It's All About the Data: The Internet of Things Raises Connectivity to a New Level

<< back Page 3 of 3

Bookmark and Share

IoT requires a new type of data and IT management talent and new types of platforms. “Enterprise NoSQL databases will be the platform for these applications, but who will develop them?” asked Gorbet. “Analysts predict that the industry will need 4.2 million IoT application developers by 2020. This is a huge opportunity for developers who understand how to work with NoSQL technologies.”

Experts agree there is plenty of work ahead in assuring connectivity between devices, sensors, and networks. “Enterprises are still finding it difficult to connect from the enterprise and cloud to devices in their network,” said Ross Mason, founder and VP of product strategy for MuleSoft. “When IoT products, devices, and services aren’t connected, valuable data becomes siloed. An ‘Internet of Things’ becomes a ‘Just a Bunch of Things’ initiative when these so-called connected devices can’t even connect to each other.”

Shovel-Ready Components

There are some pieces of data environments that are ready for IoT, starting with siloed industrial systems consisting of legacy applications, including SCADA and ICS, telematics, other control systems, and homegrown applications, said Brian Gilmore, solution expert for the Internet of Things and industrial data for Splunk. “These are probably the most shovel-ready for enterprise-scale IoT implementation,” he said. “With the right platforms and software, these systems can be responsibly connected and monitored with limited complexity—even when there are multiple generations of systems and proprietary communications protocols.”

The network is the single most important part of building an IoT-enabled environment.

Other components that can be readily connected to an emerging IoT include the “dark devices on every enterprise network—printers, IP phones, Wi-Fi hotspots, and other networking equipment,” Gilmore added. “Collection and analysis of the diagnostic data from these devices can give enterprises incredible insight into utilization, availability, and capacity, and can help optimize the operations of the enterprise as well as protect the enterprise from cyber-threats.”

Connected products that keep streaming data back to the company that produced them is an additional area of opportunity for IoT development. “Real-time access to data from connected products provides unprecedented intelligence into consumer behavior and enables businesses to provide additional proactive services to customers and to proactively service and update the products in the field,” Gilmore said.

Stitching It Together

There are several ways to stitch together a network of IoT-ready devices or applications, but open APIs may be the most effective route to assuring greater connectivity between enterprises and their devices. “APIs allow disparate systems and data to talk to each other, creating a seamless flow of information from one source to another,” said Mason.

Along with APIs, IoT requires the latest technology approaches, including cloud, mobile, and collaboration tools. “These are absolutely necessary for an IoT strategy,” said Lasser-

Raab. “Without the wireless connectivity and mobile devices, we wouldn’t have the distributed networks that connect much of today’s world. Without cloud, data storage and analytics wouldn’t be possible. Without collaboration and social tools, we couldn’t share this information in real time. The network is the backbone of IoT, and it is the single most important part of building an IoT-enabled environment.”

The ideal IoT architecture combines three key capabilities, according to Andrew Rogers, CTO and founder of SpaceCurve. “First, it must be able to continuously ingest and store extreme large volumes of data at rates of millions of records per second. Second, the architecture needs to be designed for complex spatially organized data models and support massively parallel spatiotemporal analytics; this has traditionally been a weakness of big data platforms when used for IoT applications. Lastly, the architecture needs to support real-time analytics that can seamlessly blend the live, streaming data with data that has already been stored.”

There has been a lot of attention given to the non-relational data coming into enterprises in recent years, and IoT promises to multiply the volume and variety of this data many-fold. Ideally, an IoT architecture needs to “allow for operational and non-operational data, historical data, and business data to be potentially hosted in a private cloud and the ability to run real-time analytics through a BI platform,” advised Shamlan Siddiq, VP of architecture and application development at NTT DATA. “While big data is a buzzword, it would be prudent not to ignore its practical uses depending on the data size that needs to be processed and rendered through NoSQL database type solutions. The presentation layer should also involve a multi-channel experience that includes devices, web, social, and sensors. It is additionally important to ensure that the architecture include a distributed message bus, data APIs, orchestration layer for data services, and an abstraction layer that connects to the underlying storage, and infrastructure.”

In a matter of time, the industry will converge around common platforms and standards to support IoT initiatives. “Today’s patchwork of separately evolved, vendor-specific, and proprietary infrastructure will be replaced, over time, with interoperable platforms,” Banerjee predicts.

Bringing all these disparate pieces together will require a well-designed and agile architecture internal to enterprises, incorporating the latest and most flexible technologies. “Interacting with fast data generated by the IoT demands a different process than interacting with data at rest,” said Jarr. “The ideal environment will transform IoT data into actionable insights, converting fast streams of data—often, hundreds to millions of times per second—to an action in real time, and do it on a per person per event basis.”

The potential for IoT to create massive amounts of information “spotlights the need for organizations to develop an information governance plan relating to IoT data,” said Philip Favro, senior discovery counsel for Recommind. “Any such plan should include a strategy for identifying information that must be kept for business, litigation, or regulatory purposes while segregating other materials for eventual deletion. It should also include actionable steps to ensure compliance with the privacy expectations of domestic and cross-border data protection laws. Such a proactive approach will help organizations dodge disaster and be prepared when the IoT onslaught materializes.”

In addition, today’s new generation of databases is “fast enough to process streaming data in real time with analytics and decision making, greatly simplifying the effort and architecture required by developers,” said Jarr. Employing analytics against offline systems is no longer a valid option, he added. “By the time the data has been analyzed, the opportunity is gone. The infrastructure needs to perform in the same manner as the instantaneous behavior associated with IoT, and support the ability to make real-time recommendations that are event-triggered.”

<< back Page 3 of 3

Related Articles

HP has introduced a new offering called HP Haven Predictive Analytics to accelerate and operationalize large-scale machine learning and statistical analysis.

Posted February 18, 2015