Dan's Take

The Internet of Things: Take Your Time

Before jumping in, find out what the water's like.

The folks of Liaison Technologies (LT) and I exchanged a number of messages while I was conducting research for an upcoming Virtualization Review article on the Internet of Things (IoT). They believe that enterprises are rushing far too quickly into IoT without giving the move proper thought or consideration.

The following comments come from this lively exchange of ideas:

Dan Kusnetzky: I advocate that enterprises consider what they're really trying to accomplish before entering the design phase of any project. What's your view?
LT: In the rush to implement IoT and capitalize on the terabytes of data being made available from devices and smart machines, many companies have overlooked a major component of the process: exactly how will they will integrate and utilize the data?

What are your thoughts on managing the ever-growing herd of intelligent devices that are part of any IoT program?
Devices are popping up everywhere, and it's easy to collect and store the data, but very difficult to make it meaningful and actionable.

Why? The lack of interoperability, integration and harmonization of Big Data originating from multiple devices and applications is a major obstacle. The number of sources, formats and tools for capturing and analyzing data is growing exponentially, but each device or application creates its own proprietary system. Resulting data silos must be bridged in order to gain a complete picture and real business insight from all the data now available to an organization. Unstructured and semi-structured data each require different systems, and unless a company goes out of its way to develop a polyglot solution, it's impossible to sync up with current databases.

IoT success requires complex event processing and data streaming in real time; data comes in and you need to be able to react instantly. However, this requires an intricate dance between integration and data management that's simply not facilitated well by conventional solutions that handle the critical functions of integration and data management separately.

What significant challenges do enterprises face here?
Conventional schema on write systems are also unable to keep pace with the influx of IoT data. Instead, organizations need a schema-on-read approach that enables data to be stored in a central repository, and then accessed and read across virtually any application and in the format required by each particular use case. 

The ability to integrate and map various data sources and applications together is critical for companies to take full advantage of the powerful analytics tools and insights available. In order for companies to realize the full promise of IoT and Big Data, they must focus on integrating and managing data from the full gamut of data sources in a single, unified, comprehensive and harmonized platform.

IoT devices are producing a huge amount of data. Do enterprises need to capture, clean, store and analyze all of it?
It's also important to note that not all data from IoT devices is useful or necessary. Many companies want to collect all the data, whether or not they have a valid use case. Companies must begin to think in terms of use cases to maximize data collection and analysis efficiency. For example:

  • Apple Watch or FitBit data can be used to improve health care delivery and population health. Beyond just collecting data at the individual level, it can be aggregated in a de-identified way to spot population trends in symptoms, behavior and disease states.
  • In the energy industry, IoT devices could be used to provide real-time performance and environmental data from wind turbines to understand their capability in generating power, their wind speed and other data points to optimize performance.
  • In the financial industry, credit card companies can use streaming data and IoT to make instant decisions about transactions that could be suspicious or fraudulent, to protect customers and the credit card companies.

Once enterprises have opened up the IoT data spigot, what are the key issues you see?
There's been an overwhelming focus on developing IoT devices and the requisite platforms, applications, databases and APIs that make them work, but little or no focus on how to integrate and manage the data from these devices across the silos they create in order to make this data more useful and impactful. 

It's time to enable these systems and data to be integrated and leveraged for greater business insight through a unified Data Platform as a Service (dPaaS) that enables both the integration and management of data from virtually any data source, device, API or application.

Allowing a growing herd of intelligent devices to access corporate resources appears to open up a huge potential for security problems. What's your view?
Data security and compliance are also paramount considerations for companies preparing to leverage the IOT. As companies roll out new devices, they must comply with stringent regulatory standards such as HIPAA in the healthcare industry that are related to data capture, use and sharing. A dPaaS solution that unifies both integration and data management in a single, inherently-compliant platform eases the burden and costs of maintaining compliance for organizations.

Achieving success with IoT data integration and management requires a cross-functional effort. The CIO, CEO, researchers, domain experts, data scientists, and others must all work together to devise appropriate use cases, bring the data in sync, and integrate it for maximum usefulness and benefit across the organization.

Dan's Take: Try to Stay Dry
I'm reminded of something that I've seen attributed to Benjamin Franklin, Alan Laken and a few others: "Failing to plan is planning to fail."

Enterprise decision makers often get caught up trying to implement the "trend of the minute" and fail to really think through how a new approach or technology would fit into their IT infrastructure. Sometimes the new technology really won't fit or add significant value, and should be avoided until it matures. It would be far wiser to consider what needs to be done before launching into a project to do it. It would also be better to consider how this new technology or approach fits into the overall enterprise IT architecture, rather than picking a tool first. If selected first, tools have a way of limiting enterprise thinking, tying them into narrow approaches and, sometimes, restricting future choice.

IoT is getting a lot of attention right now. Like past trends, some are jumping in, rather than just putting a toe in to see how it feels. The jumpers face the possibility of getting all wet, rather than accomplishing their goals.

About the Author

Daniel Kusnetzky, a reformed software engineer and product manager, founded Kusnetzky Group LLC in 2006. He's literally written the book on virtualization and often comments on cloud computing, mobility and systems software. He has been a business unit manager at a hardware company and head of corporate marketing and strategy at a software company.


Subscribe on YouTube