Under-provisioned Network Research Articles

A new genetic-based approach for solving [formula omitted]-coverage problem in directional sensor networks

In recent years, the use of directional sensor networks (DSNs) has continued to rise increasingly, which is due to their extensive use in many situations. In such networks, the main problem is how to cover the targets distributed in a defined area and simultaneously prolong the network lifetime as much as possible. The reason of this problem is the limitation of both sensing angle and energy resource of sensors in such networks. This problem gets more complex in cases where targets need to be covered by more than one sensor direction (i.e., each target needs to be monitored for at least k times). This problem is generally known as target k-coverage problem which its NP-completeness has been already proved in the literature. The k-coverage problem can be considered in over-provisioned and under-provisioned environments. In both of these environments, especially in the latter, it is important to create a balanced coverage, as these environments do not have enough sensors to monitor all targets for k times. Thus, in this paper, we proposed a genetic-based algorithm to solve the problem in over-provisioned environments, then developed the proposed algorithm in another way to solve the problem in under-provisioned networks so that it uses the minimum number of sensors. many experiments were performed to test the efficiency of the proposed algorithm, and the obtained results showed the high capacity of the proposed algorithm in solving the k-coverage problem in both environments.

An SDN-based framework for improving the performance of underprovisioned IP Video Surveillance networks

In the coming years, the dominance of video in global Internet traffic is expected to intensify due to the ongoing migration from analog CCTV to IP-based surveillance systems. As opposed to analog, IP cameras can be integrated into pre-existing computer networks and thus, are more cost-effective. However, since video is bandwidth-intensive, contention among the streams from multiple IP cameras and traffic from other applications results to a severely degraded network performance. Software-Defined Networking (SDN) is a relatively new paradigm which aims to build dynamically configurable networks. The decoupling of forwarding and control functions in SDN architecture enables a centralized controller to create a map of the network topology by utilizing the information collected from the switches. This paper proposes an SDN-based framework to enhance the performance of IP Video Surveillance (IPVS) systems deployed over underprovisioned networks. As a means of allowing the controller to infer a stream's QoS metrics and to execute bitrate adjustment or rerouting as necessary, two video quality indicators were also formulated by utilizing the statistics messages available in OpenFlow, a widely recognized protocol in SDN. Moreover, our design is built on the idea that the video streams in surveillance systems offer different utilities depending on the captured event. Experimental results show that employing the proposed framework improved the video streams' overall packet loss, latency, jitter and throughput by 88%, 36%, 11% and 5% respectively. The comparison of the video streams' QoS metrics also suggests that the framework is capable of prioritizing the reception of selected streams. Furthermore, we also demonstrate that the proposed framework can be easily extended to handle the case of an IPVS system wherein the streams are subjected to dynamic priority assignment.

Maximizing heterogeneous coverage in over and under provisioned visual sensor networks

We address “heterogeneous coverage” in visual sensor networks where coverage requirements of some randomly deployed targets vary from target to target. The main objective is to maximize the coverage of all the targets to achieve their respective coverage requirement by activating minimal sensors. The problem can be viewed as an interesting variation of the classical Max-Min problem (i.e., Maximum Coverage with Minimum Sensors (MCMS)). Therefore, we study the existing Integer Linear Programming (ILP) formulation for single and k-coverage MCMS problem in the state-of-the-art and modify them to solve the heterogeneous coverage problem. We also propose a novel Integer Quadratic Programming (IQP) formulation that minimizes the Euclidean distance between the achieved and the required coverage vectors. Both ILP and IQP give exact solution when the problem is solvable but as they are non-scalable due to their computational complexity, we devise a Sensor Oriented Greedy Algorithm (SOGA) that approximates the formulations. For under-provisioned networks where there exist insufficient number of sensors to meet the coverage requirements, we propose prioritized IQP and reduced-variance IQP formulations to capture prioritized and group wise balanced coverage respectively. Moreover, we develop greedy heuristics to tackle under provisioned networks. Extensive evaluations based on simulation illustrate the efficiency and efficacy of the proposed formulations and heuristics under various network settings. Additionally, we compare our methodologies and algorithm with two state-of-the-art algorithms available for target coverage and show that our methodologies and algorithm substantially outperform both the algorithms.

On balanced k-coverage in visual sensor networks

Given a set of directional visual sensors, the k-coverage problem determines the orientation of minimal directional sensors so that each target is covered at least k times. As the problem is NP-complete, a number of heuristics have been devised to tackle the issue. However, the existing heuristics provide imbalance coverage of the targets—some targets are covered k times while others are left totally uncovered or singly covered. The coverage imbalance is more serious in under-provisioned networks where there do not exist enough sensors to cover all the targets k times. Therefore, we address the problem of covering each target at least k times in a balanced way using minimum number of sensors. We study the existing Integer Linear Programming (ILP) formulation for single coverage and extend the idea for k-coverage. However, the extension does not balance the coverage of the targets. We further propose Integer Quadratic Programming (IQP) and Integer Non-Linear Programming (INLP) formulations that are capable of addressing the coverage balancing. As the proposed formulations are computationally expensive, we devise a faster Centralized Greedy k-Coverage Algorithm (CG kCA) to approximate the formulations. Finally, through rigorous simulation experiments we show the efficacy of the proposed formulations and the CG kCA.

