Multi-access edge computing (MEC) is a new leading technology for meeting the demands of key performance indicators (KPIs) in 5G networks. However, in a rapidly changing dynamic environment, it is hard to find the optimal target server for processing offloaded tasks because we do not know the end users’ demands in advance. Therefore, quality of service (QoS) deteriorates because of increasing task failures and long execution latency from congestion. To reduce latency and avoid task failures from resource-constrained edge servers, vertical offloading between mobile devices with local-edge collaboration or with local edge-remote cloud collaboration have been proposed in previous studies. However, they ignored the nearby edge server in the same tier that has excess computing resources. Therefore, this paper introduces a fuzzy decision-based cloud-MEC collaborative task offloading management system called FTOM, which takes advantage of powerful remote cloud-computing capabilities and utilizes neighboring edge servers. The main objective of the FTOM scheme is to select the optimal target node for task offloading based on server capacity, latency sensitivity, and the network’s condition. Our proposed scheme can make dynamic decisions where local or nearby MEC servers are preferred for offloading delay-sensitive tasks, and delay-tolerant high resource-demand tasks are offloaded to a remote cloud server. Simulation results affirm that our proposed FTOM scheme significantly improves the rate of successfully executing offloaded tasks by approximately 68.5%, and reduces task completion time by 66.6%, when compared with a local edge offloading (LEO) scheme. The improved and reduced rates are 32.4% and 61.5%, respectively, when compared with a two-tier edge orchestration-based offloading (TTEO) scheme. They are 8.9% and 47.9%, respectively, when compared with a fuzzy orchestration-based load balancing (FOLB) scheme, approximately 3.2% and 49.8%, respectively, when compared with a fuzzy workload orchestration-based task offloading (WOTO) scheme, and approximately 38.6%% and 55%, respectively, when compared with a fuzzy edge-orchestration based collaborative task offloading (FCTO) scheme.

Vehicular edge computing (VEC) is one of the prominent ideas to enhance the computation and storage capabilities of vehicular networks (VNs) through task offloading. In VEC, the resource-constrained vehicles offload their computing tasks to the local road-side units (RSUs) for rapid computation. However, due to the high mobility of vehicles and the overloaded problem, VEC experiences a great deal of challenges when determining a location for processing the offloaded task in real time. As a result, this degrades the quality of vehicular performance. Therefore, to deal with these above-mentioned challenges, an efficient dynamic task offloading approach based on a non-cooperative game (NGTO) is proposed in this study. In the NGTO approach, each vehicle can make its own strategy on whether a task is offloaded to a multi-access edge computing (MEC) server or a cloud server to maximize its benefits. Our proposed strategy can dynamically adjust the task-offloading probability to acquire the maximum utility for each vehicle. However, we used a best response offloading strategy algorithm for the task-offloading game in order to achieve a unique and stable equilibrium. Numerous simulation experiments affirm that our proposed scheme fulfills the performance guarantees and can reduce the response time and task-failure rate by almost 47.6% and 54.6%, respectively, when compared with the local RSU computing (LRC) scheme. Moreover, the reduced rates are approximately 32.6% and 39.7%, respectively, when compared with a random offloading scheme, and approximately 26.5% and 28.4%, respectively, when compared with a collaborative offloading scheme.

The moving object detection refers to the detection of physical moving objects from a video, which is applied in video surveillance, object recognition, object counting, human-computer interaction, and so on. Moreover, nowadays, real-time moving object detection is used as services in the cloud, edge, and fog computing. However, the existing methods do not meet the trade-off between accuracy and complexity. To address these issues, we present a background subtraction-based moving object detection method, called Fast-D. In this paper, we look at the `non-smoothing color feature' to make the moving object detection more robust in real-time. Each color feature is given equal significance during the classification of a pixel. Background model and threshold are initialized for each pixel. And then, the background model and threshold are updated dynamically when there are changes in the background of the video. Adaptive post-processing is considered to discard salt and pepper noise and fill holes in the detected moving object silhouettes. The evaluation of our proposed method on four complex datasets exhibits the significance.

Cloud computing has become the cornerstone of modern technology, propelling industries to unprecedented heights with its remarkable and recent advances. However, the fundamental challenge for cloud service providers is real-time workload prediction and management for optimal resource allocation. Cloud workloads are characterized by their heterogeneous, unpredictable, and fluctuating nature, making this task even more challenging. As a result of the remarkable achievements of deep learning (DL) algorithms across diverse fields, scholars have begun to embrace this approach to addressing such challenges. It has become the defacto standard for cloud workload prediction. Unfortunately, DL algorithms have been widely recognized for their vulnerability to adversarial examples, which poses a significant challenge to DL-based forecasting models. In this study, we utilize established white-box adversarial attack generation methods from the field of computer vision to construct adversarial cloud workload examples for four cutting-edge deep learning regression models, including Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), 1D Convolutional Neural Network (1D-CNN) and attention-based models. We evaluate our study with three widely recognized cloud benchmark datasets: Google trace, Alibaba trace, and Bitbrain. The findings of our analysis unequivocally indicate that DL-based cloud workload forecasting models are highly vulnerable to adversarial attacks. To the best of our knowledge, we are the first to conduct systematic research exploring the vulnerability of DL-based models for workload forecasting in the cloud data center, highlighting the inherent hazards to both security and cost-effectiveness in cloud data centers. By raising awareness of these vulnerabilities, we advocate the urgent development of robust defensive mechanisms to enhance the security of cloud workload forecasting in a constantly evolving technical landscape.

Deep learning-based action recognition in videos has obtained much attention because of achieving remarkable performance in diverse applications. However, due to the heterogeneous background and noisy spatio-temporal cues, extracting highly discriminative features is still quite challenging. To deal with this problem, numerous methods have been published based on the attention mechanism and skeleton modality. Instead of focusing on data pre-processing, we shed light on the feature map and concentrate on extracting highly discriminative features. First, we introduce B atch-wise E ntropy S upervised S tream (BESS) to extend feature discrimination similar to the uncertainty of the corresponding batch. Secondly, to obtain a more generalized model, we propose a S tream to H armonize the feature discrimination by A ugmenting both F eatures (HAFS) of ResNext101 and BESS. These two streams are hallucinated by the distillation and feature fusion technique effectively into HAFS. We introduce a new metric to assess the characteristics of the feature map. This metric depicts the relationship between the feature discrimination and recognition accuracy. Finally, we comprehensively evaluate our approach on two benchmark datasets, HMDB51 and UCF101. Experimental results demonstrate that, extending and then harmonizing the feature discrimination is one of the effective ways of generating highly discriminative features. Experimental outcomes indicate the superiority of our proposed technique over the existing state-of-the-art methods.