Comparative analysis of modern fault tolerance technologies

DOI: 10.31673/2412-9070.2022.042933

Abstract

On-demand cloud applications is one of the rapidly developing technologies that highly demanded on the market. With the increase of computer systems and requirements for real-time business intelligence, potential customers of such services turn to providers of «computing on demand» services. In the data centers that used for such services, computer systems face a large amount of processed data, which in turn leads to an imbalance in the load and provokes system failures. With infrastructure as a service (IaaS) usage, customers expect efficient performance of their tasks at an optimal price. In situation, when a client submits a task for work, it is deployed on a computer system in the data center of the service provider and there is no a hundred percent assurance that task would be completed and not failed because of lack of resources or software failure. Fault tolerance and high availability issues are currently the most acute in that field. The modern set of software tools for providing fault tolerance is quite wide. In general, they are all called to ensure the trouble-free operation of the computer system, but in fact they have significant differences in conceptual terms. High availability solutions provide fully automated failover to a backup system so that users and applications can continue working without disruption. HA solutions must have the ability to provide an immediate recovery point. At the same time, they must provide a recovery time capability that is significantly better than the recovery time that you experience in a non-HA solution topology. Analyzed and compared modern fault tolerance and high availability technologies. Compared modern models of High Availability system and tools for its implementation. Identified the shortcomings of modern models and tools for fault tolerance implementing. Developed recommendations for the usage of approaches that would help to eliminate reviewed shortcomings in different cases.

Keywords: fault tolerance; high availability; load balancing; cloud computing; cloud application; infrastructure as a service; software as a service; computer system; business inelegance; data center; container orchestration; service level agreement; scaling; data sharing; system deploying; cluster; node.

References
1. A dynamical and load-balanced flow scheduling approach for big data centers in clouds / F. Tang, L. T. Yang, C. Tang [et al.] // IEEE Transactions on Cloud Computing. 2016. 6(4). P. 915–928.
2. Non-cooperative power and latency aware load balancing in distributed data centers / R. Tripathi, S. Vignesh, V. Tamarapalli [et al.] // Journal of Parallel and Distributed Computing. 2017. 107. P. 76–86.
3. Hiles A. The Complete Guide to IT Service Level Agreements: Aligning IT Service to Business Needs (3rd Edition) (Service Level Management). Rothstein Publishing. Р. 31–33.
4. Blokdyk G. High availability: Second Edition - CreateSpace Independent Publishing Platform. Р. 62–66.
5. Aashish Khaira. A book on decision making for event-driven maintenance: Steps Towards High Availability & Low Downtime // LAP LAMBERT Academic Publishing. Р. 14–17.
6. Wittig A., Wittig M. Amazon Web Services in Action // Manning. Р. 442–450.
7. Gustavo Rau de Almeida Callou. Planning of Sustainable Data Centers with High Availability: An Integrated Modeling Approach to Evaluate and Optimize Sustainability, Dependability and Cost of Data Center Systems // LAP LAMBERT Academic Publishing. Р. 111–112.
8. Rastogi G., Sushil R. Analytical literature survey on existing load balancing schemes in cloud computing // IEEE. Р. 5–7.
9. Piedad F., Hawkins M. High Availability: Design, Techniques, and Processes – Pearson. Р. 42–47.
10. Hideto Saito, Hui-Chuan Chloe Lee, Ke-Jou Carol Hsu. Kubernetes Cookbook: Practical solutions to container orchestration, 2nd Edition //Packt Publishing. Р. 265–266.
11. Бондарчук А. П. Розрахунок максимальних значень інтенсивності потоків даних між окремими вузлами інфокомунікаційної мережі // Сучасний захист інформації. 2015. №. 2. С. 89–92.