With the proliferation of online social networking (OSN) and mobile devices, preserving user privacy has become a great challenge. While prior studies have directly focused on OSN services, we call attention to the privacy leakage in mobile network data. This concern is motivated by two factors. First, the prevalence of OSN usage leaves identifiable digital footprints that can be traced back to users in the real-world. Second, the association between users and their mobile devices makes it easier to associate traffic to its owners. These pose a serious threat to user privacy as they enable an adversary to attribute significant portions of data traffic including the ones with NO identity leaks to network users' true identities. To demonstrate its feasibility, we develop the Tessellation methodology. By applying Tessellation on traffic from a cellular service provider (CSP), we show that up to 50% of the traffic can be attributed to the names of users. In addition to revealing the user identity, the reconstructed profile, dubbed as "mosaic," associates personal information such as political views, browsing habits, and favorite apps to the users. We conclude by discussing approaches for preventing and mitigating the alarming leakage of sensitive user information.
Proceedings of the 2014 Conference on Internet Measurement Conference, 2014
Cloud storage services such as Dropbox, Google Drive, and Microsoft OneDrive provide users with a... more Cloud storage services such as Dropbox, Google Drive, and Microsoft OneDrive provide users with a convenient and reliable way to store and share data from anywhere, on any device, and at any time. The cornerstone of these services is the data synchronization (sync) operation which automatically maps the changes in users' local filesystems to the cloud via a series of network communications in a timely manner. If not designed properly, however, the tremendous amount of data sync traffic can potentially cause (financial) pains to both service providers and users. This paper addresses a simple yet critical question: Is the current data sync traffic of cloud storage services efficiently used? We first define a novel metric named TUE to quantify the Traffic Usage Efficiency of data synchronization. Based on both real-world traces and comprehensive experiments, we study and characterize the TUE of six widely used cloud storage services. Our results demonstrate that a considerable portion of the data sync traffic is in a sense wasteful, and can be effectively avoided or significantly reduced via carefully designed data sync mechanisms. All in all, our study of TUE of cloud storage services not only provides guidance for service providers to develop more efficient, trafficeconomic services, but also helps users pick appropriate services that best fit their needs and budgets.
2011 Proceedings of 20th International Conference on Computer Communications and Networks (ICCCN), 2011
YouTube is one of the most popular video sharing websites in the world. In order to serve its glo... more YouTube is one of the most popular video sharing websites in the world. In order to serve its globally distributed users, it requires a massive-scale video delivery system. A major part of the whole system is to decide exactly what server machine is going to serve a client request at any given time. In this paper, we analyze DNS resolutions and video playback traces collected by playing half a million YouTube videos from geographically distributed PlanetLab nodes to uncover loadbalancing and server selection strategies used by YouTube. Our results indicate that YouTube is aggressively deploying cache servers of widely varying sizes at many different locations around the world with several of them located inside other ISPs to reduce cost and improve the end-user performance. We also find that YouTube tries to use local "per-cache" load-sharing before resorting to redirecting a user to bigger/central cache locations.
Proceedings of the 2011 ACM SIGCOMM conference on Internet measurement conference, 2011
This paper investigates the roles of front-end (proxy) servers in improving user-perceived perfor... more This paper investigates the roles of front-end (proxy) servers in improving user-perceived performance of dynamic content distribution. Using Bing and Google search services as two case studies, we perform extensive network measurement and analysis to understand several key factors that affect the overall user-perceived performance. In particular, we develop a simple model-based inference framework to indirectly measure and quantify the (directly unobservable) "frontend-to-backend fetching time" comprised of the query processing time at back-end data centers and the delivery time between the back-end data centers and front-end servers. We show that this fetching time plays a critical role in the end-to-end performance of dynamic content delivery.
We build a distributed active measurement infrastructure to uncover the internals of the YouTube ... more We build a distributed active measurement infrastructure to uncover the internals of the YouTube video delivery system. We deduce the key design features behind the YouTube video delivery system by collecting and analyzing a large amount of video playback logs, DNS mappings and latency data and by performing additional measurements to verify the findings. We find that the design of the YouTube video delivery system consists of three major components: a "flat" video id space, multiple DNS namespaces reflecting a multilayered logical organization of video servers, and a 3-tier physical cache hierarchy. Further, YouTube employs a set of sophisticated mechanisms to handle video delivery dynamics such as cache misses and load sharing among its globally distributed cache locations and datacenters.
