It is naive to think that the role of integration in Internet of Things (IoT) is largely confined to machine-to-machine (M2M) integration.
Unless specific data collected by IoT devices is processed and analysed to derive meaningful insights and/or develop innovative applications, it is difficult to attach any significant commercial connotation to enterprise IoT initiatives.
The cost and complexity of integration continues to be a major roadblock to wider IoT adoption.
IoT integration is a complex and multifaceted issue
As per the Eclipse IoT Working Group’s IoT Developer Survey 2015, “interoperability,” “integration with hardware,” and “M2M connectivity” are among the top-four concerns about the development of IoT solutions.
Of the 356 respondents, 31 percent selected “interoperability” as the top concern, and this figure increased to 39 percent in the case of developers focusing on enterprise software - this is good evidence considering that respondents were experienced developers.
Moreover, as per the World Economic Forum Industrial Internet Survey 2014, “lack of interoperability or standards” is the greatest barrier inhibiting Industrial Internet of Things (IIoT) adoption.
The same survey indicated the “need to converge on standards to support better interoperability” as the second most important action item for hardware, software, and service providers in the IoT ecosystem.
IoT interoperability issues are compounded by an unrelenting tussle between interested parties, including IoT platform vendors, major software vendors, industry consortia, communications service providers, and device manufacturers, pushing the case for a specific M2M/IoT communications standard.
Some industry pundits and devoted proponents have failed to see the big picture, and continue to add confusion to the ongoing debates.
Although certain M2M/IoT communications standards have an edge over others in the first phase of adoption, Ovum advises IoT initiative leaders to refrain from any “religion” around specific standards.
IoT dramatically increases the number of integration points involved in the process of realising a logical output from a complex ecosystem of connected things, users, and enterprise IT stack.
The data collected at various endpoints of integration flows needs to be analysed along with the “context” of usage to deliver actionable insights (e.g., real-time intelligence based on identified patterns/trends), and functionality based on interactions between connected devices and applications.
This calls for elastic scalability to support processing of a large number of low-latency messages per unit time (potentially millions to billions of messages).
In simpler terms, IoT involves high data velocity and event throughput, with acceptable latency measured in microseconds.
A complex IoT solution stack can involve device-to-device, device-to-server, server-to-server, device-to-user, and server-to-backend communications, which require different communications standards, integration platforms, and infrastructure resources.
There is no standard IoT solution stack capable of supporting a wide range of IoT uses cases.
Enterprises therefore continue to use a combination of traditional middleware and emerging M2M integration approaches for meeting the requirements of specific IoT use cases.
Cloud-based infrastructure and platform services offer elastic scalability and can support high-throughput event processing, one of the key requirements for realising value from IoT.
Although both commercial and open-source middleware vendors are developing IoT offerings to fill the gap between M2M and enterprise integration, these are early days, and no vendor enjoys any significant competitive advantage.
By Saurabh Sharma - Research Analyst, Ovum
