Migration challenges of legacy software to the cloud: a socio-technical perspective

Cogent Business & Management, Год журнала: 2025, Номер 12(1)

Опубликована: Май 17, 2025

Язык: Английский

---

QEST: Quantized and Efficient Scene Text Detector Using Deep Learning

ACM Transactions on Asian and Low-Resource Language Information Processing, Год журнала: 2022, Номер 22(5), С. 1 - 18

Опубликована: Март 26, 2022

Scene text detection is complicated and one of the most challenging tasks due to different environmental restrictions, such as illuminations, lighting conditions, tiny curved texts, many more. Most works on scene have overlooked primary goal increasing model accuracy efficiency, resulting in heavy-weight models that require more processing resources. A novel lightweight has been developed this article improve efficiency detection. The proposed relies ResNet50 MobileNetV2 backbones with quantization used make lightweight. During quantization, precision changed from float32 float16 int8 for making In terms inference time Floating-Point Operations Per Second, method outperforms state-of-the-art techniques by around 30–100 times. Here, well-known datasets, i.e., ICDAR2015 ICDAR2019, utilized training testing validate performance model. Finally, findings discussion indicate efficient than existing schemes.

Язык: Английский

---

Power-Aware Fog Supported IoT Network for Healthcare Infrastructure Using Swarm Intelligence-Based Algorithms

Mobile Networks and Applications, Год журнала: 2023, Номер 28(2), С. 824 - 838

Опубликована: Март 3, 2023

Язык: Английский

---

Multi-granularity fusion resource allocation algorithm based on dual-attention deep reinforcement learning and lifelong learning architecture in heterogeneous IIoT

Information Fusion, Год журнала: 2023, Номер 99, С. 101871 - 101871

Опубликована: Июнь 7, 2023

Язык: Английский

---

GFL-ALDPA: a gradient compression federated learning framework based on adaptive local differential privacy budget allocation

Multimedia Tools and Applications, Год журнала: 2023, Номер 83(9), С. 26349 - 26368

Опубликована: Авг. 30, 2023

Язык: Английский

---

Paving the Path to a Sustainable Digital Future With Green Cloud Computing

Advances in computer and electrical engineering book series, Год журнала: 2024, Номер unknown, С. 44 - 66

Опубликована: Янв. 25, 2024

Amidst an era marked by a relentless surge in digital data and computational demands, the imperative for eco-conscious sustainable computing solutions has reached unprecedented significance. This study delves into emerging realm of green cloud (GCC), pivotal catalyst cultivating greener tomorrow. To nurture frontier, this research investigates various GCC strategies encompassing efficient center designs, resource optimization techniques, innovative virtualization practices. Additionally, authors scrutinize real-world instances industry leaders embracing energy sources. Furthermore, they shed light on obstacles within eco-friendly while illuminating forthcoming trends triumphant integration technologies. offers profound insights researchers, students, stakeholders alike.

Язык: Английский

---

Anomaly Detection in Fog Computing Architectures Using Custom Tab Transformer for Internet of Things

Electronics, Год журнала: 2022, Номер 11(23), С. 4017 - 4017

Опубликована: Дек. 3, 2022

Devices which are part of the Internet Things (IoT) have strong connections; they generate and consume data, necessitates data transfer among various devices. Smart gadgets collect sensitive information, perform critical tasks, make decisions based on indicator connect interact with one another quickly. Securing this is most vital challenges. A Network Intrusion Detection System (IDS) often used to identify eliminate malicious packets before can enter a network. This operation must be done at fog node because devices naturally low-power do not require significant computational resources. In same context, we offer novel intrusion detection model capable deployment nodes detect undesired traffic towards IoT by leveraging features from UNSW-NB15 dataset. Before continuing training models, correlation-based feature extraction weed out extra information contained within data. helps in development that has low overall load. The Tab transformer proposed well existing dataset outperforms traditional Machine Learning ML models developed as previous efforts made was designed only handling continuous As result, obtained performance 98.35% when it came classifying normal abnormal However, model’s for predicting attacks involving multiple classes achieved an accuracy 97.22%. problem imbalanced appears cause issues underrepresented classes. evaluation results were given indicated opened new avenues research detecting anomalies nodes.

Язык: Английский

---

Introducing an adaptive model forauto‐scalingcloud computing based on workload classification

Concurrency and Computation Practice and Experience, Год журнала: 2023, Номер 35(22)

Опубликована: Апрель 12, 2023

Summary With the increasing expansion of cloud computing services, one main goals researchers is to solve its major challenges. Cloud service providers must satisfy level agreement for customers and prevent resource wastage as much possible. Without a precise, optimal, dynamic policy, this unattainable. The key idea ability acquire resources you need them release when no longer them, named “Cloud Elasticity.” Elasticity trade‐off between acquisition release, if optimization done best, will be fully achieved provider have least waste resources. used machine learning techniques predict user workload decide scale up/out A challenging issue different characteristics users' workloads. results show that each prediction algorithm works well on class workloads not all. Hence, in study, new architecture has been suggested automatically classify based their sequential statistical characteristics. First, are extracted then trained neural network classifies user's workload. developed adaptive model chooses best suitable among LR, SVM, ARIMA indicate 10% improvement forecast error.

Язык: Английский

---

Live Migration of Virtual Machines Based on Dirty Page Similarity

IEEE Transactions on Cloud Computing, Год журнала: 2024, Номер 12(2), С. 563 - 579

Опубликована: Март 20, 2024

Pre-copy-based Virtual Machine (VM) live migration seamlessly migrates the running VM to target physical server by pre-copying memory pages and realizing updates through loop iterations. This method, which has high reliability robustness, can effectively achieve load balancing reduce energy consumption. It is widely used in industry manage cluster resources. However, it also involves many problems, such as dirty resulting from repeated transmission convergence failure of iterative transmission. Hence, pre-copy cannot efficiently allocate To resolve these a technology based on similarity proposed this paper. The access priority historical was determined calculating weight Hamming distance. A priority-based delay scheme for low decrease frequent pages, increase speed live-migration copy process, overall time VMs. comparative analysis experimental results six dimensions showed that method achieved better efficiency than conventional strategy.

Язык: Английский

---

Green drives: Understanding how environmental propensity, range and technological anxiety shape electric vehicle adoption intentions

Technological Forecasting and Social Change, Год журнала: 2024, Номер 210, С. 123859 - 123859

Опубликована: Ноя. 5, 2024

Язык: Английский

---