What Soracom’s AI integration means for the future of cellular IoT

By: Eleanor Hecks, Business and Technology Writer July 19, 2024

Collected at: https://www.rcrwireless.com/20240719/ai-ml/what-soracoms-ai-integration-means-for-the-future-of-cellular-iot-reader-forum

Soracom has integrated artificial intelligence capabilities into its Internet of Things (IoT) offerings, marking a new era for device connectivity and telecommunications. What are the implications of this decision? Will it permanently affect cellular IoT services?

Soracom brings AI capabilities to IoT ecosystems

Soracom — the Japan-based global connectivity provider — is bringing generative AI capabilities to internet-connected ecosystems. IoT-focused telecommunication firms can use any of its three tools to process data in real time or connect their devices to the leading hyperscaler platforms’ machine learning tools.

Each of the three services delivers unique features, meaning telecom companies can use whichever aligns with their needs. They can choose from Soracom Query, Soracom Harvest Data Intelligence or Soracom Relay to process, analyze, or transfer IoT data.

Many businesses have incorporated machine learning or generative models into internet-connected devices. Some have even developed standalone, out-of-the-box software solutions for telecommunications. However, Soracom’s AI integration into cellular IoT services is novel because it focuses on backend operations instead of automating administrative or customer-facing functions.

What can Soracom’s three new AI services do?

Soracom Query allows you to use structured query language (SQL) queries from customer lifecycle intelligence (CLI) or business intelligence (BI) tools to mine IoT-generated data without deploying servers. In other words, you can pull information from a database without utilizing separate storage. Soracom refers to it as a “managed data warehouse capability.”

Since Soracom Query has automatic data loading — the incremental processing of incoming information without additional setup — running complex analytical SQL queries is simple, even if your data sets are sizable. If you are interested, this service is available in technical preview.

Soracom Harvest Data Intelligence is a generative model — the GPT-3.5-Turbo algorithm provided by Microsoft Azure OpenAI Service, to be precise — that analyzes time series information. In other words, it processes and interprets a sequence of data points recorded in intervals to identify notable trends, patterns, outliers and anomalies.

You can prompt the AI to generate insights, such as what to do with the data or if any deviations are present in the time series. You can also give it commands, like telling it to describe trends. This service is available in public beta.

With Soracom Relay, cameras running real-time streaming protocol and real-time transport protocol can transmit video and audio files on demand to Soracom’s Harvest Files or cloud storage systems like Amazon Web Service’s Kinesis Video Streams. This service is available in technical preview.

The implications for the future of cellular IoT services

These three AI-powered products will likely have lasting impacts on cellular IoT services, as their benefits are substantial. For instance, since Soracom Query enables telecom firms to scale BI and CLI SQL query processing dynamically, their infrastructure expenses decrease.

Additionally, since it enables efficient processing of sizable data sets, decision-makers can extract critical insights sooner. Consequently, they can rapidly adapt their data mining and utilization strategies to reflect evolving device needs, use cases, or market changes.

Enhanced security is another integration benefit, as Soracom Harvest Data Intelligence can flag potential cyber threats by bringing attention to anomalies. Considering a single data breach costs $4.24 million on average, this service could generate significant cost savings, potentially prompting technological innovations or accelerating research and development.

While some decision-makers may not prioritize cybersecurity, bad actors pose a significant threat. In one real-world case, a telecom company lost about 33% of its value and roughly 100,000 customers after experiencing a data breach. Since cellular IoT services acquire and transmit millions of data points daily, security is critical.

One benefit of Soracom Relay is its cost-effective scalability. It can dynamically adjust resolution, bitrate, and video encoding depending on network conditions, demand, and server volume to optimize bandwidth and storage resources, significantly increasing data acquisition and transfer efficiency.

Also, it may help firms identify anomalies in real time. For example, it may draw attention to a lost data packet, enabling retransmission. At the very least, its activity provides an observable log so there is no worry about data loss. Its processing may even offer insights, helping readjust streaming parameters like bit rate, resolution and keyframe interval.

Scalability is among these tools’ main implications, as they rely on autonomous models. Considering 30.9 billion internet-connected devices will exist by 2025, cellular IoT services’ ability to remain dynamic is crucial. After all, data set size positively correlates to device count, meaning accounting for future transmission demand and storage flexibility is critical.

Challenges that may arise From Soracom’s AI integration

The potential challenges that may arise from Soracom’s AI integration revolve around the typical pitfalls of machine learning technology, the main being inaccuracy. Unlike robot process automation or scripts — other similar automation tools — algorithms can act autonomously, make biased decisions or hallucinate information.

Telecom firms hoping to use AI to extract valuable insight or automate transmission processes may face errors because AI is often mistake-prone. Bad actors can even intentionally manipulate them. Adversarial examples — inputs meant to fool an algorithm — caused one machine learning model to misclassify images just over 69% of the time, according to one study.

Optimizing acquisition, processing, analysis and transmission efficiency will be crucial as cellular IoT services expand. While most implications of Soracom Query, Soracom Harvest Data Intelligence or Soracom Relay are positive, overreliance could spell trouble. Enterprise leaders would be wise to test the limitations and capabilities of these technical previews thoroughly before committing to one.

The future of AI technology in cellular IoT services

As the IoT and AI fields continue expanding, out-of-the-box services like Soracom’s will become more prevalent. Currently, their implications are mostly positive — they stand to optimize many backend processes, reduce infrastructure strain and improve insight-driven action. However, you should approach these tools cautiously, considering they’re still in technical preview.