Everyone talks about the Internet of Things’ exciting innovations. But far less attention is paid to its potential pitfall: data overload.
That’s an area Tiffany Sargent focuses on as a principal engineer and IoT senior solutions architect with Intel Corporation’s Federal Sales and Marketing Team. In helping federal agencies adopt IoT technologies, Sargent hones in on a critical differentiator: the ability to transfer the right information, to the right people, at the right time.
In other words, the delivery of more information, less data.
“In the future … it really means changing data into information – that is an important point,” says Sargent. “As we talk to different agencies, we are seeing the same need over and over again.”
As is, the current approach to data ingestion rests on a typical (and very costly) approach: Sensors ingest data – and whether valuable or not — transfer the complete data set across the entire system into the cloud. Only then does analysis and processing occur.
Intel is working to change that status quo by collaborating with partners to create a new strategy and location for computing data –at the edge — and only then, passing on relevant information to the cloud.
That’s where technologies such as Intel IOT Gateways come in. “There’s an opportunity for agencies not simply to collect and send data, unprocessed, directly to the cloud, but first to ‘compute’ – using technology such as a smart device, or an edge platform like a gateway,” says Sargent.
“This [departure] turns data into useful information closest to the point of data ingestion,” says Sargent, “so you are able to do a compute, and only send a fraction of data to the cloud — that data, which is now information, needed for relevant decision-making.”
But key challenges loom in IoT system design. “There are whole new areas of complexity and constraints that need to be figured out in IoT architecture implementation,” says Sargent.
The first challenge centers on connectivity. “How do we make sure that as architectures and systems are designed, the right connectivity choices are made?” says Sargent.
Manageability is also critical. “As more IoT devices show up, how do we make sure we can identify them, then manage them seamlessly and at scale. Simply pushing an application update takes on new meaning,” says Sargent.
Then there’s the security aspect. Adopting this new technology not only changes the management of data and its associated costs, but also changes the security strategy. “Now all the compute doesn’t necessarily have to happen in the cloud, but at the earliest point of data ingestion,” says Sargent.
“However, this doesn’t mean simply securing a device – it is about the entire system,” adds Sargent. “For example, if a device is going to join the Internet of Things system, can it be verified and validated in a secure way and then, is the data coming from that device going to be securely transmitted?” To address such concerns, security needs to be included starting with the initial design of any architecture, and implementations need to be carefully thought through, says Sargent.
Finally, interoperability can’t be ignored. “It’s not just interoperability where you know all the systems are hooked together; its complexity extends into data interoperability – knowing if the right data is being fused together to create contextually relevant decisions,” says Sargent.
Tackling all four challenges requires a two-pronged strategy.
“From a technical approach, Intel has created reference architectures and platforms that allow collaboration to be easier with its partners and customers,” says Sargent. “Intel leverages the strengths of these partners to engineer solutions – with recommended architectural features around connectivity, manageability, security and interoperability.”
“On the business side,” adds Sargent, “to create interoperability you need to partner with companies that have expertise in modernizing and extending existing technology solutions and infrastructures — IoT implementations will not be rip and replace but a crawl, walk and run, leveraging new technologies that replace parts of older, obsolete systems.”
“What we are seeing,” says Sargent, “is an entire ecosystem of partners emerge for any integration to be successful – this is not a ‘one company does it alone’ implementation.”
Toward that end, Intel has partnered with next-generation data management providers, including newer names like the U.S. -based software company Cloudera.
In addition, Intel has participated extensively with the National Institute of Standards (NIST) Cyber Physical Systems Public Working Groups, creating cyber physical systems (CPS) specifications, while also working in parallel with publications developed through other industry consensus consortia and standards bodies, such as IIC, OCF, JTC1, IEEE, ESTI, AIOTI and AII.
Meanwhile, NIST is taking a lead role in harmonizing CPS specifications with these other efforts via the industry liaison working groups, driving a scope of standards for smart cities and other public IoT implementations.
With clearer paths towards interoperability come clearer results – and patterns. “Now we see families of problems and we know what the solutions are,” says Sargent. “We now also have more robust building blocks that we can reapply – the end state will be faster, and better-known results, as an overall IoT community.”
Minus, the data overload.