Multicast routing in inter-datacenter wide area networks is to compute a multicast tree from a source datacenter to multiple destination datacenters. Software-defined networking enables optimal delay-sensitive routing for multicast sessions in inter-datacenter WANs. Delay-sensitive routing requires the controller to obtain the up-to-date latency of every link. However, the communication overhead and time for collecting the latency information of all links can be prohibitively high for the controller to perform real-time scheduling in a WAN. Inspired by the sparsity of link load information in WANs, we propose compressive latency monitoring, a real-time method for collecting a near-accurate latency information of all links. The controller only needs to collect the latency information of a small subset of significant links and then estimates the latency information of insignificant links. We use real-world network topologies to simulate multicast routing both under the compressive latency monitoring and global latency monitoring. Simulation results show that the compressive latency monitoring can achieve nearly good network throughput performance as the global monitoring with much less communication overhead and in a more timely fashion.

Data center networks employ parallel paths to perform load balancing. Existing traffic splitting schemes propose weighted traffic distribution across multiple paths via a centralized view. An SDN controller computes the traffic splitting ratio of a flow group among all the paths, and implements the ratio by creating multiple rules in the flow table of OpenFlow switches. However, since the number of rules in TCAM-based flow table is limited, it is not scalable to implement the ideal splitting ratio for every flow group. Existing solutions, WCMP and Niagara, aim at reducing the maximum oversubscription of all egress ports and reducing traffic imbalance, respectively. However, the transmission time of flow groups, which measures the quality of cloud services, is sub-optimal in existing solutions that ignore heterogeneous network bandwidth. We propose and implement NAMP, a multipathing scheme considering the network heterogeneity, to efficiently optimize the transmission time of flow groups. Experimental results show that NAMP reduces the transmission time by up to 45.4% than Niagara, up to 50% than WCMP, and up to 60% than ECMP.

A Partially Overlapped WiFi Channel (POC) is a type of WiFi channel whose spectrum is partially overlapping with other carriers. It has been empirically demonstrated that the throughput of heterogeneous cognitive network can be improved by utilizing POCs. POC detection is a prerequisite to POC utilization. Unfortunately, the existing Clear Channel Assessment (CCA) methods such as energy-based detection and preamble detection cannot accurately detect the POC in heterogeneous cognitive networks. As a result, POCs will not be used by most WiFi users. The spectrum in POCs is therefore under-utilized and wasted. In this article, we propose to detect a POC by statistically analyzing the bit-level information inside the payload of WiFi frames. The proposed approach is based on a series of measurements on bit errors under real-world IEEE 802.11ac channels. A POC can be accurately detected by analyzing the correlation between an unknown WiFi channel and a given POC in terms of their bit-error vectors. Our approach is evaluated by detecting fifty [1/2]-overlap WiFi channels among a hundred different real-world WiFi channels. The final results show that, our approach can achieve an accuracy of 96% and a false positive rate of 8% on POC detection, which is much better than the existing CCA methods.

Hybrid software-defined network (SDN) is a network where legacy routers and SDN routers coexist during the incremental deployment of SDNs. The existing SDN router placement methods mainly focus on maximizing the traffic engineering performance under a limited budget. Traffic engineering requires real-time link-load information. However, the latency for collecting the global link-load information can be prohibitively long in a wide area network due to the long interior gateway protocol convergence time. Inspired by the sparsity of the link load, we propose a novel compressive traffic monitoring method for collecting real-time load information of all links. In this method, the controller only needs to collect the load of a small subset of important links and then estimates the link load of the rest. The minimal number of SDN routers is placed to cover these important links. We use real-world topologies and traffic matrices to evaluate our method. Experiment results show that our method can quickly estimate the global link load at an error rate of 5% within sub-seconds. Compared with the state-of-the-art methods, our method has better adaptability to the dynamic traffic changes and can reduce the maximal link usage by 39%.

The wireless network that integrates a heterogeneous cellular network and a WLAN is referred to as a WLAN/cellular integrated network (WCIN). In a fourth generation WCIN, in order to achieve collision-avoidance between the LTE-A and IEEE 802.11 family, some interference-free mechanisms, such as carrier sense adaptive transmission and listen-before-talk, are proposed. However, the interference-free medium access mechanisms leave some unexploited partially overlapped 802.11ac channels, which result in a waste of spectrum. In this paper, we propose an interference-tolerant medium access method to optimize the WCIN throughput by utilizing those partially overlapped channels (POCs). First, we show the feasibility of enhancing WLAN throughput by utilizing 802.11ac POCs in a WCIN. Second, we mathematically model the partial overlap when an 802.11ac channel is partially overlapping with an LTE-A component carrier. Third, we propose an interference-tolerant medium access mechanism to optimize the WCIN throughput. The interference-tolerant one ensures that the interference to LTE-A users is tolerable in a given WCIN. Finally, we construct a hardware-in-the-loop testbed to evaluate and compare our proposed mechanism with three other state-of-the-art mechanisms. The experimental results show that our approach achieves at most 39% more throughput than one of the state-of-the-art collision-avoidance mechanisms regarding the entire WCIN.

Nowadays, many organizations or corporations, such as banks, hospitals, schools, etc., use third-party platform to send notifications or information to their customers or users. This communication model can be generalized to the publish/subscribe model. The publisher will send the content to the subscriber whenever there is any new content that matches the interest of the subscriber. However, the content may contain users' privacy which is sent in plain text by the third-party platform. When the platform is not fully trusted, users' privacy is at the risk. In this paper we propose a secure protocol to protect users' privacy against the untrusted third-party platform. Our privacy preserving protocol has three main advantages: 1) There is no need to set up a secure synchronous communication channel between the subscriber and the publisher in our protocol. 2) There is no need to introduce any Certificate Authority (CA) for certifying publishers or subscribers in our protocol. 3) There is no need for subscribers or publishers to remember any additional key for each subscription or publication, which makes our protocol scalable to large pub/sub system.

The controller placement problem tries to place k controllers in a given network to optimize performance metrics such as propagation latency, load distribution, network reliability and failure resilience. However, the quality of service (QoS) is always a primary concern of the network operators in the placement of SDN controllers. Since the SDN controllers are responsible to provide services for switches, the response time of controllers is an important QoS parameter of network operators. In this paper, we introduce the QoS-Guaranteed Controller Placement problem: Given a network topology and a response time bound, how many controllers are needed, where to place them, and which switches are assigned to each of the controller. We propose three heuristic algorithms: incremental greedy algorithm, primal-dual-based algorithm and network-partition-based algorithm. The proposed algorithms are tested and compared on the Internet Topology Zoo, a dataset of public available network topologies. Simulation results show that the proposed incremental greedy method slightly outperforms the other two methods on all input topologies.

In the current Internet architecture, application service providers (ASPs) own users' data and social groups information, which made a handful of ASP companies growing bigger and bigger and denied small and medium companies from entering this business. We propose a new architecture, called Application Independent Information Infrastructure (AI3). The design goals of AI3 are: 1) Decoupling users' data from ASPs and users' social relations from ASPs, such that ASPs become independent from users' data and social relations. 2) Open architecture, such that different ASPs can interoperate with each other. This demo is to show a prototype of AI3. The demo has four parts: 1) ASPindependent data management in AI3; 2) ASP-independent management of users' social relations in AI3; 3) inter-domain data transport and user roaming; 4) real-time communications by using AI3. The demo video can be watched at: http://www.cs.cityu.edu.hk/~jia/AI3_DemoVideo.mp4