Paper: Evaluating Solutions for Achieving High Availability or Near Zero Downtime

Publisher: IEEE

Authors: Antra Malhotra; Amr Elsayed (Salem); Randolph Torres; Srinivas Venkatraman

Link: https://ieeexplore.ieee.org/document/10214005

Background:

High availability is critical for cloud-based applications, as downtime can result in significant revenue losses and damage to reputation. Load balancing is an essential component of high availability, as it enables the distribution of incoming traffic across multiple servers or data centers.

In this paper, we present an evaluation of various solutions for achieving high availability or near-zero downtime in cloud computing systems. Our research focuses on the critical aspect of load balancing, which is essential for ensuring the reliability and performance of cloud-based applications.

We analyzed several popular load-balancing solutions, including HAProxy, Amazon Elastic Load Balancer (ELB), and PgBouncer. To achieve high availability, we recommend a combination of solutions that provide redundancy and failover capabilities. We also propose a framework for evaluating the performance and reliability of load-balancing solutions in cloud computing systems.

Methodology:

We conducted a comprehensive evaluation of several load-balancing solutions, including HAProxy, Amazon ELB, and PgBouncer. We analyzed their performance, reliability, and scalability in cloud computing systems.

Results:

Our evaluation shows that each solution has its strengths and weaknesses. For example, HAProxy provides excellent performance and scalability but lacks built-in redundancy. Amazon ELB offers robust redundancy capabilities but can be complex to implement. PgBouncer provides a balance of performance and redundancy but may not be suitable for large-scale applications.

Conclusion:

Achieving high availability or near zero downtime in cloud computing systems requires a combination of solutions that provide redundancy and failover capabilities. Our evaluation shows that no single solution can guarantee 100% uptime, but a well-designed framework can help ensure the reliability and performance of load-balancing solutions.

This paper contributes to the growing body of research on cloud computing and high availability by providing a comprehensive evaluation of various load-balancing solutions.

Future Work:

In future work, we plan to explore new approaches for achieving high availability in cloud computing systems. We will also investigate the application of machine learning techniques to improve the performance and reliability of load-balancing solutions