Deep neural networks (DNNs) typically have a single exit point that makes predictions by running the entire stack of neural layers. Since not all inputs require the same amount of computation to reach a confident prediction, recent research has focused on incorporating multiple “exits” into the conventional DNN architecture. Early-exit DNNs are multi-exit neural networks that attach many side branches to the conventional DNN, enabling inference to stop early at intermediate points. This approach offers several advantages, including speeding up the inference process, mitigating the vanishing gradients problems, reducing overfitting and overthinking tendencies. It also supports DNN partitioning across devices and is ideal for multi-tier computation platforms such as edge computing. This article decomposes the early-exit DNN architecture and reviews the recent advances in the field. The study explores its benefits, designs, training strategies, and adaptive inference mechanisms. Various design challenges, application scenarios, and future directions are also extensively discussed.
Many corporations rely on disaster recovery schemes to keep their computing and network services running after unexpected situations, such as natural disasters and attacks. As corporations migrate their infrastructure to the cloud using the infrastructure as a service model, cloud providers need to offer disaster-resilient services. This article provides guidelines to design a data center network infrastructure to support a disaster-resilient infrastructure as a service cloud. These guidelines describe design requirements, such as the time to recover from disasters, and allow the identification of important domains that deserve further research efforts, such as the choice of data center site locations and disaster-resilient virtual machine placement.
30 September 2015 in Journal of Network and Systems Management
In an Internet of Things (IoT) environment, sensors collect and send data to application servers through IoT gateways. However, these data may be missing values due to networking problems or sensor malfunction, which reduces applications’ reliability. This work proposes a mechanism to predict and impute missing data in IoT gateways to achieve greater autonomy at the network edge. These gateways typically have limited computing resources. Therefore, the missing data imputation methods must be simple and provide good results. Thus, this work presents two regression models based on neural networks to impute missing data in IoT gateways. In addition to the prediction quality, we analyzed both the execution time and the amount of memory used. We validated our models using six years of weather data from Rio de Janeiro, varying the missing data percentages. The results show that the neural network regression models perform better than the other imputation methods analyzed, based on the averages and repetition of previous values, for all missing data percentages. In addition, the neural network models present a short execution time and need less than 140 KiB of memory, which allows them to run on IoT gateways.
A cost-effective approach to gather information in a smart city is to embed sensors in vehicles such as buses. To understand the limitations and opportunities of this model, it is fundamental to investigate the spatial coverage of such a network, especially in the case where only a subset of the buses have a sensing device embedded. In this paper, we propose a model to select the right subset of buses that maximizes the coverage of the city. We evaluate the model in a real scenario based on a large-scale dataset of more than 5700 buses in the city of Rio de Janeiro, Brazil. Among other findings, we observe that the fleet of buses covers approximately 5655 km of streets (approximately 47% of the streets) and show that it is possible to cover 94% of the same streets if only 18% of buses have sensing capabilities embedded.