Jun 26, 2016Enterprises small and large currently face the fundamental challenge of digitizing their business and integrating business processes with their physical footprint and virtual presence on the web. Accessing and analyzing sensor network data provides real-time visibility of an enterprise's capital and human assets. Digitizing workflow and processes propagates the chain of value creation down to all levels of an enterprise, including sales, marketing, operation, customer relationship management (CRM), supply chain management (SCM), information technology, and research and development—uplifting an enterprise's revenue and margin with a typical payback, in most cases, of less than a year.
Smartphones: Driving the Internet of Things
At present, the Internet of Things (IoT) is primarily powered by crowdsourcing data from smartphones owned and operated by a company's personnel or customers. Smartphones are the most ubiquitous remote sensing hardware platforms, offering more than a dozen types of sensors tightly integrated with a mobile device's operating system. Once a smartphone is paired with a new generation of wearable sensors, a rich dataset can be used to locate, monitor and control a vast range of indicators for businesses.
The number of connected IoT devices is forecasted to surpass 25 billion in 2020. Real-time sensor data, in conjunction with open-collaboration software platforms, has already been transformational in key niche markets.
In enterprise markets, wireless sensor networks (WSNs) are already a critical component in high-value use cases. Location-aware sensing can enable an enterprise to digitize business processes over space and time to match demand and supply at each step in the production or delivery of services. In sectors such as automotive, aerospace, logistics, precision farming, transportation and health care, sensor networks have an established track record of providing a high return on these investments. It's noteworthy that so-called "unicorn" companies like Uber or Lyft are hosted sensing services for customers and drivers leveraging their smartphones. The IoT is a key driver pushing innovation toward a fully autonomous car. This transformation made in transportation is now catalyzing other crowdsourced business models and new ways of doing business with sensor data.
Sensor Technology: Growth and Innovation
Buyers are faced with bewildering choices: IoT hardware for sensing, an edge appliance for collecting sensor data at a local site, and software to process, store and analyze the data. Unleashing business value with actionable insight from IoT data is highly dependent on seamless integration of all three components to effectively consume sensor data and provide an end user with the correct information at the proper time. With the broad array of sensor types being offered, these choices have created a great deal of confusion on the part of buyers to purchase the right "thing" for their IoT deployments.
There is also a mismatch in supplier and customer expectations in terms of cost, installation, capability, support and services for deploying and maintaining large wireless sensor networks. Enterprises with mission-critical data cannot afford any downtime. The engineering and deployment of large wireless networks requires an expert chain of support to assure full coverage and 24-7 fault-tolerant operation. In order to assure success, buyers must require their suppliers to have solid customer support and services throughout the lifecycle of their IoT investment.
Hardware is proliferating for multi-modal sensors with GPS, gyro, accelerometer, temperature, camera and a variety of other modalities. Given the array of options to design and deploy wireless sensor networks, an end user's first decision is the choice of sensor device to deploy. The right hardware selection begins by understanding the technology's underlying capabilities versus expected metrics to fulfill a user's particular business needs. The chosen hardware should always be based on industry-wide standards, with availability from multiple suppliers. End users should avoid proprietary single-sourced devices.
Moving forward, we will continue to see a great deal of innovation in sensor technology at a brisk pace. There will be a world of heterogeneous sensor networks, as there is no such thing as a single "panacea" sensor to solve all problems. At the device level, there is rapid ongoing progress in nanotechnology and condensed matter physics to offer printable sensors with batteries, transducing different forms of energy from one type to another, such as mechanical to electrical and chemical to electrical. These new generations of sensors will shorten the product lifecycles of existing generations. With the rapidly evolving IoT landscape, a buyer is compelled to use best-of-breed technology in remote sensing. Suppliers need to assure buyers that they will support new and emerging sensor types, to prevent their being boxed in with a single technology with no viable upgrade path.
Passive sensors that harvest energy to operate will be instrumental in transitioning from an analog to digital business era. Passive ultrahigh-frequency (UHF) RFID technology has matured throughout the past decade to reliably unleash the digital identity of any physical asset, on demand and over the air. In a fully digitized enterprise, every physical asset—from conference rooms and hotel rooms to cars and warehouse bays—can autonomously report capacity, utilization and availability in real time.
With the fusion of active sensor types (ranging from cameras to simple Bluetooth devices) with passive sensors and RFID tags, machine-to-machine (M2M) interaction is changing business processes well beyond industrial instrumentation. With physical objects sensing their environment and communicating their state, software platforms can make intelligent, autonomous decisions based on real-time spatial-temporal events. In most applications, sensor fusion is formed by a dedicated server at the edge of a network, sometimes referred to as an IoT gateway, for the purpose of "fog computing" (a term coined by Cisco). A buyer needs to evaluate the total cost of ownership for enterprise-wide deployment for a total solution. The overall cost is largely driven by the suite of software components to complete an end-to-end solution.
