AtomBeam has performed extensive testing of IoT data for various data types. In each case, the AI-based AtomBeam software beats complex compression algorithms (such as zlib).

IoT sensor data is 1) highly structured, 2) uses limited bandwidth network with low power usage – LPWANS (e.g. Sigfox, LoRa, NB-IoT or LTE-M), and 3) short bursty data packets/messages. For this type of data, there have been essentially no alternatives. Standard compression algorithms simply can’t handle the short packets/message because compression requires generally over 8,000 bits of information before it can property apply its algorithms.
AtomBeam is a new alternative, that is radically efficient for IoT data. Using its AtomIzer AI engine, IoT data is learned and then streamed irrespective of the data packet size. Compaction ratios average over 70% compared with compression in the 15-20% range.