MODERN MODELS AND METHODS OF RESOURCE MANAGEMENT OF DISTRIBUTED COMPUTER SYSTEMS

DOI: 10.31673/2412-4338.2024.019902

Abstract

The allocation of resources in heterogeneous distributed computer systems is a challenging task, constrained by factors such as task diversity and the decision process for optimal node selection. Traditional scheduling methods face limitations in addressing these complexities. This research proposes an AI-based optimization approach that leverages neural networks and deep learning techniques to efficiently allocate tasks across diverse nodes.
The core component is a neural network responsible for assigning tasks to nodes based on attributes like computational efficiency, security, fault tolerance, and data transfer latency. Node attributes representing current state are continuously monitored and used to train the neural network, allowing it to learn node capabilities. When a new task arrives, the trained network matches it to the most suitable node by comparing task requirements to learned node attributes.
Extensive experiments compared the performance of feedforward neural networks (FFNN) and convolutional neural networks (CNN) across five datasets of varying sizes (100-2000 rows representing potential nodes). The FFNN demonstrated superior overall accuracy and consistency, achieving 90-98.6% validation accuracy, while the CNN exhibited fluctuating performance.
The proposed AI-based scheduling approach provides an adaptive framework for optimally assigning heterogeneous tasks in distributed environments. Key advantages include adaptability to changing system conditions through continuous training, flexible task-node mapping based on learned capabilities, scalability leveraging deep learning, and optimized resource utilization by fitting tasks to suitable nodes.
However, the experiments revealed no clear superior neural network architecture across all dataset scales. Further research aims to develop a hybrid or adaptive architecture that can dynamically adjust structure and parameters based on input data characteristics, combining strengths of feedforward and convolutional networks for efficient resource allocation tailored to specific datasets.

Keywords: distributed computer systems, resource allocation, neural networks, deep learning, task scheduling, heterogeneous systems, adaptive optimization.

Resources
1. Yang, S.S., 2016. Centralized and decentralized methods of efficient resource allocation in cloud computing (Doctoral dissertation, Arizona State University)
2. Hu, Z., Mukhin, V., Kornaga, Y. et al. The Analytical Model for Distributed Computer System Parameters Control Based on Multi-factoring Estimations. J Netw Syst Manage 27, 351–365 (2019). https://doi.org/10.1007/s10922-018-9468-x
3. Zhengbing Hu, Vadym Mukhin, Yaroslav Kornaga, Yaroslav Lavrenko, Oksana Herasymenko,"Distributed Computer System Resources Control Mechanism Based on Network-Centric Approach", International Journal of Intelligent Systems and Applications(IJISA), Vol.9, No.7, pp.41-51, 2017. DOI:10.5815/ijisa.2017.07.05
4. Luong, N.C., Hoang, D.T., Gong, S., Niyato, D., Wang, P., Liang, Y.C. and Kim, D.I., 2019. Applications of deep reinforcement learning in communications and networking: A survey. IEEE Communications Surveys & Tutorials, 21(4), pp.3133-3174.
5. Wang, S., Tuor, T., Salonidis, T., Leung, K.K., Makaya, C., He, T. and Chan, K., 2019. Adaptive federated learning in resource constrained edge computing systems. IEEE Journal on Selected Areas in Communications, 37(6), pp.1205-1221.
6. Huang, K., Bi, S. and Zhang, Y.J., 2020. Deep reinforcement learning for online computation offloading in wireless powered mobile-edge computing networks. IEEE Transactions on Mobile Computing, 19(11), pp.2581-2593.
7. Ouyang, W., Xu, R., Wang, X., Zhang, L. and Chang, P., 2019. Adaptive user-managed service placement for mobile edge computing: An online learning approach. IEEE Transactions on Mobile Computing, 20(3), pp.753-766.
8. Liu, J., Mao, Y., Zhang, J. and Letaief, K.B., 2016. Delay-optimal computation task scheduling for mobile-edge computing systems. In IEEE International Symposium on Information Theory (ISIT) (pp. 2760-2764). IEEE.
9. Behmandpoor, P., Patrinos, P. and Moonen, M., 2022. Federated Learning Based Resource Allocation for Wireless Communication Networks. In 2022 30th European Signal Processing Conference (EUSIPCO). IEEE.
10. Zhou, Z., Chen, X., Li, E., Zeng, L., Luo, K. and Zhang, J., 2019. Edge intelligence: Paving the last mile of artificial intelligence with edge computing. Proceedings of the IEEE, 107(8), pp.1738-1762.
11. Wang, X., Han, X., Wang, C.X., Zhang, Q. and Chen, X., 2019. In-edge AI: Intelligentizing mobile edge computing, caching and communication by federated learning. IEEE Network, 33(5), pp.156-165.
12. Dinh, C.T., Tang, J., La, Q.D. and Quek, T.Q., 2017. Offloading in mobile edge computing: Task allocation and computational frequency scaling. IEEE Transactions on Communications, 67(8), pp.5389-5404.
13. Wang, X., Han, Y., Leung, V.C. and Niyato, D., 2020. Convergence of edge computing and deep learning: A comprehensive survey. IEEE Communications Surveys & Tutorials, 22(2), pp.869-904.
14. Mao, Y., You, C., Zhang, J., Huang, K. and Letaief, K.B., 2017. A survey on mobile edge computing: The communication perspective. IEEE Communications Surveys & Tutorials, 19(4), pp.2322-2358.
15. Yu, S., Wang, X. and Langar, R., 2017. Computation offloading for mobile edge computing: A deep learning approach. 2017 IEEE 28th Annual International Symposium on Personal, Indoor, and Mobile Radio Communications (PIMRC) (pp. 1-6). IEEE.
16. Zhang, X., Zhu, H., Han, Y. and Li, Z., 2020. Joint Resource Allocation and Task Offloading in Edge Computing via Deep Reinforcement Learning. In Proc. IEEE Global Communications Conference (GLOBECOM) (pp. 1-6). IEEE.
17. Sellami, B., Hakiri, A., Ben Yahia, S. and Berthou, P., 2020. Deep Reinforcement Learning for Energy-Efficient Task Scheduling in SDN-based IoT Network. In 2020 IEEE 19th International Symposium on Network Computing and Applications (NCA) (pp. 1-4). IEEE.
18. Cui, D. et al., 2020. Cloud Workflow Task and Virtualized Resource Collaborative Adaptive Scheduling Algorithm Based on Distributed Deep Learning. In 2020 IEEE International Conference on Advances in Electrical Engineering and Computer Applications (AEECA) (pp. 137-140). IEEE.