ABSTRACT. Bandwidth reservation has been increasingly used to provide QoS for various network applications. To accommodate a high-priority bandwidth reservation request (BRR), the bandwidth scheduler sometimes needs to preempt existing bandwidth reservations that have been made for BRRs with a lower priority, which is traditionally known as connection preemption. When such preemption is unavoidable, one primary goal of bandwidth scheduling is to minimize the disruption to existing reservations. In this paper, we study the problem of bandwidth reservation preemption for two types of BRRs, bandwidth- and data transfer-oriented, respectively, on one given link of the scheduling network with two different objectives: (i) minimize the number and then the total bandwidth of existing bandwidth reservations to be preempted, and (ii) minimize the total bandwidth and then the number of existing bandwidth reservations to be preempted. We prove these four problems to be NP-complete and propose a heuristic algorithm for each. We also design baseline heuristic algorithms for performance comparison. Extensive simulation results show that the proposed heuristic algorithms outperform those in comparison.

ABSTRACT. Different end-to-end Transmission Control Protocol (TCP) algorithms widely in use behave differently under network congestion. The TCP congestion control itself has grown increasingly complex which in practice makes predicting TCP per-connection states from passive measurements a challenging task. In this paper, we present a robust, scalable and generic machine learning-based model which may of interest for network operators that experimentally infers the underlying variant of loss-based TCP algorithms within a flow from passive traffic measurements collected at an intermediate node. We believe that our study has also a potential benefit and opportunity for researchers and scientists in the networking community from both in academia and industry who want to assess the characteristics of TCP transmission states related to network congestion.We validate the robustness and scalability approach of our prediction model through several controlled experiments. It turns out, surprisingly enough, that the learned prediction model by leveraging knowledge from the emulated network performs reasonably well when it is applied on a real-life scenario setting bearing similarity to the concept of transfer learning in the machine learning community. The accuracy of our experimental results both in an emulated network, realistic and combined scenario settings demonstrate that our model is effective and has considerable potential.

ABSTRACT. Short connections are generally used to transfer small-size messages, which contribute a large part of workload in modern applications. The maximum sustainable short connection rate, which is called short connection capacity, is an important index for admission control, Web QoS control, and energy saving. A capacity estimation mechanism aims to find the workload just saturating the server, and it relies on both workload information and system information. Past researches point out that kernel space network stack becomes the bottleneck when a huge number of concurrent short connections coexist. On the other hand, high performance user level network stacks have been proved to eliminate such bottleneck, thus become hotspots in both academia and industry. However, they also bring challenges for estimating short connection capacity, making traditional methods ineffective. Therefore, it is important to find a new method to estimate short connection capacity in high performance user level network stacks.

In this paper, we prove that the effective CPU utilization is an adaptive index to different workload patterns and application complexities, which can reflect the server state. Then we design and implement an online capacity estimator on the Seastar platform. We conduct experiments to verify the effectiveness of our online capacity estimator. The results show that our estimator can actually estimate the capacity online. When server is near saturated, the 90th percentile relative estimating error is no more than 9.18%. Furthermore, our capacity estimator only introduces no more than 1.38% of capacity loss in our experiments.

ABSTRACT. As the cornerstone of future network research, network traffic classification plays an important role on network management, cyberspace security and quality of service. Recently, many researches have used Machine Learning technologies for traffic classification. Most of them only focus on classifying the samples into predefined classes but ignoring the ``strangers''. In this paper, we use stranger to represent the traffic not belonging to any predefined application, and propose a novel scheme named FlowCop to achieve stranger detection in network traffic classification. By constructing multiple one-class classifiers, FlowCop can divide testing traffic into N classes and a stranger class. Since samples of stranger class are not required during the training stage, FlowCop works in an inexperienced way to detect strangers, just like the cops searching the crowd for strangers. Besides, for accurate classification and low overhead, a feature subspace algorithm is proposed to select outstanding features for each one-class classifier. To verify the effectiveness of FlowCop, we make contrast experiments on two real-world datasets. The results show that FlowCop can not only identify the predefined traffic flows but also detect the strangers. It outperforms four state-of-the-art approaches on both precision and recall.

ABSTRACT. Redundant data transmission, which is very common in computer networks, degrades bandwidth efficiency and wastes energy. To reduce such redundancy, many redundancy elimination (RE) techniques have been proposed. Most of them require strict history synchronization between the sender and receiver, and therefore assume the existence of bidirectional links to maintain the synchronization. This paper presents MinMax, a new RE mechanism that is specially designed for unidirectional lossy links. Using an existing fingerprinting algorithm, MinMax eliminates a region of payload in an outgoing packet only when the region is common in a certain number of previous packets. Thus over a lossy link, MinMax significantly reduces decompression failures at the receiver without resorting to any forms of feedback and retransmission. Experiments on manipulated data and real-world traffic show that MinMax is able to maintain fair bandwidth savings while inducing negligible decompression failures even when the packet loss rate of the link is as high as 20%.

