Abstract Service‐oriented architecture (SOA) has been widely used to design enterprise applications in the past two decades. The services in SOA are becoming complex with the increase in changing user requirements and SOA is still seen as monolithic from a deployment perspective. Monolithic services make the application complex, and it becomes difficult to maintain. With the evolution of microservices architecture, software architects started migrating legacy applications to microservices. However, existing migration approaches in the literature mostly focus on migrating monolithic applications to microservices. To the best of our knowledge, very few works have been done in migrating SOA applications to microservices. One of the major challenges in the migration process is the extraction of microservices from the existing legacy applications. To address this, we propose an approach to extract the candidate microservices using graph based algorithms. In particular, four algorithms are defined: (i) construction ofservice graph(SG), (ii) construction oftask graph(TG) for each service of the a SOA application, (iii) extraction of candidate microservices using the SG of SOA application, and (iv) construction of a SG for a microservices application to retain the dependencies between the generated microservices. We chose a SOA‐based web application to demonstrate the proposed microservices extraction approach and extracted the microservices. Additionally, we have evaluated the extracted microservices and compared them with SOA based services.

Service oriented architecture (SOA) has been widely used in the design of enterprise applications over the last two decades. Though SOA has become popular in the integration of multiple applications using the enterprise service bus, there are few challenges related to delivery, deployment, governance, and interoperability of services. To overcome the design and maintenance challenges in SOA, a new architecture of microservices has emerged with loose coupling, independent deployment, and scalability as its key features. With the advent of microservices, software architects have started to migrate legacy systems to microservice architecture. However, many challenges arise during the migration of SOA to microservices, including the decomposition of SOA to microservice, the testing of microservices designed using different programming languages, and the monitoring the microservices. In this paper, we aim to provide patterns for the most recurring problems highlighted in the literature i.e, the decomposition of SOA services, the size of each microservice, and the detection of anomalies in microservices. The suggested patterns are combined with our experience in the migration of SOA-based applications to the microservices architecture, and we have also used these patterns in the migration of other SOA applications. We evaluated these patterns with the help of a standard web-based application.

Background Merkel cell carcinoma (MCC) is an aggressive form of skin cancer mostly caused by the Merkel cell polyomavirus [1]. MCC although rare in incidence is associated with poor prognosis. Gene expression studies (GES) have identified a number of genes that are associated with risk of developing MCC. However, their clinical utility to predict risk, response to treatment, or treatment toxicity, remains undefined. There is a need to better understand the biology of MCC and to develop potential new targets with regard to cancer risk and prognostic value [2]. Gene expression profiling and pathway association analysis on GES data can elucidate relevant biological processes, and identify new candidate target genes. Materials and methods We sought to identify differential gene expression and functional associations in MCC. Differentially expressed genes with a P value ≤ 0.05 and a fold change ≥ 2.5 between prognostic groups were identified and pathway association analysis was done for a total of 38 differentially expressed genes from a gene expression profiling study [3] with prognostic groups of MCC patients. The over representation of gene-based associations in each pathway was calculated using Fisher’s exact test. Results Pathways involved in neuropathic pain signaling, glutamate receptor signaling and synaptic long term potentiation were found to be highly enriched with associations. These results suggest that gene expressions associated with these pathways may contribute to MCC development and prognosis.