Receiver adaptation layered multicast video has been pro- posed to address the issue of network b... more Receiver adaptation layered multicast video has been pro- posed to address the issue of network bandwidth heterogeneity in multi- party video communication over networks. An important issue in this ap- proach is how to conduct layer adaptation to achieve the "optimal level" of subscription to maximize the perceived video quality at receivers while uti- lize bandwidth efficiently. In the current work, optimal layers are typically determined by congestion signals fed back from networks. This approach implies that instantaneous congestion is a key parameter; however, in ran- dom traffic, such an implication somewhat under-estimates other parame- ters. In order to investigate the performance of "layer adaptation" in random traffic for improving the network-based multi-party video communication system, in this paper we present an analytical studying of the issue of re- ceiver layer adaptation for layered video transmission. In this work, we use the notion of congestio...
Abstrucf-Video delivery from a server to a client across a network is an important component of m... more Abstrucf-Video delivery from a server to a client across a network is an important component of many multimedia applications. While delivering a video stream across a resource constrained network, loss of frames may be unavoidable. Under such circumstances, it is desirable to find a servcr transmission schedule that can efficiently utilize the network resources while maximizing the perceived quality-of-service (QoS) at the client. To address this issue, in this paper we introduce the notion of selective ,frume discurd at the server and formulate the opfimul selective frume discitrd problem using a QoS-based cost function. Given network bandwidth and client buffer constraints, we develop an O ( N log N ) algorithm to find the minimum number of frames that must be discarded in order to meet these constraints. The correctness of the algorithm is also formally established. Since the computational complexity of the optimal algorithm for solving the optimal selective frame discard problem is prohibitively high in general, we also develop several efficient heuristic algorithms for selective frame discard. These algorithms are evaluated using JPEG video traces.
Broadband Cable Networks (BCNs) bring high-speed Internet access to home and make emerging multim... more Broadband Cable Networks (BCNs) bring high-speed Internet access to home and make emerging multimedia streaming applications feasible. However, bandwidth contention is still a challenging problem in providing efficient IP-based Video-On-Demand (VOD) service on BCNs, due to the lack of effective approaches to exploit the unique characteristics of BCNs. To address the bandwidth contention issue, we propose an efficient video scheduling technique, called full-sharing scheduling in this paper. This technique fully exploits the unique characteristics of BCNs to reduce the bandwidth consumption of video sessions sharing a cable channel of fixed capacity, thereby maximizing the number of simultaneous video sessions on the single channel. Furthermore, we analyze the expected bandwidth and the session blocking probability of a video under the full-sharing scheduling. Based on this analysis, we design an efficient video assignment mechanism for maximizing the profit of a VOD system in scheduling videos on BCNs. Through both analysis and simulation, we show that our approach minimizes the bandwidth consumption of video sessions compared with the previous approaches and has significant advantages on BCNs. The proposed approach is also directly applicable on other broadcast/multicast networks
The expected commute times for a strongly connected directed graph are related to an asymmetric L... more The expected commute times for a strongly connected directed graph are related to an asymmetric Laplacian matrix as a direct extension to similar well known formulas for undirected graphs. We show the close relationships between the asymmetric Laplacian and the so-called Fundamental matrix. We give bounds for the commute times in terms of the stationary probabilities for a random walk over the graph together with the asymmetric Laplacian and show how this can be approximated by a symmetrized Laplacian derived from a related weighted undirected graph.
