Does my aaS look big in this? Microsoft Azure targets IoT amateurs

Microsoft has launched its IoT-as-a-Service offering, seemingly cashing in the gap between IoT euphoria and a lack of in-house expertise to make any sense of it.

Time Series Insights claims to give customers a near real-time global view of data across various event sources and quickly validate IoT solutions. It essentially an IoT interface for those organizations who do not have the in-house capabilities to build it themselves. If the promise of IoT is to spot hidden trends, spot anomalies and conduct root-cause analysis, through insight derived from mountains of data, it could be a useful play to capitalize on the current skills gap, and the desire for executives to appear like they know what they are talking about.

“Today more than ever, with increasing connected devices and massive advances in the collection of data, businesses are struggling to quickly derive insights from the sheer volume of data generated from geographically dispersed devices and solutions,” said Joseph Sirosh Corporate VP at Microsoft’s Data Group, on the company blog.

“In addition to the massive scale, there is also a growing need for deriving insights from the millions of events being generated in near real time. Any delay in insights can cause significant downtime and business impact.”

The service is now available in preview, and also includes a number of useful APIs to allow developers to integrate it into existing workflows. Alongside collating all data from existing sensors, customers can also upload historical data from various sources to build the bigger picture.

Another interesting update for the IoT portfolio is the introduction of Azure Stream Analytics on edge devices, which is also in preview. It’s a simple idea, but one which has been difficult to achieve to-date, however the ability to run complex event processing solutions on the edge will help improve the efficiency of operations.

Multiple IoT scenarios require real-time response, resiliency to intermittent connectivity, handling of large volumes of raw data, or pre-processing of data at source to ensure regulatory compliance. The ability to deploy and operate analytical intelligence physically closer to the devices is another very useful update for customers.

“In industrial IoT scenarios, the volume of data can be too large to be sent to the cloud directly due to limited bandwidth or bandwidth cost,” said Santosh Balasubramanian Principal Program Manager for Azure Stream Analytics.

“For example, the data produced by jet engines (a typical number is that 1TB of data is collected during a flight) or manufacturing sensors (each sensor can produce 1MB/s to 10MB/s) may need to be filtered down, aggregated or processed directly on the device before sending it to the cloud. Examples of these processes include sending only events when values change instead of sending every event, averaging data on a time window, or using a user-defined function.”

In a cloud world which is becoming increasingly more competitive, it would appear Microsoft is finding its niche to push forward its burgeoning IaaS ambitions. AWS has scale and accessibility as its USP, IBM is using its Watson cognitive AI solution to advance its claim and we’re not too sure where Google Cloud is targeting, but the Azure team seem to be using IoT as its differentiator.