Jose Moura

Review and analysis of networking challenges in cloud computing

Cloud Computing offers virtualized computing, storage, and networking resources, over the Internet, to organizations and individual users in a completely dynamic way. These cloud resources are cheaper, easier to manage, and more elastic than sets of local, physical, ones. This encourages customers to outsource their applications and services to the cloud. The migration of both data and applications outside the administrative domain of customers into a shared environment imposes transversal, functional problems across distinct platforms and technologies. This article provides a contemporary discussion of the most relevant functional problems associated with the current evolution of Cloud Computing, mainly from the network perspective. The paper also gives a concise description of Cloud Computing concepts and technologies. It starts with a brief history about cloud computing, tracing its roots. Then, architectural models of cloud services are described, and the most relevant products for Cloud Computing are briefly discussed along with a comprehensive literature review. The paper highlights and analyzes the most pertinent and practical network issues of relevance to the provision of high-assurance cloud services through the Internet, including security. Finally, trends and future research directions are also presented.

Efficient access of mobile flows to heterogeneous networks under flash crowds

Future wireless networks need to offer orders of magnitude more capacity to address the predicted growth in mobile traffic demand. Operators to enhance the capacity of cellular networks are increasingly using WiFi to offload traffic from their core networks. This paper deals with the efficient and flexible management of a heterogeneous networking environment offering wireless access to multimode terminals. This wireless access is evaluated under disruptive usage scenarios, such as flash crowds, which can mean unwanted severe congestion on a specific operator network whilst the remaining available capacity from other access technologies is not being used. To address these issues, we propose a scalable network assisted distributed solution that is administered by centralized policies, and an embedded reputation system, by which initially selfish operators are encouraged to cooperate under the threat of churn. Our solution after detecting a congested technology, including within its wired backhaul, automatically offloads and balances the flows amongst the access resources from all the existing technologies, following some quality metrics. Our results show that the smart integration of access networks can yield an additional wireless quality for mobile flows up to thirty eight percent beyond that feasible from the best effort standalone operation of each wireless access technology. It is also evidenced that backhaul constraints are conveniently reflected on the way the flow access to wireless media is granted. Finally, we have analyzed the sensitivity of the handover decision algorithm running in each terminal agent to consecutive flash crowds, as well as its centralized feature that controls the connection quality offered by a heterogeneous access infrastructure owned by distinct operators.

Successful Cooperation between Backoff Exponential Mechanisms to Enhance IEEE 802.11 Networks

The apparently never ending widespread of wireless networks requires an enhanced multimedia application support. In this paper, we show how to improve the performance of EDCA in terms of throughput, frame delay and station scalability, through the effective cooperation of two backoff mechanisms: EDM and CWD. Simulation results confirm our previous analytical study and enable the discussion about how to improve relevant Quality of Service metrics, considering different channel bit rate and network load.

What's DECR impact on wireless communications?

In this paper, we study how DECR influences delay on IEEE 802.11 networks and also how they are affected by frame size variation. Impact on some relevant network parameters are studied, namely throughput, idle time, collision time and cumulative fraction function for delay. Several results, obtained from an extensive number of simulations, are presented and discussed, comparing DECR with other MAC proposals, for different scenarios varying parameters like, traffic load, channel bit rate and frame size. Results show that DECR outperforms other proposals, using available wireless channel resources more efficiently.