International journal of future generation communication and networking, Dec 31, 2016
Against the background of a surge in software-defined network (SDN) as well as urgent demands for... more Against the background of a surge in software-defined network (SDN) as well as urgent demands for LTE-WIFI internetworking and IPv6 network integration, the paper integrates LTE mobile telecommunication technology, wireless local area network and 5G (the fifth generation mobile communication system) into one, in an attempt to complete routing, distribution, and transmission of different radio frequency signals (i.e. LTE, WIFI) on the SDN basis. As a result, large-area coverage of LTE-WIFI network signals is realized in a way that meeting the demands of reliability, safety, flexibility and extendibility for mobile internet. What is more, different complementary wireless telecommunication systems are closed combined with each other. Through the use of SDN integration technology, the paper designs network integration architecture of IP-LTE mobile internet and WIFI, and also studies on the application of SDN-based IP optical integration network, which has broad prospects.
International journal of future generation communication and networking, Dec 31, 2016
Against the backdrop of Internet of Things on campus, this paper makes Internet of things cover t... more Against the backdrop of Internet of Things on campus, this paper makes Internet of things cover the entire campus through wired and wireless accesses. As a result, campus sees the fulfillment of information transfer via computers, mobile phones, devices and other articles, which is the embodiment of intelligence. Besides, OPNET is used for modeling and simulation of wireless network structure in SDN intelligent campus, with a view to analyzing the feasibility and advantages of various campus network topologies. Then, a reasonable scheme is presented responding to different types of topologies, which is conducive to better building an intelligent campus.
2021 IEEE 23rd Int Conf on High Performance Computing & Communications; 7th Int Conf on Data Science & Systems; 19th Int Conf on Smart City; 7th Int Conf on Dependability in Sensor, Cloud & Big Data Systems & Application (HPCC/DSS/SmartCity/DependSys)
To keep pace with the exploding data volume raised from geographical distribution edge networks, ... more To keep pace with the exploding data volume raised from geographical distribution edge networks, more and more edge servers have been built in recent years. As the computing power and storage capacity are different on each server, requests have to be transferred from one server to another before finally being responded and returned back to users. Such server-toserver transmission naturally introduce non negligible latency, which inevitably affects the quality of service (QoS). To eliminate this transmission latency, Internet Service Providers (ISPs) are building or renting more edge servers (both computing servers and storage servers) to reduce transmission distance and enhance the server configuration, which brings great costs. Fortunately, through a large number of real trace analysis, we found that it is possible to reduce server number while keeping the QoS! In this paper, we first disclose three key characteristics from Kuaishou Company: (1) Unbalanced request frequencies on different servers; (2) Imprecise latency measure on server-to-server transmission; (3) Nonlinear latency reduction to server number increment. Based on these findings, we propose a frequencyaware edge storage server deployment strategy Frend that is an improved Genetic Algorithm to optimize the number of edge storage servers by the internal diffusion capability that is a new latency measure called S2SL. Through a series of experiments using real application data, we demonstrate that while achieving the same S2SL, Frend can reduce the number of required edge storage servers by up to 56% compared with the state-of-the-art Anveshak method.
2021 IEEE International Conference on Big Data (Big Data), 2021
As mobile networks proliferate, we are experiencing a strong diversification of services, which r... more As mobile networks proliferate, we are experiencing a strong diversification of services, which requires greater flexibility from the existing network. Network slicing is proposed as a promising solution for resource utilization in 5G and future networks to address this dire need. In network slicing, dynamic resource orchestration and network slice management are crucial for maximizing resource utilization. Unfortunately, this process is too complex for traditional approaches to be effective due to a lack of accurate models and dynamic hidden structures. We formulate the problem as a Constrained Markov Decision Process (CMDP) without knowing models and hidden structures. Additionally, we propose to solve the problem using CLARA, a Constrained reinforcement LeArning based Resource Allocation algorithm. In particular, we analyze cumulative and instantaneous constraints using adaptive interior-point policy optimization and projection layer, respectively. Evaluations show that CLARA clearly outperforms baselines in resource allocation with service demand guarantees.
2018 IFIP Networking Conference (IFIP Networking) and Workshops, 2018
Much attention of the research community has focused on performance analysis of cache networks un... more Much attention of the research community has focused on performance analysis of cache networks under various caching policies. However, the issue of how to evaluate and compare caching policies for cache networks has not been adequately addressed. In this paper, we propose a novel and general framework for evaluating caching policies in a hierarchical network of caches. We introduce the notion of a hit probability/rate matrix, and employ a generalized notion of majorization as the basic tool for evaluating caching policies for various performance metrics. We discuss how the framework can be applied to existing caching policies, and conduct extensive simulation-based evaluation to demonstrate the utility and accuracy of our framework.
