Reinaldo Gomes

Efficient network service level agreement monitoring for cloud computing systems

Cloud computing defines an infrastructure to provide virtual network services on demand. One of the challenges to cloud computing systems is how to manage service level agreements (SLA) while guaranteeing the service level objectives (SLOs) negotiated between provider and customer. Deviations in the expected behavior of network service levels may be generically named traffic anomalies. The awareness of anomalies in cloud services may provide useful information for both customer and provider, in order to support management accounting to their business strategies. In this paper we present a novel approach to monitor and account network SLA deviations by designing a new architecture, as well as implementing and deploying it as an open source engine. It has been integrated into a production middleware to federate clouds, the Jit Clouds. Nevertheless, this monitoring mechanism is loosely coupled, being extensible to any sort of clouds. The architecture proposed was validated on a federated cloud platform in production.

VANETs' research over the past decade: overview, credibility, and trends

Since its inception, Vehicular Ad hoc Networks (VANETs) have been attracting much attention from both academia and industry. As for other wireless networking areas, scientific advancements are mainly due to the employment of simulation tools and mathematical models. After surveying 283 papers published in the last decade on vehicular networking, we pinpoint the main studied topics as well the most employed tools, pointing out the changes in research subject preference over the years. As a key contribution, we also evaluate to what extent the research community has evolved concerning the principles of credibility in simulation-based studies, such as repeatability and replicability, comparing our results with previous studies.

Evaluating live virtual machine migration overhead on client's application perspective

A key feature on virtualization technology is live migration. Through live migration, it is possible to exchange virtual server hosts between physical machines without stopping its services and with low downtime during migration. However, virtualization also has some drawbacks, for instance, the overhead generated from the virtualization layer. Many research evaluated overhead effect in different scenarios, but few investigated overheads in the migration process, on clients side. Our research evaluated live migration process overhead, on real and virtual environments, noticed from the client side regarding two different services: web server and database. YCSB and ab Benchmark were adopted as workloads for those services. In the findings, based on our experimental setting, we could observe that almost all tests on real environment overcame those on virtual, with both benchmarks. The impact of the live migration in the services was evident, proving to be more effective in real machines than on virtual machines. We found the DB service accommodated better to the virtual environment and to migration then Web service.

Evaluating performance degradation in NoSQL databases generated by virtualization

There are many services in Cloud Systems currently on the market for hosting servers with different purposes. These environments are IT infrastructures frequently deployed in large data centers using Virtual Machines. Moreover, an increasing number of users are using massive online application resources with hosted databases on virtual machines. In light of this, there is a critical need to evaluate databases performance in virtual environments. This work aims to evaluate overhead generated by virtualization in two common NoSQL databases, Cassandra and MongoDB, by different virtualization techniques - full virtualization and paravirtualization. For this, we used the YCSB benchmark (Yahoo! Cloud Serving Benchmark) to drive performance tests, with the results being evaluated through statistical analysis. As findings, the performed experiments in both virtualization techniques demonstrated that the experimental design comprises most of the factors in MongoDB, different from Cassandra which comprises of just a few. In paravirtualization scenarios the environment factor was more sensitive to experimental variations overcoming factors such as threads and numbers of transaction. Both databases in full virtualization scenarios reached significant variation, however the threads and transactions factors were more significant.

IP traffic classifiers applied to DiffServ networks

The future Internet scenario consists of a higher number of users and applications, which demand more resources from the communication infrastructure. Techniques for providing performance and scalability, such as Traffic Engineering (TE), will always be necessary even if the transmission rate is very high, because of such demands. Quality of Service is one of the solutions that can be used to improve the traffic engineering in the Internet, with the most referenced architecture: DiffServ. In general, TE needs traffic classification to accurately identify the input traffic and manage it properly. However, the current DiffServ port traffic classifier is considered outdated. This paper presents a performance evaluation of machine learning traffic classification solutions applied to DiffServ, and investigates their benefits on network performance. For a backbone network with 40 nodes, the performance of the network can increase up to 15% for both data and voice traffic.

