The last few years have seen an increased adoption of cloud infrastructure, which has in turn led to a growth in large-scale distributed architectures in data centers to accommodate cloud resource elasticity and resiliency better. Selecting the right approach to build secure, scalable, and reliable cloud infrastructure within a budget is always a challenge. This text focuses on offering practical solutions for designing and building a secure, scalable, and reliable cloud-based infrastructure where auto-scaling and multi-region deployments are the two key approaches to offer high availability. It covers designing secure and scalable microservices using cloud platforms. The content will provide an understanding of public cloud architecture, the design of microservices running on the cloud, and also the design patterns used in the cloud era. With real-world examples, you will learn how microservices can enable scalable distributed systems. Furthermore, you will be walked through multi-region deployments, auto-scaling, and traffic management in cloud environments, using a sample environment setup and useful tips and tricks for monitoring. Finally, you will see a mock implementation of cloud infrastructure on-premise for a private cloud or single-node cloud. By the end of this text, you will be able to build, manage, and deploy a highly scalable and reliable cloud-ready solution [1].

Modern organizations leverage highly distributed, global deployments to provide high availability and resiliency for cloud-first applications. By hosting these applications across multiple geographic locations and relying on highly available services, organizations can prevent disruption to their business and reduce complexity by employing the scale of infrastructure offered by major cloud providers. Global deployments in the cloud are built on well-known models such as failover, load balancing, and scalability. However, traditional methods used to recover from regional failure—while effective—can be complex. Typical multi-region recovery and high availability system architectures have latency and cost risks that should be considered when facing other limitations such as deployment models in the cloud. This document describes the different traffic management techniques that can be applied to multi-region strategies, focusing on trade-offs and costs. The introduction of new traffic management techniques being applied to the traditional global architectures now allows organizations to adopt cloud services more efficiently. Traffic management is much more straightforward in some environments, while others have started to leverage their traffic management platform via routing. In multi-region deployments, active-active and active-passive are the most common architectural models, allowing organizations to seamlessly handle failover, scalability, and global distribution based on business goals and requirements. However, traffic management for these infrastructures is critical to ensure just data distribution and efficiency, maintaining costs under control and workloads rerouted when necessary. Using the new traffic management techniques will allow organizations to evolve system architectures easily based on business requirements, taking advantage of cost benefits from multiple infrastructures. In these scenarios, traffic management becomes a crucial backbone of success to ensure that traffic is being efficiently and intelligently distributed [1].