Proceedings of the 24th International Conference on World Wide Web, 2015
Reachability is crucial to many network operations in various complex networks. More often than n... more Reachability is crucial to many network operations in various complex networks. More often than not, however, it is not sufficient simply to know whether a source node s can reach a target node t in the network. Additional information associated with reachability such as how long or how many possible ways node s may take to reach node t. In this paper we analyze another piece of important information associated with reachability -which we call pivotality. Pivotality captures how pivotal a role that a node k or a subset of nodes S may play in the reachability from node s to node t in a given network. We propose two important metrics, the avoidance and transit hitting times, which extend and generalize the classical notion of hitting times. We show these metrics can be computed from the fundamental matrices associated with the appropriately defined random walk transition probability matrices and prove that the classical hitting time from a source to a target can be decomposed into the avoidance and transit hitting times with respect to any third node. Through simulated and real-world network examples, we demonstrate that these metrics provide a powerful ranking tool for the nodes based on their pivotality in the reachability.
We establish asymptotic upper and lower bounds on the asymptotic decay rate of persession queue l... more We establish asymptotic upper and lower bounds on the asymptotic decay rate of persession queue length tail distributions for a multiple-queue system where a single constant rate server services the queues using the generalized processor sharing (GPS) scheduling discipline. In the special case where there are only two queues, the upper and lower bounds match, yielding the optimal bound proved in [15]. The dynamics of bandwidth sharing of a multiple-queue GPS system is captured using the notion of partial feasible sets, and the bounds are obtained using the sample-path large deviation principle. The results have implications in call admission control for high-speed communication networks.
Proceedings of the 3rd annual ACM workshop on Mining network data, 2007
With the widespread adoption of SIP-based VoIP, understanding the characteristics of SIP traffic ... more With the widespread adoption of SIP-based VoIP, understanding the characteristics of SIP traffic behavior is critical to problem diagnosis and security protection of IP Telephony. In this paper, we propose a general methodology for profiling SIP-based VoIP traffic behavior at multiple levels: SIP server host, server entity and individual user levels. Using SIP traffic traces captured in a production VoIP service, we illustrate the characteristics of SIP-based VoIP traffic behavior in an operational network and demonstrate the effectiveness of our general profiling methodology. In particular, we show how our profiling methodology can help identify performance anomalies through a case study.
Proceedings of the 8th ACM Conference on Embedded Networked Sensor Systems, 2010
The ability to move energy around makes it feasible to build distributed energy storage systems t... more The ability to move energy around makes it feasible to build distributed energy storage systems that can robustly extend the lifetime of networked sensor systems. eShare supports the concept of energy sharing among multiple embedded sensor devices by providing designs for energy routers (i.e., energy storage and routing devices) and related energy access and network protocols. In a nutshell, energy routers exchange energy sharing control information using their data network while sharing energy freely among connected embedded sensor devices using their energy network. To improve sharing efficiency subject to energy leakage, we develop an effective energy charging and discharging mechanism using an array of ultra-capacitors as the main component of an energy router. We extensively evaluate our system under six real-world settings. Results indicate our charging and discharging control can effectively minimize the energy leaked away. Moreover, the energy sharing protocol can quantitatively share 113J energy with 96.82% accuracy in less than 2 seconds.
Netflix is the leading provider of on-demand Internet video streaming in the US and Canada, accou... more Netflix is the leading provider of on-demand Internet video streaming in the US and Canada, accounting for 29.7% of the peak downstream traffic in US. Understanding the Netflix architecture and its performance can shed light on how to best optimize its design as well as on the design of similar on-demand streaming services. In this paper, we perform a measurement study of Netflix to uncover its architecture and service strategy. We find that Netflix employs a blend of data centers and Content Delivery Networks (CDNs) for content distribution. We also perform active measurements of the three CDNs employed by Netflix to quantify the video delivery bandwidth available to users across the US. Finally, as improvements to Netflix's current CDN assignment strategy, we propose a measurement-based adaptive CDN selection strategy and a multiple-CDN-based video delivery strategy, and demonstrate their potentials in significantly increasing user's average bandwidth.
