As you approach your first Internet of Things project, here are some important ideas and practices to keep in mind.
- Understand the big picture.
Start with a question, specifically: What specific problem are you trying to solve? IoT often addresses failure points.Grasp the specific use case for your project, which could range from a self-driving car to locations of equipment to preventive maintenance to irrigating a vineyard. The more specific the better.Derived insight response time is a critical IoT concept, referring to the data or guidance needed to make a decision. The shorter the response time, the closer the computational analysis to the sensor should be.
- Pick a strong technical lead.
IoT projects are quite technical,Large technology projects are complex and often require that multiple technologies and components are designed and integrated into a tailored solution. Such projects contain a lot of uncertainty and rarely go according to plan.” The project lead needs a strong technical lead to manage or clarify this complexity.
- Phase the project.
Break the project into manageable phases or start with a proof of concept stage where you can try, at relatively low risk, to prove that your IoT project plan will work.
- Test often.
Do this especially for interfaces between different elements in IoT such as sensors and communication devices that collect the raw data. Also, try software cleansing the raw data to ready it for analytics.
- Understand (or select) an IoT reference architecture.
With your technical lead, address which specific technologies you will use to address the seven-layer architecture emerging as a best practice. The most comprehensive version, funded by the European Union, is from the Internet of Things Architecture (IoT-A) research project. IT consulting firm Sirius Computer Solutions has a more concise summary (complete with Cisco graphic) in the white paper Four Key Strategies to Kickstart Your IoT Revolution.
- Pick your project team.
With all the technologies required in IoT, you’ll need diverse talents: Hardware and software engineers, embedded device experts, wireless whizzes, storage and integration pros, database and server clustering specialists, etc. Have your technical lead help you identify the skills and number of hours needed. IoT experts are still rare, so don’t be afraid to add outside consultants to fill in gaps and transfer skills, even at a steep price.
- Pick your project infrastructure.
There are a lot of decisions to be made—on sensors, storage, servers, integration, data analytics, communications and multiple types of software, including a database. It’s common practice to opt for an open source database.
- Maximize sharing.
Communicate with team members regularly and urge them to talk to each other about what’s going on, their concerns, findings, etc. Using online project management software can be a huge facilitator for interacting and tracking progress—especially when team members are located remotely or in different cities.
- Build team competencies.
Your organization’s first IoT project won’t be its last. Your project team will learn a lot on your debut IoT project. And that knowledge that proves va luable to your team and the company on future IoT projects.
- Learn about fog computing.
Fog computing puts processing power into smart devices that are placed close to the sensors, which generate the data. “Fogging” improves efficiency and reduces the amount of data that must be sent over the network or into the cloud. Consider cloud infrastructure as both highly available and elastic. (In meteorology, fog is simply a cloud that is close to the ground—hence this metaphor.)
- Get ready to deal with your team’s frustration.
It’s tough doing cutting-edge work, but you can inspire your team to move from frustration to exhilaration. You are all part of a strategic initiative, so make that a motivator.