With new uses and applications of IoT, there also comes new methods of delivering and managing connectivity for devices—as well as fresh ways to monetize the data they carry.

By Vikram K It’s boom times for the Internet of Things. “Things” are being added to the connected global ecosystem at an incredible rate—and not just the overall number of things, but the diversity of use cases is growing rapidly as well. Gartner predicts that the number of endpoints will reach 20.8 billion units by 2020. With new uses and applications of IoT, there also comes new methods of delivering and managing connectivity for devices—as well as fresh ways to monetize the data they carry. While the evolution of devices in the ability to collect data is captivating, the challenge for organizations will be in creating a flexible infrastructure through which they can connect and manage the changing world of things. When Things get complicated Every organization will have unique requirements for managing IoT data based on factors such as how fast, how secure, or how critical the data is. Even within a single organization, different M2M data will need different technologies and flexible architectures to be valuable. And in the vast majority of cases, the data will also need to be shared with other IT instances, like analytics platforms and other applications. The automotive industry provides a great case study for such variances. Today’s newer “connected cars” collect a wide range of operator and performance data. But how are car manufacturers using it? One way is to transform the customer experience, providing constant connect points between the auto maker and the consumer. For example, car makers might use some of this data to warn about impending weather, or alert drivers approaching a border to have their relevant documents ready or risk being fined or turned away. By monitoring how customers are using their cars, the auto maker can better identify more valuable features and functionalities—while the customer gets a better driving experience. The collection and analysis of data on how drivers use their cars can happen over a long period of time, but there are other uses of data in the connected car that require real-time analysis and action, like the risk of collision. This has a very different set of data collection and processing requirements. Auto manufacturers —and any other organization deploying an IoT solution—need a flexible IoT platform to help them meet the requirements of these very different use cases. They also need to address certain challenges, all of which can be classified within these categories:

1. Connectivity. The most efficient choice for transmitting M2M data will change based on the device and how the data is used. There are new options in connectivity, including cellular networks, LTE, and cutting-edge low-power, long-range networks.
2. Compute. Already many organizations have learned the hard way that as data volume scales from megabytes to terabytes to petabytes, their ability to ingest, validate, and load that data can’t keep up. Compute resources must scale easily and flexibly—and critical real-time data must be processed at the edge for faster time to action.
3. Security and governance. If your devices move around, they need business rules to govern their behavior based on changing location or conditions. And of course, sensitive data has its own set of requirements.
4. Analytics. Where and how you process data is becoming a critical pivot point in IoT architectures. Low-latency requirements often invoke data processing and edge analytics, for example. However, other uses require you to collect data from different locations over longer periods of time, while transmitting to a data center or the cloud for later processing and analysis.
5. Partner Ecosystem. There is no single company that can deliver the full breadth of an IoT solution. A successful IoT deployment requires many capabilities that can only be fulfilled by teaming with multiple partners for efficiency. Auto makers, for example, may use a third party to deliver service and performance insights to consumers.
6. Services. Business stakeholders and data analysts need a methodology for interacting with and managing deployed devices—as well as people and processes—so they can monitor for efficiency and performance.