Proceedings of the ACM SIGCOMM 2013 conference on SIGCOMM, 2013
With the proliferation of online social networking (OSN) and mobile devices, preserving user priv... more With the proliferation of online social networking (OSN) and mobile devices, preserving user privacy has become a great challenge. While prior studies have directly focused on OSN services, we call attention to the privacy leakage in mobile network data. This concern is motivated by two factors. First, the prevalence of OSN usage leaves identifiable digital footprints that can be traced back to users in the real-world. Second, the association between users and their mobile devices makes it easier to associate traffic to its owners. These pose a serious threat to user privacy as they enable an adversary to attribute significant portions of data traffic including the ones with NO identity leaks to network users' true identities. To demonstrate its feasibility, we develop the Tessellation methodology. By applying Tessellation on traffic from a cellular service provider (CSP), we show that up to 50% of the traffic can be attributed to the names of users. In addition to revealing the user identity, the reconstructed profile, dubbed as "mosaic," associates personal information such as political views, browsing habits, and favorite apps to the users. We conclude by discussing approaches for preventing and mitigating the alarming leakage of sensitive user information.
Proceedings of the 2014 Conference on Internet Measurement Conference, 2014
Cloud storage services such as Dropbox, Google Drive, and Microsoft OneDrive provide users with a... more Cloud storage services such as Dropbox, Google Drive, and Microsoft OneDrive provide users with a convenient and reliable way to store and share data from anywhere, on any device, and at any time. The cornerstone of these services is the data synchronization (sync) operation which automatically maps the changes in users' local filesystems to the cloud via a series of network communications in a timely manner. If not designed properly, however, the tremendous amount of data sync traffic can potentially cause (financial) pains to both service providers and users. This paper addresses a simple yet critical question: Is the current data sync traffic of cloud storage services efficiently used? We first define a novel metric named TUE to quantify the Traffic Usage Efficiency of data synchronization. Based on both real-world traces and comprehensive experiments, we study and characterize the TUE of six widely used cloud storage services. Our results demonstrate that a considerable portion of the data sync traffic is in a sense wasteful, and can be effectively avoided or significantly reduced via carefully designed data sync mechanisms. All in all, our study of TUE of cloud storage services not only provides guidance for service providers to develop more efficient, trafficeconomic services, but also helps users pick appropriate services that best fit their needs and budgets.
2011 Proceedings of 20th International Conference on Computer Communications and Networks (ICCCN), 2011
YouTube is one of the most popular video sharing websites in the world. In order to serve its glo... more YouTube is one of the most popular video sharing websites in the world. In order to serve its globally distributed users, it requires a massive-scale video delivery system. A major part of the whole system is to decide exactly what server machine is going to serve a client request at any given time. In this paper, we analyze DNS resolutions and video playback traces collected by playing half a million YouTube videos from geographically distributed PlanetLab nodes to uncover loadbalancing and server selection strategies used by YouTube. Our results indicate that YouTube is aggressively deploying cache servers of widely varying sizes at many different locations around the world with several of them located inside other ISPs to reduce cost and improve the end-user performance. We also find that YouTube tries to use local "per-cache" load-sharing before resorting to redirecting a user to bigger/central cache locations.
Proceedings of the 2011 ACM SIGCOMM conference on Internet measurement conference, 2011
This paper investigates the roles of front-end (proxy) servers in improving user-perceived perfor... more This paper investigates the roles of front-end (proxy) servers in improving user-perceived performance of dynamic content distribution. Using Bing and Google search services as two case studies, we perform extensive network measurement and analysis to understand several key factors that affect the overall user-perceived performance. In particular, we develop a simple model-based inference framework to indirectly measure and quantify the (directly unobservable) "frontend-to-backend fetching time" comprised of the query processing time at back-end data centers and the delivery time between the back-end data centers and front-end servers. We show that this fetching time plays a critical role in the end-to-end performance of dynamic content delivery.