Buyer Considerations: Security and Hosted Services
Many companies currently offer cloud-hosted middleware for the IoT, as well as some on-premise software solutions. These offerings bridge the sensor data stream from the sensor layer to enterprise back-end ERP/CRM systems. When choosing a cloud-hosted solution, a buyer must take into account some key considerations.
One key concern with using the IoT has been security. For a bulletproof system, an enterprise can choose an on-premises option, in which all data is collected and maintained locally on each single site in a classified militarized zone or fully decoupled subnets. With the right choice of software, authentication and access control may be enforced even at the finest granular level of each flow sourced by every sensor.
After more than a decade of effort by open source and industry, big data architecture for storing large volumes of unstructured data has been solved. Options range from hosted big-data services, from the likes of Amazon, Microsoft or Google, to open-source options such as Apache Hadoop for storing and processing data up to web scale. There is also an emerging trend for distributed sensor databases residing at the edge of a network and accessed only on an as-needed basis across an enterprise. Buyers should select a supplier to provide them with a migration path to distributed databases as their business needs evolve.
During the past four years, cloud computing has undergone an accelerating growth cycle, primarily catalyzed by the open-source community and growing demands for different types and classes of web hosting services. In legacy cloud-computing software services for virtualization, physical hardware is dedicated with pre-defined memory size and computing resources to application software. The extension of ideas in hypervisor in the Unix operating system has ignited a paradigm change with the introduction of the Docker container. Instead of hardware virtualization, a new approach is centered on virtualizing the components of the application software itself.
The software container embeds all of the components typically in a server, such as file systems and libraries. In addition to providing a much-needed boost in scalability and throughput using the same hardware platform, the underlying model becomes independent of the cloud-hosting service provider. This is a key enabler for the application to run on any hosting service, such as Amazon Web Services (AWS), Microsoft Azure or Google. This model relaxes the constraint of being host-service dependent and running only on a single service.
Another frequently overlooked—but key—capability is that startup time for containers takes less than a second. This is highly beneficial for sensor data streams with non-stationary flow characteristics. From an operational perspective, distributed software applications with Docker substantially reduce the number of systems required, due to each container's small footprint. All of these benefits combined lend themselves to a win-win business scenario for both a buyer and a supplier to pay for exactly what a customer uses, as opposed to a theoretical physical hardware, sized up for maximum offered load.
In selecting the proper software, a buyer needs to know if the solution conforms to modern software design patterns for web services and cloud hosting—and, most importantly, whether it is designed and developed by experts in wireless sensor networks. Knowledge of WSN in different propagation environments, as well as the proper engineering of these networks for mission-critical business processes, is a requirement for the successful deployment of a large network of IoT devices.
Ease of configuring, provisioning on demand, and real-time monitoring and control of large WSNs require specialized capability in the software to assure scalability within an enterprise to grow from point solutions to multiple locations with a global footprint.
When IoT data is mashed up with data streams from public sources (weather data, for example) or private (such as Nielsen ratings), it can enable an enterprise to optimize operational efficiency and, ultimately, the bottom line. With the right choice, a buyer should be able to seamlessly complement his or her data with public or private data feeds, for the purposes of visualization, data mining, and behavioral or prediction analysis.
Internet of Things: Transforming the Business Model
Growing capabilities of virtualization in computer science, coupled with IoT technologies, are coming together to transform the enterprise into a digital projection of its business model. Currently, the IoT creates the ability to digitize, sell and deliver physical assets as easily as with virtual goods. Services presently characterized as platform-as-a-service (PaaS)—infrastructure-as-a-service (IaaS) evolving from the software-as-a-service (SaaS) model—provide buyers with alternatives to minimize up-front capital expenditures for digitization of their business with the IoT.
We are entering an exciting period of IoT adoption with accelerated growth. Innovative disruptions are to be expected—from the sensor device level to the edge of the network to the very core of the network, with web-scale-size services supported by a rich set of data mining and machine-learning tools. All of this is uniquely transforming the process of digitizing businesses with wireless sensor technologies.
Ramin Sadr, Ph.D., is the founder of Mojix, a provider of RFID and IoT solutions. With more than 25 years of experience in the communications industry as an entrepreneur, researcher, executive and lecturer, Sadr has launched and led successful high-tech businesses based on his technology innovations. He holds numerous patents and has received 15 NASA achievement and recognition awards for his contributions to the U.S. space program.