ABSTRACT. Online Social networks (OSNs) are now one of the main resources for people to keep abreast of current news and to exchange opinions about new products and social trends, etc. However, unethical use of OSNs also provides a convenient conduit to the diffusion of malicious rumors and misinformation, thus it is of significant importance to discover rumor diffusion and detect the rumor source. This is a very challenging task, as shown in many existing works, e.g., even in the regular tree graphs, the accuracy of detecting the information source from a diffusion snapshot cannot exceed 31\%. To overcome this issue, in this work, we propose a novel system framework for information source detection in OSNs and investigate a new rumor source detection problem, called $K$-Minimum Distance Rumor Source Detection ($k$-MDRSD). Specifically, given a rumor spreading snapshot, our target is to find a small set of rumor candidates which can be used as initial seeds for further iterative query or investigation. To this end, we introduce a notion, called distance error, for rumor candidate sets and formulate the $K$-MDRSD problem. Resorting to methods from Combinatorics, we develop a near optimal algorithm for $K$-MDRSD. By experimental simulation, we show that the proposed $k$-MDRSD significantly improves the likelihood of detecting rumor sources or trend setter in OSNs.

ABSTRACT. Social sensing has emerged as a new sensing paradigm where humans (or devices on their behalf) collectively report measurements about the physical world. This paper focuses on an optimized task allocation problem in multi-attribute social sensing applications where the goal is to effectively allocate the tasks of collecting multiple attributes of the measured variables to human sensors while respecting the application's budget constraints. While recent progress has been made to tackle the optimized task allocation problem, two important challenges have not been well addressed. The first challenge is "online task allocation": the task allocation schemes need to respond quickly to the potentially large dynamics of the measured variables (e.g., temperature, noise, traffic) in social sensing. Delayed task allocation may lead to inaccurate sensing results and/or unnecessarily high sensing costs. The second challenge is the "multi-attribute constrained optimization": minimizing the overall sensing error given the dependencies and constraints of multiple attributes of the measured variables is a non-trivial problem to solve. To address the above challenges, this paper develops an Online Optimized Multi-attribute Task Allocation (OO-MTA) scheme inspired by techniques from machine learning and information theory. We evaluate the OO-MTA scheme using an urban sensing dataset collected from a real-world social sensing application. The evaluation results show that OO-MTA scheme significantly outperforms the state-of-the-art baselines in terms of the sensing accuracy.

ABSTRACT. Peer-to-peer(P2P) social networks rely on voluntary resource contributions of peers, understanding and maximizing the effects of resource allocation mechanisms on resource contribution of peers have been a focus in social networks. In most of proposed research, the resource sharing dilemma is always modeled by a two-player donor-recipient game in which peers are limited to binary decision (e.g., contribute or not). However, in addition to contributing to multiple recipients simultaneously, a peer also can determine its contribution level in networks. In this paper, we first formulated the resource sharing transaction among a group of peers as a multi-player donor-recipient game with multiple strategies which signify contribution willingness of peers. Then, we studied the influences of two reciprocity based allocation mechanisms in which peers are served based on their total and direct contributions, on the evolution of peers' contribution strategies. Moreover, the influences of some common behaviors of peers (e.g., leave-rejoin and irrational behaviors, slandering behaviors in reporting others' contribution) are also studied. The research is expected to provide valuable information for resource allocation mechanism design in social networks.

ABSTRACT. While communication technologies are evolving rapidly, there is still the nontrivial matter of a communication systems being green. Although some energy-aware solutions have been proposed for the telecommunications sector, they are not designed with the ultimate goal of being environment-friendly. In this paper, we investigate the problem of achieving energy efficiency in IP networks by taking into account not only the energy consumption of the network but also the impact of various energy sources, e.g., renewable energies. We propose a new green networking approach in which we classify network nodes into clusters and select one header node in each cluster according to the energy generation cost and the carbon emission per unit of energy. We develop a routing scheme using IP routing only on header nodes and conducting packet forwarding using a carefully designed identifier on other nodes to achieve a greener communication system. We validate our solution with a variety of simulations using real-world renewable energy statistics, and the results show that our approach is superior to other existing solutions, particularly in terms of energy and cost efficiency.

ABSTRACT. Ternary content addressable memories (TCAMs) have been used to implement high-speed regular expression (Regex) matching for deep packet inspection. However, one major drawback of TCAMs is their high power consumption, which is becoming critical with the increasing number of Regex patterns. In this paper, we present GreenCAM, a power-efficient approach to TCAM-based Regex matching for reducing the power consumption of TCAMs. We introduce a simple but effective idea of separating input characters from states in TCAMs to develop a three-stage architecture for low-power Regex matching. For this architecture, we propose two techniques of character indexing and table partitioning to minimize the number of TCAM blocks activated for each transition lookup in average and worst cases. Experiments on real Regex pattern sets show that GreenCAM requires only two active TCAM blocks per transition lookup on average, and achieves significant power reductions of up to several orders of magnitude as well as higher matching throughput compared to previous schemes.

