Kleber Vieira Cardoso

Increasing throughput in dense 802.11 networks by automatic rate adaptation improvement

Rate control algorithms for commercial 802.11 devices strongly rely on packet losses for their adaptation. As a result, they give poor performance in dense networks because they are not able to distinguish packet losses related to channel error from packet losses due to collision. In this paper, we evaluate automatic rate adaptation algorithms in IEEE 802.11 dense networks. A certain number of works in the literature address this problem, but they demand modifications of the IEEE standard, or depend on some special feature not available in off-the-shelf devices. In this context, we propose a new automatic rate control algorithm which is simple, easy to implement, standards-compliant, and well-suited for crowded 802.11 networks. Our approach consists of measuring the contention level, inferring the collision probability, and choosing transmission rates which maximize throughput. Results from simulation and real experiments show throughput improvement of up to 100% from our mechanism.

Dimensioning virtualized wireless access networks from a common pool of resources

Resource sharing in mobile wireless networks has been employed to reduce costs, extend coverage, and ease the entry of new players in the market. The introduction of programmability and virtualization is expected to amplify these benefits of resource sharing. In this paper, we study a new virtualization-based paradigm for resource sharing in mobile wireless networks. Specifically, we consider the problem of resource allocation, particularly when user demands are uncertain. We formulate several two-stage sequential stochastic allocation schemes that provide tradeoffs between cost and user satisfaction. These allocation schemes are studied under different resource provider pricing models. Our simulations demonstrate that: First, while reducing cost significantly, virtualization considerably improves user satisfaction, and virtualization gains increase with the number of operators that share resources. Second, the improvements in cost, user satisfaction, and resource usage increase substantially with the level of user clustering.

Load balancing routing for path length and overhead controlling in Wireless Mesh Networks

Wireless Mesh Networks (WMN) provide ubiquitous Internet access for mobile users by integrating wired and wireless networks. As gateways towards wired networks are potential bottlenecks, load balancing routing plays a central role in the performance of WMN. With regard to this matter, a number of load balancing routing heuristics have been proposed. However, none of them tackles four key aspects of the load balancing routing problem at the same time, which are achieving low computational costs, and reducing the average path length and routing overhead, while distributing the flows uniformly. To fill this gap, the aim of this paper is to introduce a new load balancing routing heuristic, called BPR - Bottleneck, Path length and Routing overhead, which offers an efficient online solution by taking into consideration all these aspects of the load balancing routing problem. We have carried out a simulation study to compare the performance of BPR with a recent related work. BPR obtains bottleneck values within desirable bounds, while reducing the average path length. As a result, BPR also notably reduces the number of route updates in the network, i.e. the routing overhead. Finally, we show that BPR is simple and has low demand for processing requirements.

Filling the gap between Software Defined Networking and Wireless Mesh Networks

Software Defined Networking (SDN) has emerged as a new paradigm that highly increase the network management flexibility through simple but powerful abstractions. The key idea is decoupling the control plane, which makes the forward decisions, from the data plane, which effectively makes the forward. However, the OpenFlow, the main SDN enabler, is designed mainly by wired networks characteristics. As consequence, Wireless Mesh Networks (WMNs) is not suitable for operating as control plane and many wireless networks features are neglected in the OpenFlow, e.g.: power control and network ID. In addition, there are few effort research to extend SDN to wireless networks and these existing works focus on very specific issues of this integration. In this paper, we propose an architecture to extent the OpenFlow functionalities in order to proper deal with wireless networks, including an approach for transporting the control plane over wireless multihop networks. The extensions include new rules, actions, and commands, which bring the network management flexibility to the wireless context. We validated our proposal by implementing and testing some extensions in a small real world testbed. As a proof of concept, we illustrate the OpenFlow capability of isolation between research and production traffics in a wireless backhaul.

Virtualization for Load Balancing on IEEE 802.11 Networks

In IEEE 802.11 infrastructure networks composed by multiple APs, before a station can access the network it needs to make a decision about which AP to associate with. Usually, legacy 802.11 stations use no more than the signal strength of the frames received from each AP to support their decision. This can lead to an unbalanced distribution of stations among the APs, causing performance and unfairness problems. This work proposes a new approach that combines the number of associated stations and the current load of each AP plus the virtualization of client wireless interfaces. In this approach, stations frequently switch of association among APs and stay on each one of them for a time interval that is calculated based on the number of associated stations and the channel current load. Simulation results confirm the improvement obtained in the load balancing and fairness on network capacity allocation, while keeping the maximum network utilization.

