EVALUATION OF EFFICIENCY AND PERFORMANCE OF SERIALIZATION FORMATS FOR DISTRIBUTED SYSTEMS

The conducted study allows us to evaluate the impact of various serialization formats on the performance of inter-service communication, focusing on serialization speed, data bandwidth efficiency, and latency in environments integrating middleware, characteristic of microservice architectures. Through an empirical analysis of a wide range of serialization formats and comparisons with traditional standards, it is demonstrated that the compactness of serialized data formats is more critical for reducing end-to-end latency than serialization speed itself. Despite high serialization speed, protocols such as FlatBuffers and Cap'n Proto show lower performance in distributed environments due to larger message sizes, in contrast to the more balanced performance observed in protocols like Avro, Thrift, and Protobuf. The purpose of the article is to review existing data formats and message processing and transmission protocols, and through practical experiments, demonstrate the importance of optimizing message sizes to enhance network efficiency and bandwidth capacity. Keywords: data encoding, performance evaluation, message transmission protocols, distributed system, data formats.