How reliable is the Internet of Things infrastructure?

By Jane Irene Kelly

The Internet of Things, the rapidly evolving state of computing where everything from refrigerators to energy grids are connected to the Internet and communicating with each other, will create new challenges for data centers, says a recent Gartner report.

Storage capacity, scalability and inbound data-center bandwidth requirements are just some issues organizations are already facing as they struggle to manage growing volumes of data, including smart sensors, tablet computers, wearable devices, applications and more.

Efficient systems management will be critical to maintaining data centers in an Internet of Things world, where the convergence of IT systems and operations technology systems – two areas that have long lived blissfully apart – will be crucial.

Solutions such as software-defined networking (SDN) promise to help IT teams manage converged network infrastructure easily and centrally, allow for dynamic and massive scalability and elasticity of data centers to support Internet of Things applications, and enable automation of such tasks as configuration and policy management.

These solutions are emerging, but many organizations aren’t paying attention because they’re too focused on collecting big data.

What can we automate?

Mason Katz, chief technology officer and co-founder of StackIQ, a provider of software-as-a-service for managing big data computing systems, says big data distraction is preventing organizations from asking one important question that will help them prepare for the Internet of Things: “What can we automate?”

“If you’re already underwater with systems management because of big data, things are only going to get worse if you don’t take action now,” he says. “Stop worrying about optimizing systems. Focus IT resources on developing apps to manage unstructured data.”

Katz says that apps are needed not only to transform big data into meaningful information for the business, but also to make network infrastructure smarter and more efficient to meet the demands of the Internet of Things.

Some experts say that proactively developing a refined strategy for handling unstructured data will help to alleviate already growing pressure on the data center. They say that companies should focus on data not devices.

Yet organizations should also not fixate on accumulating large quantities of data for its own sake; rather, they need to be selective. Experts say that it’s important for companies to ask themselves what data they’ll need to achieve their goals.

What cloud are you?

Once organizations decide what data they want, Katz says the next step is deciding where and how that data should be managed: in a public cloud, a private cloud, or a hybrid cloud.

“The sheer volume of data the Internet of Things will produce, the low latency and quality of service that users will expect, and security and compliance concerns for the business mean everything needs to move back into the data center,” Katz explains. “Big data is making systems huge, so you can’t manage everything efficiently or automatically through traditional data centers. You need the cloud.”

Determining the business model

Catherine Spence, technical coordination committee chair for the Open Data Center Alliance, a consortium of leading global IT organizations, and principal engineer for Intel IT’s cloud program, says companies also need to determine what their business model will be in the Internet of Things because it will have a direct impact on their systems management approach.

“From a provider standpoint, businesses need to decide who will be managing their Internet of Things applications,” she says. “If you decide to take the primary responsibility for managing the data being produced, then you need to be prepared to massively scale out your data center and bring it back again at different times of the day based on how users use devices and applications.”

Spence sees automation as an important investment area for organizations that want to prepare for the Internet of Things. “Highly automated systems need to sit behind the apps. Real-time apps will be core to the Internet of Things, and the quality of service needs to be there. Devices will be constantly producing small communications that need to be prioritized, and that requires automation.”

“The Internet of Things will only be possible through good infrastructure — invisible, fully automated infrastructure that essentially takes care of itself,” says Katz. “It’s not ***, but it’s what’s needed. To get there, organizations wandering aimlessly in the big data ’candy store’ need to stop and think about how many IT resources they’re putting into their data center right now, and whether those resources could be working smarter to help them prepare for the Internet of Things.”

About the Author: Power More