ABSTRACT. Software-defined Networking (SDN) brings new vitality to traditional network technology as its nice property of network programmability makes our network more open and flexible. By using interfaces of SDN controllers, different applications with diverse network functions can deploy their needed flow rules into SDN switches. However, some of these flow rules would probably produce conflicts that result in invalidation of network functions and cause security issues. To address this issue, we design a novel approach, Transaction-based flow rule Conflict Detection and Resolution (TCDR), which can isolate the flow rules of different network functions to avoid interference between different network functions. Meanwhile, our proposed method introduces a transaction-based authentication to guarantee the legality of flow rules. Finally, we implement a prototype of our solution, and evaluate its effectiveness and efficiency. The performance evaluation shows that TCDR can reject illegal flow rules and avoid many flow rule conflicts with a small overhead.

ABSTRACT. Extremely dense wireless network topologies gradually become a reality, especially through wireless sensors networks and more recently nanonetworks. Electromagnetic nanonetworks are expected to allow a very large amount of extremely small and capability-limited devices to communicate with each others. In nanonetworks, even in a communication range limited to tens of centimeters, thousands of neighbors can be found. Information diffusion and routing protocols would greatly benefit from having an accurate estimation of the density of nodes. However, in this context, most traditional wireless communication protocols are not suited. We propose Density Estimator for Dense Networks (DEDeN), a distributed algorithm able to provide the required density estimation. It allows confidence tuning and can cope with an extreme range of local densities. A formal analysis of DEDeN is provided and corroborated by extensive simulations. DEDeN interest is then demonstrated through application to two information diffusion protocols tailored for very dense networks, and also to a routing protocol specific to nanonetworks.

ABSTRACT. Countermeasure selection is a major part of the Intrusion Response System (IRS). Many cost-sensitive schemes have been proposed to select the optimal countermeasure to maximize security utility by attuning attack damage and response cost. However, existing schemes for IRS are not suitable to select combined countermeasures as they ignore the interaction between different countermeasures for multiple attack paths. To address this problem, in this paper, we propose a combined countermeasures selection scheme based on probabilistic attack tree (PAT). Firstly, we employ Bayesian networks to calculate the confidence of atomic attacks in the PAT. Secondly, the occurrence probability of each attack path is calculated and multiple possible attack paths are identified. Thirdly, we quantify the damage of the identified attack path and formulate the countermeasure selection for single attack path as a multi-objective optimization problem. Finally, given the security utilities of the countermeasures for different attack paths, a greedy strategy is designed to select the combined countermeasures and maximize overall security utility. The experimental results demonstrate the effectiveness of the proposed scheme.

ABSTRACT. Limited battery capacity has been the main bottleneck for smartphones. Users are required to charge their smartphones frequently to keep them alive. Access to a charging facility, however, may not be possible especially when users are outside. This has caused users to charge their devices at every opportunity with as much power as possible. While this results in overcharging of devices unnecessarily, it might have brought an opportunity for the realization of power sharing among mobile devices. In this paper, we introduce the concept of crowdcharging which aims to provide mobile users with ubiquitous power access through crowdsourcing. We first discuss the feasibility of crowdcharging from users' perspective and present some analysis and survey results showing the interest and need. We then look at the software and hardware challenges to build such a system. To this end, we have developed a mobile app that builds a mobile social network environment among the users and manages the entire process of power sharing between the mobile devices. We present the software implementation details using P2P wireless energy sharing and provide initial lab results with actual wireless charging hardware.

ABSTRACT. The drawbacks encountered in today’s cloud computing infrastructures have led to a paradigm shift towards in-network processing, where resources in the core and at the edge of the network are leveraged to perform computations. This can lead to decreased costs and better quality of service for users, e.g., when latency-critical applications are executed close to data sources and users. Deploying applications or parts thereof on these infrastructures requires to place operators (i.e., functional components of applications) on available resources in the network. Solving large instances of this problem in an optimal way is known to be computationally hard and, thus, practically unfeasible. While other heuristic approaches exist, they are mostly aimed at placing functionalities on homogeneous nodes or make unrealistic assumptions for edge computing environments. To address this issue, this paper is the first to study the placement problem in the context of a 3-tier architecture consisting of cloud, fog and edge devices. We provide a comprehensive model and propose a heuristic approach to the problem, in which we introduce constraints on the placement decision to limit the possible solution space, leading to a decrease in the solving time for the problem. These constraints exploit the characteristics of our 3-tier network architecture. To demonstrate the feasibility of the approach, we present a general framework that supports different types of heuristics. We validate the approach by implementing example heuristics for each type. We show that our approach can scale to large instances, i.e., it can significantly reduce the resolution time to find a placement solution while introducing only a small optimality gap.