Without a clear understanding of what data resides where, the Internet of Things (IoT) could be overrun with masses of data that has no business value, says software provider Yotta. Adrian Bridgwater spoke to CTO Manish Jethwa.
Businesses risk being overwhelmed by IoT data that has little value to the enterprise, claims software provider, Yotta.
This is why Internet of Things (IoT) apps need to be engineered and architected for the data landscape in which they will run, according to Manish Jethwa, CTO of the infrastructure asset management software firm.
“The application itself needs to provide an interface that is capable of applying a structure around the data it comes into contact with, such as its hierarchy, so that it can establish which information to classify as useful for business insight.
“More, the use of data queues – with high uptime specifications – can help resolve resiliency issues by providing a temporary store for data, which the application can process when it has the capacity,” he explained.
The madness of data crowds
Jethwa believes that developers should push applications towards a better appreciation of data structure and queuing principles, especially on the journey towards smart cities or heavyweight Industrial Internet of Things (IIoT) applications. In these areas, cloud-based metadata analytics will be essential to stop organisations being overwhelmed by the sheer volume of data being gathered by sensors.
“For sensors and probes that operate with low power limitation, the cloud provides the first opportunity to analyse data, and identify important trends or patterns using elastic resources,” he said. “In this way, the analysis process to extract metadata can act as a filter, only passing on key information that requires action.”
Microservice modular boundaries
Jethwa believes that microservices will be critical in driving data to the right place on the IoT data map. But there are hurdles to clear in installing and managing them, he said. “The challenges of maintaining microservices are the same as with any distributed system, in that managing multiple services on a regular basis demands efficient operations management skills.
“Microservices provide strong modular boundaries, but maintaining consistency with other components poses a challenge. Realistically, problems like these can only be solved through the use of ‘distribution automation’, using tools to connect directly to cloud services platforms, such as AWS and Azure.”
What are the key elements to consider when developing a set of software layers that are intended to operate as an IoT network?
“With a typical IoT network, the business requirements for the data being collected, and for the network engineering, are most likely being developed by separate organisations. So it is quite possible that the data provided by IoT sensors may not be what the business requires.
“Again, microservices can serve as a translation mechanism in the data stream to make the data more useful. But it is vital that the business case for the data is assessed beforehand, to ensure that the commercial need to use the data exists.”
Internet of business says
Wise words. Simply powering up IoT applications and turning on the pipe will be of little use to organisations if they then drown in a torrent of useless data. Jethwa’s commentary provides welcome granular detail, as well as stressing the need for strategic business use cases.
What happens next will perhaps come down to the new breed of data-developers, who will engineer the ‘right complexity’ into real-world IoT systems. Programming the IoT requires a developer who has an appreciation for data architecture and topography, and an understanding of just how fast the data pipe is filling up the business.
Read more: Why HarperDB’s HTAP release unlocks the IoT data value chain
Read more: Why data streaming matters in the IoT
Read more: Huawei’s roadmap leads to IoT automation services
IoTBuild is coming to San Francisco, CA on March 27 & 28, 2018 – Sign up to learn all you need to know about building an IoT ecosystem.