We build a distributed active measurement infrastructure to uncover the internals of the YouTube ... more We build a distributed active measurement infrastructure to uncover the internals of the YouTube video delivery system. We deduce the key design features behind the YouTube video delivery system by collecting and analyzing a large amount of video playback logs, DNS mappings and latency data and by performing additional measurements to verify the findings. We find that the design of the YouTube video delivery system consists of three major components: a "flat" video id space, multiple DNS namespaces reflecting a multilayered logical organization of video servers, and a 3-tier physical cache hierarchy. Further, YouTube employs a set of sophisticated mechanisms to handle video delivery dynamics such as cache misses and load sharing among its globally distributed cache locations and datacenters.
Receiver adaptation layered multicast video has been pro- posed to address the issue of network b... more Receiver adaptation layered multicast video has been pro- posed to address the issue of network bandwidth heterogeneity in multi- party video communication over networks. An important issue in this ap- proach is how to conduct layer adaptation to achieve the "optimal level" of subscription to maximize the perceived video quality at receivers while uti- lize bandwidth efficiently. In the current work, optimal layers are typically determined by congestion signals fed back from networks. This approach implies that instantaneous congestion is a key parameter; however, in ran- dom traffic, such an implication somewhat under-estimates other parame- ters. In order to investigate the performance of "layer adaptation" in random traffic for improving the network-based multi-party video communication system, in this paper we present an analytical studying of the issue of re- ceiver layer adaptation for layered video transmission. In this work, we use the notion of congestio...
Abstrucf-Video delivery from a server to a client across a network is an important component of m... more Abstrucf-Video delivery from a server to a client across a network is an important component of many multimedia applications. While delivering a video stream across a resource constrained network, loss of frames may be unavoidable. Under such circumstances, it is desirable to find a servcr transmission schedule that can efficiently utilize the network resources while maximizing the perceived quality-of-service (QoS) at the client. To address this issue, in this paper we introduce the notion of selective ,frume discurd at the server and formulate the opfimul selective frume discitrd problem using a QoS-based cost function. Given network bandwidth and client buffer constraints, we develop an O ( N log N ) algorithm to find the minimum number of frames that must be discarded in order to meet these constraints. The correctness of the algorithm is also formally established. Since the computational complexity of the optimal algorithm for solving the optimal selective frame discard problem is prohibitively high in general, we also develop several efficient heuristic algorithms for selective frame discard. These algorithms are evaluated using JPEG video traces.
Broadband Cable Networks (BCNs) bring high-speed Internet access to home and make emerging multim... more Broadband Cable Networks (BCNs) bring high-speed Internet access to home and make emerging multimedia streaming applications feasible. However, bandwidth contention is still a challenging problem in providing efficient IP-based Video-On-Demand (VOD) service on BCNs, due to the lack of effective approaches to exploit the unique characteristics of BCNs. To address the bandwidth contention issue, we propose an efficient video scheduling technique, called full-sharing scheduling in this paper. This technique fully exploits the unique characteristics of BCNs to reduce the bandwidth consumption of video sessions sharing a cable channel of fixed capacity, thereby maximizing the number of simultaneous video sessions on the single channel. Furthermore, we analyze the expected bandwidth and the session blocking probability of a video under the full-sharing scheduling. Based on this analysis, we design an efficient video assignment mechanism for maximizing the profit of a VOD system in scheduling videos on BCNs. Through both analysis and simulation, we show that our approach minimizes the bandwidth consumption of video sessions compared with the previous approaches and has significant advantages on BCNs. The proposed approach is also directly applicable on other broadcast/multicast networks
The expected commute times for a strongly connected directed graph are related to an asymmetric L... more The expected commute times for a strongly connected directed graph are related to an asymmetric Laplacian matrix as a direct extension to similar well known formulas for undirected graphs. We show the close relationships between the asymmetric Laplacian and the so-called Fundamental matrix. We give bounds for the commute times in terms of the stationary probabilities for a random walk over the graph together with the asymmetric Laplacian and show how this can be approximated by a symmetrized Laplacian derived from a related weighted undirected graph.
Uploads
Papers by Zhi-Li Zhang