Migrating from Monoliths to Cloud-Based Microservices: A Banking Industry Example

Abstract

As more organizations are placing cloud computing at the heart of their digital transformation strategy, it is important that they adopt appropriate architectures and development methodologies to leverage the full benefits of the cloud. A mere "lift and move" approach, where traditional monolith applications are moved to the cloud will not support the demands of digital services. While, monolithic applications may be easier to develop and control, they are inflexible to change and lack the scalability needed for cloud environments. Microservices architecture, which adopts some of the concepts and principles from serviceoriented architecture, provides a number of benefits when developing an enterprise application as compared to a monolithic architecture. Microservices architecture offers agility and faster development and deployment cycles, scalability of selected functionality, and the ability to develop solutions using a mixture of technologies. Microservices architecture aims to decompose a monolithic application into a set of independent services which communicate with each other through open APIs or highly scalable messaging. In short, microservices architecture is more suited for building agile and scalable cloud-based solutions. This chapter provides a practicebased view and comparison between the monolithic and microservices styles of application architecture in the context of cloud computing, and proposes a methodology for transitioning from monoliths to cloud-based microservices.

A service graph based extraction of microservices from monolith services of service‐oriented architecture

“…SG provides the details of basic metrics which are used to determine other metric values. NoS value is given by the count of nodes in the SG, NoS = n. Coupling of services (CS) value is given by the degree of each node as given in Equation (2).…”

“…Distributed system has emerged from the monolithic style of client‐server architecture to service‐oriented architecture (SOA), and then to the trending microservices. Monolithic applications are built as a large block of code and deployed as a single archive file 2 . Since the applications are built as a single piece of code, the entire application must be deployed for every update.…”

Migrating Monoliths to Microservices

“…Monolithic applications are self-contained, consist of a single code base, include every single functionality, and are easily implemented [34]. Modularity, however, is not considered as a design principle [35]; therefore, it functions within a monolith collectively sharing resources on the host system, which limits scalability [32]. Developing, maintaining and changing monolith applications also becomes increasingly difficult and slow, as they tend to grow in size and complexity [44].…”

“…Despite the compelling advantages of microservices, migrations are rare [35]. This may be rooted in a perceived risk and uncertainty surrounding the migration.…”

Self-Adaptive Microservice-based Systems -- Landscape and Research Opportunities

“…Despite its benefits, as the number of deployed microservices increases so does its complexity, bringing some challenges to this architectural style [8]. Therefore, some approaches are necessary to monitor the microservices' status (and the containers where they execute) at runtime in order to provide transparency, elasticity, resilience and deployment management.…”