11. Technical Strategies and Architectural Patterns for Migrating Legacy Systems to the Cloud

With the advent of Internet, information has crossed the realms of books and gone digital, requiring data to be easily accessible and delivered anywhere speedily. There are myriads of formats in which data is currently available, such as videos, images, documents and Web pages. Accordingly, handling datasets in various formats has made the task of designing scalable and reliable application really challenging. Building the applications of tomorrow would need architects and developers to construct applications that can meet the needs that would demand handling high volumes of data and deliver substantial throughput. In today’s enterprises, there are legacy applications which may have been developed several years or even decades ago. At the time, the business may have been in its infancy, and so applications may have been designed to satisfactorily handle workloads prevalent in those times with some average growth factors built in. However, owing to the new emerging trends in the technology space, such as mobile and big data, the workloads at which businesses operate today have grown manifold. Also, the monolithic legacy systems serving those workloads have failed to keep pace, often struggling to deliver the SLAs (Service-Level Agreement). Although cloud is not a panacea for all kind of new demands, we believe that with some appropriate architectural restructuring, existing applications may go a long way in serving the demands of new growing businesses. Applications redesigned on the lines of parallel computing patterns such as master/worker or MapReduce and implemented on cloud platforms can be leveraged to add new life or re-energise legacy applications that can scale much better. In this chapter, we discuss an approach to transform legacy applications, designed to handle high-volume requests, by using re-engineering techniques and modern design patterns so as to effectively realise the benefits of cloud environment.