Various industrial applications deal with high-frequency data. Traditionally, these systems have analyzed high-frequency data directly on the data source or at the commanding PLC. However, currently, Industry 4.0 technologies support new monitoring scenarios for high-frequency data monitoring where raw data is transmitted in soft-real time to an Edge/Fog or Cloud node for processing, enabling centralized computing. This demands efficient communication protocols capable of handling high-frequency, high-throughput data. This paper focuses on analyzing the performance of key IIoT (Industrial Internet of Things) messaging protocols—AMQP, MQTT, KAFKA, ZeroMQ, and OPCUA—to evaluate their suitability, in terms of latency and jitter, for transmitting high-frequency data within these new scenarios. The analysis reveals MQTT, AMQP, and ZeroMQ as top performers in Edge/Fog computing, while ZeroMQ exhibits the lowest latency and jitter in Cloud computing. Finally, a guideline for protocol selection is proposed, aiding industrial enterprises in protocol selection for specific AI use cases.