METHOD OF THE HIGH-EFFICIENCY IMPLEMENTATION OF A SIDECAR COMPONENT WITH MINIMAL SYSTEM RESOURCE CONSUMPTION

Abstract

An analysis of existing solutions for implementing a sidekick is carried out, and a comparison of different technologies is made in terms of RAM usage and startup time, which are critical aspects for sidekick components. Based on the results of the analysis, Quarkus was chosen because its GraalVM-based advance-of-time (AOT) compilation capabilities provide high startup speed and low consumption of RAM and central processing unit (CPU). To solve problems with memory leaks, it is proposed to optimize data flow management, in particular, by eliminating blocking calls and using asynchronous approaches. Replacing the high-level WebClient with the low-level HttpClient for redirecting HTTP requests significantly reduced the use of RAM and avoided OutOfMemory errors. Choosing the optimal Garbage Collector in the JVM environment provides high load resistance. Elimination of memory leaks allows the garbage collector to steadily clear RAM without creating long pauses even under heavy load. The developed and proposed method, as a set of optimizations, demonstrated that modern approaches to the implementation of sidekick can significantly reduce the need for infrastructure resources. The selected technologies, tools, and techniques ensure the creation of a highly efficient, load-resistant sidekick that demonstrates minimal resource consumption and has a positive impact on system performance. The new method allows optimizing the management of RAM and data flows, eliminates blocking calls, uses lower-level tools for processing HTTP requests, and implements a streaming data transfer mechanism. The conducted research of the method demonstrates that, if implemented correctly, the sidekick allows minimizing additional infrastructure costs, preserving the advantages of the sidekick architecture - independence of the main microservices, ease of scaling, and high efficiency. The developed sidekick is resistant to loads and ensures minimal resource utilization. It can be effectively integrated into modern microservice systems.