A Top-down Multi-layer Routing Architecture for Vehicular Ad-Hoc Networks

The actual proposed routing protocols for VANETs (Vehicular Ad-hoc Networks) present different features for communication among hosts/vehicles considering the strong topology change, but most of these features are needed for routing in these specific networks. These routing protocols support vehicle traffic on a large scale, intense mobility of vehicles, connections without link breakage, etc. But as they are different protocols the routers (nodes) have to switch to a routing protocol in a certain moment, which is a problem. This paper presents a routing architecture for VANETs to face it. The most important technical features for routing in VANETs were grouped in the Routing Architecture. To validate the proposed architecture several existing protocols were unified in the architecture producing a new routing protocol for VANETs. The produced protocol is the Generic Vehicular Dynamic Source Routing (GVDSR). Simulations of the GVDSR protocol have been made on the Malaga city showing the contributions and advantages for routing performance. The proposed architecture and protocol were simulated in the Network Simulator 2 featuring better performance than the compared protocolsThe actual proposed routing protocols for VANETs (Vehicular Ad-hoc Networks) present different features for communication among hosts/vehicles considering the strong topology change, but most of these features are needed for routing in these specific networks. These routing protocols support vehicle traffic on a large scale, intense mobility of vehicles, connections without link breakage, etc. But as they are different protocols the routers (nodes) have to switch to a routing protocol in a certain moment, which is a problem. This paper presents a routing architecture for VANETs to face it. The most important technical features for routing in VANETs were grouped in the Routing Architecture. To validate the proposed architecture several existing protocols were unified in the architecture producing a new routing protocol for VANETs. The produced protocol is the Generic Vehicular Dynamic Source Routing (GVDSR). Simulations of the GVDSR protocol have been made on the Malaga city showing the contributions and advantages for routing performance. The proposed architecture and protocol were simulated in the Network Simulator 2 featuring better performance than the compared protocols.

On the evaluation of self-addressing strategies for ad-hoc networks

Ad-hoc networks are supposed to operate autonomously and, therefore, self-* technologies are fundamental to their deployment. Several of these solutions have been proposed during the last few years, covering most layers and functionalities of networking systems. Addressing is one of the critical network operations on supporting others such as routing and even security of IP-based communcations. This paper has the goal of putting together different strategies for self-addressing problem by evaluating five self-addressing protocols and pointing out their strenghs and drawbacks. At this time, we did not intend to come up with a novel approach for the problem, but to evaluate the existing ones in a non-isolated manner and also considering critical ad-hoc networking situations of partition and merging. Conclusions about the evaluated approaches and their applicability are drawn at the end of this paper.

Enabling high-level network programming: A northbound API for Software-Defined Networks

Software-Defined Networking (SDN) is a new paradigm that is attracting significant attention from both academy and industry. In recent years, increasing research efforts have addressed the conception of programming interfaces—northbound APIs—which allow business-applications to properly communicate with SDN controllers and program the network. However, most of current SDN controllers still offer low-level northbound interfaces based on the OpenFlow protocol, which incurs an undesirable but common coupling between network application and controllers code. As software developers often have to insert intricate controller-specific instructions directly into the applications code, this dramatically impairs its reusability and maintainability. In this paper, we propose an open northbound API that enables streaming video applications to easily enforce Quality of Service (QoS) requirements in a high-level fashion, without incurring any controllers code coupling or network operator intervention. The results of a statistical analysis show that our solution can increase the average bandwidth rate in comparison to traditional command-line-driven management approaches. In particular, a simulation study is conducted on a single-controlled SDN architecture using Floodlight controller, where our northbound API performs well, meeting quality of service requirements at runtime with a high statistical confidence.

Evaluation of performance and scalability of routing protocols for VANETs on the Manhattan Mobility Model

Routing plays an important role in VANETs, even more than conventional mobile ad-hoc networks, because of the high mobility, high density and high velocity of nodes/vehicles on this scenario. Thus, actual research effort is invested on the design of new routing protocols for these networks. The problem is that no considerable previous evaluation of routing was made in VANETS. This must be accurately enough to compare fairly new protocols for VANETs with ad-hoc existing routing ones. This paper presents such evaluation considering an applicable mobility model with velocity and number of nodes analysis.