Optimal Delivery of Rate-Adaptive Streams in Underprovisioned Networks

The growth of Internet video traffic imposes a severe capacity problem in today's Content Delivery Network (CDN). Rate-adaptive streaming technologies, such as the Dynamic Adaptive Streaming over HTTP (DASH) standard, reinforces this problem in the core CDN infrastructure since delivering one video means delivering multiple representations for an aggregated bit-rate that is commonly over 10 Mbps. In this paper, we explore better trade-offs between CDN infrastructure cost and Quality of Experience (QoE) of the end-users for live broadcast video streaming applications. We consider in particular underprovisioned CDN networks, our goal being to maximize the QoE for the population of heterogeneous end-users despite the lack of resources in the intermediate CDN equipments. We show that previous theoretical models based on elastic bit-rates do not fit for this context. We propose a user-centric discretized streaming model where the satisfaction of end-users is related to the context and where a stream has to be either delivered in its entirety, or not delivered at all. We first formulate an Integer Linear Program (ILP) that achieves the optimal delivery through a multi-tree delivery overlay. The evaluation of the ILP shows the benefits of this model. We then design a practical system by revisiting the three main algorithms implemented in CDN: user-to-server assignment, content placement and content delivery. At last, we use a realistic trace-driven large-scale simulator to study the performances of our system. In particular, we show that the population of users is reasonably well served (three quarters of the population do not experience degradation) even when the CDN infrastructure experiences a severe underprovisioning (less than half of the required infrastructure).

A provision-aware fair bandwidth distribution marker algorithm for DiffServ networks

The rise in demand for real-time applications on the Internet necessitates Quality of Service (QoS). Differentiated Services (DiffServ) is one of the technologies used currently to provide QoS and service differentiation. It is simple and scalable. It provides service differentiation to aggregates, mainly through the Assured Forwarding (AF) per-hop behaviour. Previous work on fair sharing of network bandwidth did not adequately address the Under-Provisioned Network (UPN) condition. In this paper, we propose a new three-colour marker, named paItswTCM (provision-aware Improved TSW based Three-Colour Marker). We compare our new algorithm with both time-sliding window markers and token-bucket-based markers using simulations. Results show that our new provision-aware marker outperforms these previous algorithms not only in the UPN condition but also for low to medium network provision levels. We conclude that to achieve proportional sharing of bandwidth, no packet type should be injected at the expense of others.

A fair bandwidth distribution mechanism in a Diffserv network

Previous works for the Assured Forwarding (AF) service in a Diffserv network have no sufficient consideration on the target rate and the effect of RTT and UDP in achieving the fairness of bandwidth share. Also previous works act like Best-effort service in an Under-Provisioned Network (UPN) condition. In this paper, in order to solve these problems, we propose the Proportionally Fair Differentiated Service Architecture (PFDSA), which is composed of tcp-microflow based Target rate and RTT Aware Three Color Marker (tmTRA3CM), udp-microflow based Three Color Marker (um3CM), Target Rate Based Dropper (TRBD), and Target Rate Adjusting Function (TRAF). In the results of comparisons on the performance among existing mechanisms and the PFDSA in the OPN (Over-Provisioned Network) and the UPN conditions, the PFDSA was able to further mitigate the RTT and UDP effects than the previous works. The PFDSA was shown to provide better fair bandwidth distribution proportional to various target rates even in the UPN condition.

The two markers system for TCP and UDP flows in a differentiated services network

There are two cruxes of issues identified in differentiated services (Diffserv) networks: one is TCP dynamics, and the other is the interaction of TCP and UDP flows for assured forwarding per hop behaviour (PHB). The saw-tooth like behaviours of TCP prevent each flow from realising its target rate, and congestion insensitive flows such as UDP penalise for TCP flows when both are given same AF class. Therefore, we argue the use of TCP-friendly building blocks and fairness modules in the Diffserv architecture regarding these issues, and propose two markers system (TMS) that is able to properly mark packets and fairly share the bandwidth to each flow for their targeted sending rates. TMS has two marking modules that are placed on the source and the edge of a Diffserv network, respectively. For sources of the network, TCP-source making module plays important roles of reducing TCP impacts in the assured services and adequately marking packets. Next, in the edge of the network, the edge-embedded marking module conducts new fairness policy based on the marking rate of flows from sources, so-called ‘marking rate-based fairness’. Finally, we present simulation results to illustrate the effectiveness of TMS scheme over several parameters. That is, TMS alleviates TCP impacts over assured service and the issue of unfair distribution of excess bandwidth in an over-provisioned networks as well as unfair degradation in an under-provisioned network for TCP and UDP flows.