olly - Fotolia

Get started Bring yourself up to speed with our introductory content.

4 strategies for scaling IoT to overcome technical challenges

IoT adoption alone can overwhelm an organization, but IoT leaders must also plan ahead to scale IoT deployments and avoid operational and technical pitfalls.

Scaling IoT projects challenges organizations' approach to such setups and existing architecture. It requires much more than additional sensors attached to more machines. IoT leaders must ensure their team and architecture can handle the increased connected devices and influx of data.

"Too many organizations treat [IoT] as a technology project, as opposed to what it really is, which is a business transformation project using technology. First rule of IoT club: Don't talk about IoT. Talk about business using IoT," said Alfonso Velosa, research vice president and analyst for IoT at Gartner.

Often, organizations will find it easier to build in tools to expand IoT from the start, instead of building on existing architecture.

Strategies to succeed at scaling IoT for operations

IoT leaders must work with their team to drive new technology adoption and expansion.

"If I'm doing business transformation using technology, I really have to keep my people first and foremost in front of me and figure out how they help drive the transformation, because they know what the problem is. Then we work together to use the technology," Velosa said.

IoT experts should follow these practices to set projects up for success.

Start small and build up

Organizations scaling IoT shouldn't aim to do it all in one attempt. A trial run on a set of sensors on a small number of machines can confirm that everything works before IT adds more sensors to all equipment, whether that is a fleet of trucks or a set of hotel elevators. Each expansion project should have very clear metrics of what defines their success and should pay for themselves to demonstrate sustainable capabilities. This also gives tech experts the opportunity to correct anything that doesn't work in a manageable way.

IoT data lifecycle

Set a clear business objective

Organizations must set a business objective and then align the technology to achieve that goal. The intent of an IoT project might be to optimize machine maintenance. The maintenance team would need software or a way to identify what type of device needs maintenance, which specific machine it is and where it is located. The reverse order of organizations collecting the data and hoping to find insights to create an optimization process from the data will not guarantee a successfully scalable IoT deployment.

Create a specialized team or center of excellence

IoT leaders should organize a group to allocate IoT project priorities and resources and share best practices, Velosa said. IoT deployments require executives or senior leaders to champion the projects to get everyone on board with the same goals. For example, IoT projects involve two departments: IT and operational technology (OT).

IT experts focus on standards, procedures and protocols, whereas an operational engineer prioritizes keeping things working reliably, potentially for decades. IoT leaders must address this IT/OT convergence of these teams and priorities.

Understand maturity

Maturity, in this situation, means how an IoT implementation ranks when it comes to meeting its objectives. Within an organization, different facilities or departments may be at different points in adopting new technology. Some facilities might have workers who decide to use IoT data and have the equipment for it already, but another facility might have analog equipment and employees who are unfamiliar with IoT. Where IT admins start an IoT implementation and scale up depends on the maturity of the facility or department.

Measuring IoT maturity

Technical challenges of scaling IoT

Engineers and IT admins have three areas that they must focus on when they design the technical aspects to scale their IoT deployment: data ingestion, data storage and organization, and data processing, according to Karthik Ranganathan, founder and CTO at Yugabyte. Each of these areas calls for specific technology and creates additional challenges, including security, latency and cost.

"It's important to think through and make sure that at any point, you're not so far away from the ability to scale that you actually need downtime," Ranganathan said.

Asset and connectivity management

Organizations must connect more machines to the network to expand their IoT deployments. This expansion requires IT admins to inventory, discover and add devices to their network and provision them to send data to the right place. Legacy machines might have different protocols that make automatically pulling data difficult. The IoT deployment might need a device, such as an IoT gateway, that can read the protocols and write the data before it is sent on to the cloud and analytics software.

Too many organizations treat [IoT] as a technology project, as opposed to what it really is, which is a business transformation project using technology.
Alfonso VelosaResearch vice president and analyst for IoT, Gartner

The further away data collection devices are from the cloud, the greater the latency. To avoid increased latency, organizations must incorporate edge computing architecture based at the data source. If latency is not a requirement, the IoT sensors could send the data back to the cloud for computing. With each new edge device connected, hackers have another entry point to the network. IT admins must have the ability to update device software over the air or in person and implement edge security best practices.

Data storage and organization

IoT sensors generate a massive amount of data, which requires storage space or else an organization might lose data. It's critical to have a distributed database to scale, which gives the advantage of more easily adding new devices without interruption, continued functionality when a failure occurs and lower communication costs for databases close to the data, Ranganathan said.

Some organizations might keep all the raw data, while others only keep outliers. This presents a tradeoff for better insights from more data at greater expense from more data storage. With distributed databases, more internal communication occurs between database nodes at the edge and the cloud. Scaling IoT introduces more data transfers over the network that IT admins must secure.

Data processing

Organizations might choose to only collect the data they can derive useful insights from, because not all the data is relevant. IT admins must use an event-based architecture for the IoT deployment to derive insights. IoT platforms can make sense of the massive amount of data IoT devices generate and often include other capabilities, such as management or security.

Dig Deeper on Internet of Things (IoT) Strategy

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.

SearchCIO

SearchSecurity

SearchNetworking

SearchDataCenter

SearchDataManagement

Close