Internet of Things – Can it be more than a Technology Phenomenon?
Folklore has it that the term ‘Internet of Things’ (IoT) was first popularized in 1999 at MIT to describe the architecture of connected RFID devices. Cisco then looked to define when the IoT came in to being as a concrete entity – defined as the year in which the number of connected devices exceeded the number of connected people (somewhere between 2003 and 2010 – see diagram).
Click to zoom
Cisco predicted that the number of devices will grow to 50 billion by 2020, a trend substantiated by many. And over the years, the use of the term ‘Internet of Things’ has been enlarged to describe connected devices across any type of communication technology, regardless of whether this is public Internet or private networks – the Internet of Everything.
Yet Cisco also reports that the Internet of Everything is failing to deliver on its revenue potential. Much like other exciting technology trends, such as semantic analytics, is it possible that this is technology-driven excitement looking for a business problem to solve? An understanding of the value chain and related business models is required. For example, Telco M2M applications have been hindered by the cost of wireless backhaul, which given the low ARPU levels for M2M services, has meant M2M services have struggled to achieve mass adoption. Although the cost of optical backhaul capacity has dropped significantly over the past 10 years, it’s still the ‘last mile’ access that is the constraining cost factor.
Technology makes the possible probable
Just as the Big Data movement was enabled by the combination of technology drivers (increasing volumes of unstructured data, low cost commodity servers and a thirst for faster answers), similarly a number of different technology advances have converged to enable the Internet of Everything:
- Low cost ‘smart’ wireless sensors
- Significant advances in energy and battery technology
- IPv6 capability and associated network devices
- Proliferation of smart phones and instrumentation devices
Yet so far we have seen the cart driving the horse. If we look at M2M again, still seen as the great hope for mobile carriers looking to move further up the value chain, it has failed so far to live up to revenue and growth expectations (at least, as forecast by some). Healthcare (and micro-sensor monitoring) applications are starting to emerge, but require external investment in order to be viable, the connected home has several conferences dedicated to it, yet not much revenue, and Smart Grid applications have been hindered by the existing distribution infrastructure.
Yet there is hope. GPS-based solutions for Intelligent Transportation applications – traffic flow and congestion prediction maps, real-time Journey Time apps – have seen an increase in adoption as M2M services. Consumers traffic apps, logistics and asset tracking, as well as partnerships with automotive and Transportation agencies for telematics. The network for IPv6 traffic is also evolving, with technologies such as Software Defined Networks (SDN) reducing the cost of network management and capacity optimization.
Real-time Performance, Health Monitoring and Analytics
The platforms for IoT/IoE applications have a several major components, the main three being the SDP for delivery and definition of the services, billing and revenue generation, and real-time performance and health monitoring with real-time analytics. The latter is effectively a streaming operational intelligence problem, requiring:
- Remote, intelligence data collection architecture, with the ability to perform remote filtering and analytics, therefore improving the signal to noise, and reducing the backhaul capacity requirements.
- Real-time analytics, predictive analytics and prescriptive analytics – the ability to extract actionable intelligence form the streams of device and sensor data, but being able to predict device failures and operational exceptions in advance, and where possible, to automate appropriate actions.
- A massively scalable platform capable of managing the data collection and reuse of all real-time data streams, and ensuring the reuse of raw and derived data streams across any number of applications executing on the platform, such as performance, health monitoring, GPS analytics, streaming aggregation and backend integration.
Total Cost of Performance for IoT Platforms
The cost of achieving real-time solution performance remains an important consideration. SQLstream’s streaming platform delivers the lowest total cost of ownership for real-time applications across all data management technologies. Both Hadoop and traditional RDBMS-based systems suffer from both a technology tipping point for latency and throughput, beyond which it is virtual impossible to improve performance, but even before this point, the cost of real-time performance makes these solutions unfeasible. There are many reasons, but the underlying architecture is simply not scalable for high velocity, low latent analytics – the principle of store first, analyze second introduces unacceptable delays.
A performance benchmark for SQLstream’s real-time operational intelligence scalability, a system for 4G wireless network performance monitoring, discusses the core streaming architecture scaling to 1.3 million records per second on a standard 4-core server, and an overall architecture of 10 million records per second over only a few servers. The Total Cost of Ownership was the lowest by some margin (based on an independent benchmark program).
If you would like to know more about real-time operational intelligence as the performance platform for the IoT, register for our next webinar “Listen to your Sensors“, where we’ll be discussing the requirements and business case for large-scale intelligence sensor network management.