MICROSERVICE AND MONOLITH PERFORMANCE COMPARISON IN TRANSACTION APPLICATION

Abstract

When it comes to building or developing an online system, developers need to choose what kind of architecture will be used for the system. When facing the challenge, a developer needs the most suitable architecture that most fit the case whether uses microservices or monolithic architecture. Both architectures offer different benefits. Microservice recently become popular because many large companies start migrating from monolith to microservices but on the other hand. many organizations are still unfamiliar with microservices. Despite microservice providing many benefits, it also has challenges. With that being said, it is common that many organizations choose to stick with monolithic architecture since it was easier to maintain, develop, and deploy. To find out the better architecture performance-wise, API provided by both application need to be tested. The test was conducted by hitting the API several times with many threads concurrently. The test result is latency and request time, success rate needs to be monitored as well since the error occurred during load testing. With the data collected it could be shown which architecture performs better. The test results obtained by testing both monolith and microservice with several scenarios are quite unexpected. From the data, it turns out that for most features, average latency from the monolith is better than microservice. Meanwhile in many other scenarios, microservice edge monolith on the close gap in terms of success rate. The latency average number result for microservice is slightly worse because there are more success requests and taking more time.