Search

Network operations involve several decision-making tasks. Some of these tasks are related to operators, such as extending the footprint or upgrading the network capacity. Other decision tasks are related to network functions, such as traffic classifications, scheduling, capacity, coverage trade-offs, and policy enforcement. These decisions are often decentralized, and each network node makes its own decisions based on the preconfigured rules or policies. To ensure effectiveness, it is essential that planning and functional decisions are in harmony. However, human intervention-based decisions are subject to high costs, delays, and mistakes. On the other hand, machine learning has been used in different fields of life to automate decision processes intelligently. Similarly, future intelligent networks are also expected to see an intense use of machine learning and artificial intelligence techniques for functional and operational automation. This article investigates the current state-of-the-art methods for packet scheduling and related decision processes. Furthermore, it proposes a machine learningbased approach for packet scheduling for agile and cost-effective networks to address various issues and challenges. The analysis of the experimental results shows that the proposed deep learning-based approach can successfully address the challenges without compromising the network performance. For example, it has been seen that with mean absolute error from 6.38 to 8.41 using the proposed deep learning model, the packet scheduling can maintain 99.95% throughput, 99.97% delay, and 99.94% jitter, which are much better as compared to the statically configured traffic profiles. [ABSTRACT FROM AUTHOR]

The context of this study examines the requirements of Future Intelligent Networks (FIN), solutions, and current research directions through a survey technique. The background of this study is hinged on the applications of Machine Learning (ML) in the networking field. Through careful analysis of literature and real-world reports, we noted that ML has significantly expedited decision-making processes, enhanced intelligent automation, and helped resolve complex problems economically in different fields of life. Various researchers have also envisioned future networks incorporating intelligent functions and operations with ML. Several efforts have been made to automate individual functions and operations in the networking domain; however, most of the existing ML models proposed in the literature lack several vital requirements. Hence, this study aims to present a comprehensive summary of the requirements of FIN and propose a taxonomy of different network functionalities that needs to be equipped with ML techniques. The core objectives of this study are to provide a taxonomy of requirements envisioned for end-to-end FIN, relevant ML techniques, and their analysis to find research gaps, open issues, and future research directions. The real benefit of ML applications in any domain can only be ensured if intelligent capabilities cover all of its components. We observed that future generations of networks are heterogeneous, multi-vendor, and multidimensional, and ML can provide optimal results only if intelligent capabilities are used on a holistic scale. Realizing intelligence on a holistic scale is only possible if the ML algorithms can solve heterogeneous problems in a multi-vendor and multidimensional environment. ML models must be reliable and efficient, support, and possess the capability to learn and share the knowledge across the network layers and administrative domains to solve issues. First, this study ascertains the requirements of the FIN and proposes their taxonomy through reviews on envisioned ideas by various researchers and articles gathered from reputed conferences and standard developing organizations using keyword queries. Second, we have reviewed existing studies on ML applications focusing on coverage, heterogeneity, distributed architecture, and cross-domain knowledge learning and sharing. Our study observed that in the past, ML applications were focused mainly on an individual/isolated level only, and aspects of global and deep holistic learning with cross-layer/cross-domain knowledge sharing with agile ML operations are not explored at large. We recommend that the issues mentioned previously be addressed with improved ML architecture and agile operations and propose an ML pipeline based architecture for FIN. The significant contribution of this study is the impetus for researchers to seek ML models suitable for a modular, distributed, multi-domain, and multi-layer environment and provide decision making on a global or holistic rather than an individual function level. [ABSTRACT FROM AUTHOR]

With the rapid increase of compute intensive tasks and workload, the need of the hour is to manage the total workload due to resource constraints and increased cost. For complex computations where multiple computer systems are required to execute a single task such as in federated cloudlet environment, load balancing is the main challenge. Load balancing means to divide the total workload between all the member nodes to get the maximum benefits from the federated environment. A cloudlet is a resourceful computer that is coupled to the Internet and is accessible to the mobile devices in the vicinity. Cloudlet Federation is the concept of a cooperative framework to share resources and load balancing among various member cloudlets. Different tasks consume different amount of energy for their execution that results in a large amount of heat dissipation. Due to heat, the performance of the systems is decreased. The more heat, the more performance degradation, so it becomes very vital to keep the energy consumption as minimum as possible. To address this problem, this paper proposes a novel scheduling strategy that will distribute the total load between all the member nodes in a federated cloudlet environment for load balancing. The proposed methodology not only considers the energy level of the system on which the task has to be executed but also considers the energy required by the tasks in queue. The proposed methodology is tested in a Cloudlet Federation environment and the results show an improved load balancing in terms of energy consumption. [ABSTRACT FROM AUTHOR]

Purpose: This retrospective study presents a comparative analysis of the overall survival and toxicities, as side effects, of docetaxel plus cyclophosphamide (TC) and doxorubicin plus cyclophosphamide (AC). The study measured their efficacies during adjuvant chemotherapy, treating Pakistani breast cancer patients by validating the results obtained, with the published analysis of the same treatment given to US patients. Patients and methods: Between June 2015 and September 2017, for four chemotherapy cycles, 189 patients out of 358 received TC (75 mg/m2 of docetaxel, 600 mg/m2 of cyclophosphamide) and 169 were treated with AC (60 mg/m2 of doxorubicin, 600 mg/m2 of cyclophosphamide). On the basis of using pathological markers to assess patients, toxicities, as side effects, (due to docetaxel, doxorubicin, and cyclophosphamide) were listed in the database of this study. Common factors with respect to common terminology criteria for adverse events version 5.0 and side effects listed in MedlinePlus, NIH US database, and from the database of this study were then separated to be included in comparison for this study. Statistically, chi-squared test was used at α=0.05. Results: There was no statistically significant difference between the proportions of patients with vomiting, extreme tiredness, diarrhea, mild anemia, stability, and overall survival because P-value >0.05. However, AC remained less toxic (P-value <0.05) by 22.6%, 25.7%, 25.3%, 12.4%, 20.8%, and 16.4% compared to TC for changes in taste, muscle pain, burning hands, change in hemoglobin level, moderate anemia, and needing blood transfusion respectively, whereas TC remained less toxic by 52.9%, 32.5%, and 26.3% for dizziness, weight loss, and sores in throat and mouth, respectively. Conclusion: At 27 months, TC was more toxic than AC, whereas both combinations had the same overall survival rate. [ABSTRACT FROM AUTHOR]

The availability of cheap depth range sensors has increased the use of an enormous amount of 3D information in hand-held and head-mounted devices. This has directed a large research community to optimize point cloud storage requirements by preserving the original structure of data with an acceptable attenuation rate. Point cloud compression algorithms were developed to occupy less storage space by focusing on features such as color, texture, and geometric information. In this work, we propose a novel lossy point cloud compression and decompression algorithm that optimizes storage space requirements by preserving geometric information of the scene. Segmentation is performed by using a region growing segmentation algorithm. The points under the boundary of the surfaces are discarded that can be recovered through the polynomial equations of degree one in the decompression phase. We have compared the proposed technique with existing techniques using publicly available datasets for indoor architectural scenes. The results show that the proposed novel technique outperformed all the techniques for compression rate and RMSE within an acceptable time scale. [ABSTRACT FROM AUTHOR]