February 29, 2016 | Written by: Kevin Allen
Share this post:
Behind the dozens of forecasts about how many billions of Internet of Things devices will come online by 2020, there lies the challenge of interconnectivity.
Gartner predicts 21 billion devices. IDC predicts 29.5 billion. Suffice to say we’re in for a data deluge. In order to get from where we are today to achieving that 15-30 percent compound annual growth rate (depending on which study you believe), we’ll need to solve this problem and create the framework for these connected devices.
“The challenge is that all these edge sensors are all so heterogeneous, so you’ve got a lot of people doing things in different ways,” Rob Hirschfeld, founder and CEO of RackN, says. “It makes it very hard to create consistent management patterns.”
Andrew Hately, Distinguished Engineer and CTO, Cloud Labs at IBM, says that Internet of Things systems are inherently open, and they’ll need to stay that way.
“[IoT] is only going to work on open tech and open ecosystems because there’s almost no single-party complete IoT solution,” he says.
In real-world IoT implementations, there are a variety of systems that must be able to communicate. Communication must be seamless from the device makers, analytics engines, all the way down to the end user. Plus, third party stakeholders may need to access the data as well.
“Almost everything in IoT requires agreement,” Hately says, “and to me, agreement in the tech world means an open technology basis.”
Open technology foundations like Cloud Foundry Foundation, OpenStack Foundation and the Linux Foundation, are viewing the growing number of IoT devices as high-volume data sources. They’re building their platforms with an eye toward connecting to and supporting a growing list of devices.
Hirschfeld says it also doesn’t make sense for the vendors to move toward proprietary systems in IoT. Vendors can differentiate themselves in the data collected, the sensors they use and the decision-making processes they implement.
“Vendors in IoT are going to drive toward open source technologies because they can’t afford to include proprietary pieces where there’s not a value added component,” Hirschfeld says.
Hirschfeld also sees an opportunity for growth and for communities to work together to improve operational management for IoT.
“When I look at those pieces and parts, a lot of the open source stuff doesn’t manage and interconnect as a system very well,” he says. “The challenge is that all these edge sensors are all so heterogeneous, so you’ve got a lot of people doing things in different ways. It makes it very hard to create consistent management patterns.”
Sarah Cooper, chief operating officer for M2Mi, sees three main areas where open technology is enabling Internet of Things
1. Infrastructure scale: Open technologies like Kafka are making large data sets more manageable, and allowing for more useful data to be collected and used. OpenStack and Cloud Foundry have also been proven at IoT production scale.
2. Distributed systems: Technologies like Spark for streaming distributed analytics components are making it easier to manage distributed systems. Riak TS from Basho allows queries of large databases for IoT and time series data separated over geographies.
3. Operating environment management: Businesses with large IoT device installs such as farmers or shipping fleets don’t always have robust IT departments to support these deployments. Open source tools like Docker help them operate in the infrastructure more easily.
Open technology isn’t the only segment needed to support Internet of Things. As the technology becomes more complex and allows for more opportunity, a growing list of companies will have to work together across the value chain in order to improve user experience.
“[IoT] requires a great deal of agreement from disparate partners that are often frenemies in the ecosystem,” Cooper says. “IoT has a challenge to get these partners to collaborate effectively.”