Singapore probes distributed analytics for IoT limitations

Local researchers are looking at technologies that can enable data to be analysed more efficiently within the limitations of Internet of Things, as Singapore aims to take the lead in the emerging market.

Singapore researchers are looking at various technologies that can enable data to be analysed more efficiently within the limitations of Internet of Things (IoT), as the country looks to take the lead in this emerging market segment.

IoT had been touted to play a crucial role in big data analytics, gathering information that could be examined to improve products and services as well as enhance a society's quality of life. There were, however, key issues that had to be addressed for these to materialise.

Shonali Krishnaswamy, head of data analytics department at A*Star's Institute for Infocomm Research (I2R), explained that a main challenge in tapping IoT for analytics was the limited size and computational power of these devices.

IoT devices were typically developed to "sense" and send information to be analysed, but with billions of such devices expected to roam the global web, the large volume of data generated meant it would be difficult to do this with scale.

"So we need to think about how to do analytics within this environment," said Krishnaswamy, who was speaking to ZDNet on the sidelines of the IoT Asia 2016 conference held in Singapore this week. She noted that analytics at the edge and distributed analytics were some key research areas that I2R was currently working on.

She explained that not all data would eventually prove useful for analytics, but most IoT devices today would send information back to the network indiscriminately. All of the data would then have to be analysed so relevant insights could be identified and extracted.

"Sending data that's not very useful is a waste of bandwidth," she said, noting that it would be pointless, for instance, for a IoT system to send alerts every few minutes simply to indicate it was functioning normally. "It's only when something goes wrong that you need to extract more data to find out what's wrong. And if millions of devices are sending data all the time, the communications network will be clogged."

The ability to distribute data intelligence or analytics, putting some of such activities on the device itself, would enable this process to be more efficient, she said.

Charles Reed Anderson, IDC's Asia-Pacific vice president and head of mobility and IoT, also noted the development of IoT at the edge. Speaking at the conference, the analyst said 35 percent of IoT-created data would be stored and analysed at the network edge by 2019.

IoT devices designed to reside at the network edge would enable organisations to aggregate and analyse data as well as control devices.

IoT devices not secured by design

Internet of Things devices are not originally built to be robust against attacks, resulting in vulnerabilities that can lead to severe repercussions if left unchecked, such as data that's being collected and acted upon in Singapore's smart nation plan.

Read More

Echoing Krishnaswamy's comments, Anderson said this would make IoT data analytics more efficient and agile, less costly, and faster.

IDC believes there will be 8.6 billion connected devices in the region, excluding Japan, by 2020, with a US$583 billion market opportunity.

I2R researchers currently also were developing algorithm that would help determine, for instance, how much computational power a particular IoT device could produce and limit the amount of analytics it would then carry out.

And, eventually, as people became more ubiquitous and mobile phones gained more intelligence, the hardware chip itself would be built with more powerful analytics capabilities, Krishnaswamy said, pointing to the development of neuromorphic computing.

During his keynote address at the conference, Singapore's Minister of Foreign Affairs and Minister-in-Charge of Smart Nation Vivian Balakrishnan also highlighted how the ubiquity of connected devices had transformed humans into "mobile units of the internet", where almost every person in the country now carried a device that was communicative or sensory.

Pervasive computing, broadband availability, artificial intelligence, robotics, and 3D printing, among others, had transformed the way people lived, worked, played, entertained, communicated, and mobilised. "This is what is driving the profound economic changes that are sweeping the world," Balakrishnan said.

With significant technological advancements made in the last year alone such as Google's Alpha Go, he underscored the need for "cultural" readiness to help societies thrive amid the developments.

"Do you have a society with a culture of openness and of learning, where you're prepared to invest in infrastructure, education...and where you're prepared to re-format your rules and standards to take advantage of these technological advancements," the Singapore minister said. Cities and nations with such culture, encompassing the right blend of openness, cooperativeness, innovation, and investment, as well as the ability to execute well that would pull ahead of its peers.

He added this was Singapore's vision and the foundation of its smart nation initiative, which heavily featured the use of IoT. Balakrishnan also pointed to the city-state's robust infrastructure in mobile and fixed broadband, population with strengths in science and math, as well as a society and government that had embraced an engineering ethos, including a prime minister who codes.

All of these, he said, placed the country in good stead to continue to innovate and tap digital technology to improve the population's quality of life.