Gartner identified advanced, pervasive and invisible analytics (APIA) as a trend for 2015. Not only is the Internet of Things (IoT) set to bring more data into the enterprise, it’s also pushing IT and operational technology (OT) toward a much different relationship paradigm.
Operational challenges of adopting IoT
The two sides of the house—IT and OT—have been organizationally separate in most companies. “They use different technologies, they generally have different goals and there has been a lot of decoupling between the business side and the operational side,” said David Barnett, vice president of products and markets at Real-Time Innovations (RTI).
Even the standards between the two sides are different. IT’s well-established standards are now meeting OT’s more proprietary, often more fragmented set of standards. The highly centralized architecture of IT must be made to connect with the decentralized approach used by OT. “It’s definitely challenging to physically integrate the systems,” Barnett said.
“Machine and sensor data is becoming a first-class data citizen in the enterprise, the way account and customer data has been for decades,” said Bill Zujewski, senior vice president of marketing at ThingWorx, a PTC business. But because machine data occurs differently than the data IT has been handling all these years, netadmins will need to find a way to manage it not just when it’s at rest, but when the data is in transit, too.
“It can be very sporadic in how things connect,” Zujewski said. Simple network disruptions could spell problems. “When the network is down, the data queues up at the edge,” Zujewski explained. As those connections are restored, the network and its data center are likely to be bombarded with enormous data sets from all directions.
As interconnectivity issues loom, security is emerging as a challenge of its own. Securing a multitude of devices, many of which weren’t originally designed to support any sort of security natively, will require a different mindset and a different set of tools than what has historically been used to protect the enterprise. “You’re not going to be able to put your traditional antivirus software on many of these IoT devices, because they don’t have the 200 MB of disk space required to put your normal antivirus software on it,” said Ido Sarig, vice president and general manager of the IoT Solutions Group at Wind River. “Some of them can’t afford the additional 90 seconds or so that booting up requires when you need to get the antivirus software in place.” Firewalls and other common security measures simply might not work. Netadmins may instead need to look at the gateways and other added-layer security solutions that are coming into the marketplace to address the issue.
Internet of Things will change IT’s relationships
Implementing processes to gather and transfer data, along with deploying analytical tools to mine that data, will now happen with input from more groups than ever. To make things function efficiently, Zujewski sees increased collaboration between IT and OT. “IT will almost become the liaison between business function and R&D,” he said.
Harvesting data for pay-per-use billing, a common scenario in healthcare and a good example of where IT’s involvement is likely to evolve, often involves IT being asked which datasets are required for the hospital to bill based on usage. To get the answer, IT may need to go back to the manufacturer to learn which type of data is needed and how to gather it so it can then be passed on to the finance group. Zujewski described IT’s role in this IoT situation as “becoming the glueware, the middleware.”
There’s a good chance IT and OT will go through some level of convergence. “That’s going to require cooperation on both sides,” Sarig said. This is already happening in some companies that have moved down the IoT path, and Sarig said it occasionally results in pushback from IT. “Part of it is that IT is feeling people are stepping on their toes and encroaching on their territory.” But where IT takes that stance—that they’re the ones with expertise in data collection and network security and the OT group just runs the plant—Sarig said, “It’s going to lead to conflict, and it’s going to lead to OT just doing their own thing.” If IT isn’t helpful, if they don’t secure their place at the table, they may simply be run over.
Fortunately, Barnett said he’s seeing many attempts to meet in the middle. “Some of the things we’ve been working on for the last few years are basically building bridges,” he explained. Most of those are technology bridges, but they go hand in hand with organizational bridges. Knowing where and how to overcome challenges, such as making connected devices speak the same language as the network, is key.
“You can’t call an IT person every time you move a device that’s connected to a different station,” Barnett said. “It just wouldn’t scale or be efficient. So you need things that are self-provisioning effectively.” This is a much different approach than what IT is used to, where laptops and other connected devices are each configured and managed. Creating a partnership between IT and OT will be necessary to deal with broad issues as well as these more garden-variety concerns.
Put the pieces in place to embrace IoT
There are a number of proactive steps netadmins can take right now to be better prepared for the challenges IoT will bring. Having a plan to deal with scalability issues is one. “Traditional IT systems are very much oriented on human users and human speed data and human data volumes,” Barnett explained. “When you’re dealing with real-world or cyber physical systems or sensor data, usually you’re talking about orders of matter magnitude more data.”
Historically, the strategy to deal with that type of scenario has been to reduce the data in order to maintain performance. Barnett said that’s not optimal with IoT, because much of that data is desirable if the organization wants to do predictive analytics. “Making sure your infrastructure is elastic and scalable is key to being able to exploit the amount of information that’s out there.”
“If you’re a leader of IT, you have to get your arms around various initiatives and projects related to connected products,” Zujewski said. Not only will IT need to get a handle on what the needs and expectations are across the organization, they also need to see the big picture for another reason. “These projects start to add up to significant workload for network administrators,” Zujewski said.
To get ahead of issues, Zujewski encouraged IT to champion a cross-functional IoT center of excellence. “Get the functions in a room—engineering, IT, service, sales, finance and product management—and say, ‘What’s our two-year connection product or IoT strategy?” This will help ensure everyone is on the same page regarding which data will be collected and why, what the bandwidth requirements are likely to be, how security will be handled and all the other components that are part of supporting IoT initiatives.
Working with their OT counterparts will enable IT to know what sorts of problems need to be solved. “What is the architecture being contemplated?” Sarig asked. Multiple IoT devices may be connected in a hub-and-spoke pattern to sending data back to a centralized location in the cloud, or the model might be peer to peer, or some sort of hybrid. “That has implications on the technologies that they need to put to work,” Sarig explained. “Some protocols are more suitable for one type of use cases, and others for the others.” Having a plan to reduce latency and ensure quality of service (QoS) will be crucial, as not all protocols have QoS capabilities that are equivalent. “Unless IT gets together with their OT counterparts in understanding the requirements, they are not going to be in a good position to help influence the decisions,” Sarig said.