On the effectiveness of push-out mechanisms for the discard of TCP packets

This paper investigates the joint use of push-out mechanisms with random early detection (RED)-like discarding policies to support service differentiation in the Internet. The efficiency and the degree of differentiation of a complete sharing with push-out, a RIO (RED with in/out) and a RIO with push-out queues are assessed. Results indicate that push-out used jointly with RED-like policies does not improve the performance.

A joint CPU-RAM energy efficient and SLA-compliant approach for cloud data centers

Cloud computing is a new paradigm that offers computing resources in a virtualized way with unprecedented levels of flexibility, reliability, and scalability. The benefits of cloud computing, however, come at a high cost in terms of energy consumption, mainly because of one of the clouds core enablers, the data center. There are a number of proposals that seek to enhance the energy efficiency of data centers. Still, most of them focus on the energy consumed by CPU and ignore other important hardware components, e.g., RAM. In this paper, we show the considerable impact that RAM can have on the total energy consumption, particularly in servers with large amounts of this memory. We then propose two new approaches for dynamic consolidation of virtual machines in cloud data centers that take into account both CPU and RAM usage. We have implemented and evaluated our proposals in the CloudSim simulator using real-world traces and compared the results with other state-of-the-art solutions. By adopting a wider view of the system, our proposals can reduce not only energy consumption but also service level agreement (SLA) violations, thus providing a better service at a lower cost.

FIT@BR - a Future Internet Testbed in Brazil

A major objective of the Brazil-EU FIBRE project is the deployment in Brazil of FIT@BR, a wide-area network testbed to support user experimentation in the design and validation of new network architectures and applications. In such a testbed, a high degree of automated resource sharing between experimenters is required, and the testbed itself must be instrumented so that precise measurements and accounting of both user and facility resources may be carried out. In this article, we describe the design and implementation of the Control and Monitoring Framework (CMF) for the FIT@BR testbed, which is based on three CMFs developed in existing testbed projects. In order to take best advantage of different testbed functionalities at different sites, FIT@BR is being created as a federated testbed, which will facilitate future interoperation with international initiatives.

Improving load balancing, path length, and stability in low-cost wireless backhauls

Futuristic wireless networks are being proposed to promote a significant improvement in performance, mainly in highly dense scenarios. However, cost constraints are also a key concern for the next generation of wireless networks. In this context, Low-cost Wireless Backhauls (LWBs) can provide relevant contributions. LWBs are based on WLAN technologies, such as Wireless Mesh Networks, in which gateways to the wired networks are potential bottlenecks, and load balancing is critical for performance. In spite of several proposals to deal with load balancing, they fail to combine three key aspects: (1) stability, (2) reduction of the average path length, and (3) throughput fairness. Our proposal addresses these aspects by adopting a joint approach for routing and channel assignment. The routing part is composed of a heuristic that employs an on-demand local solution in which load balancing is combined with the three aforementioned key aspects. We have carried out an in-depth simulation study in ns-3 to assess the impact of our proposal on traffic performance when compared with the state of the art. Our proposal is superior in most of the scenarios, in particular in terms of stability and average path length. Our proposal also increases the aggregate throughput and minimum throughput per flow in the network, especially when the number of flows is large.

Using Partial Correlation for Effective Metric Selection and Anomaly Detection in Multi-tier System

Large-scale data centers allow organizations to gain access to computer resources without incurring high costs in purchasing and maintaining IT infrastructure. In these environments, due to the large number of hardware and software involved, anomaly detection is difficult but essential for service provisioning. Computer systems hosted in data centers usually involve multiple layers and provide a large set of metrics for tracking their operation. The analysis of all available metrics generates drawbacks associated with communication, storage and processing. A more efficient way to support anomaly detection and minimize the cost of monitoring is to use stable statistical correlations among metrics that reflect the system state. We present the strategies PCTN, MST-PCTN and PCTN-MST, based on partial correlation, for selecting metrics to support anomaly detection in multi-tier systems. We evaluate the proposed strategies using an e-commerce, Web transaction benchmark. Results show that the PCTN-MST strategy allowed the construction of a monitoring network with 8% less metrics than that obtained with MST and achieved a fault coverage up to 10% larger.