Over the past 40 years, technological change has proven highly disruptive to virtually all global enterprises. From computers and networks to e-commerce and mobile devices, companies have been forced to quickly adapt or experience the pain that comes with being a follower. The Internet of Things (IoT) is likely to be the next monumental shift. The IoT is quickly becoming mainstream, and together with artificial intelligence, robotics and other emerging technologies, could end up having a bigger impact than any of the technologies that preceded it.
The problem, though, is that there are a lot of unknowns. Hardware will continue to improve, standards and protocols will continue to proliferate, and advancements will be made in where, when and how data is collected, analyzed and acted upon. For companies who are planning an IoT strategy, this can be daunting. The traditional method of architecting for change, where people analyze the current state and then try to envision a future state, just doesn’t work. There are too many unknowns. A better solution is to focus on improving the current state of business and developing an architecture that can evolve instead of having to start over each time a change occurs.
Architecting For Change
As software has developed over the last few decades, it has become much more diverse. The move from client-server to networks and the cloud has created many opportunities for integration, but it’s also made architecting solutions more difficult. By creating hooks that allow software to more easily communicate, application programming interfaces (APIs) have been an enabler for growth, but they’ve also become a source of heartburn for enterprise architects. At times, corporate systems are evolving faster and more broadly than they can anticipate. Change is no longer the exception, it’s the new normal, and companies must be ready and able to adapt.
To account for this sort of rapid transformation, companies are considering more agile forms of architecture. Typically, these are based on several key principles:
- Ontology: The architects, implementers and other constituents need to utilize a common structured vocabulary that represents the meaning of terms within a specialized context. It acts as a specific form of glossary.
- Abstraction: Components that make up a system are decoupled to the point where they don’t need to know much, if anything, about one another. That way, when a change is made to one, there’s minimal impact elsewhere.
- Standards-based: Modern architectures rely on industry, de facto, and corporate standards as a way to ensure everyone is building things the same way. Among other things, it allows new people and processes to plug in without having to learn a customized work system.
- Agnosticity: The goal here is to avoid being locked in by a specific technology, product or vendor. An architecture might be reliant on a specific technology (for now) but can change as other options become available.
Topics To Consider When Planning For IoT
In a recent Forbes Insights study of more than 500 executives, 40% of respondents said they already have significant IoT programs in place. To ensure ongoing success in this highly dynamic environment, companies need to build flexibility and modularity into their architecture. Within an IoT system, there are four main areas that must be considered:
- Hardware and software: With so much variation in devices and so many different protocols for connecting and moving data, the architecture must be highly modular so that, if needed, specific hardware components and software programs can be replaced without having to reengineer the entire solution.
- Network connectivity: While the devices in an IoT system might be diverse, they still must send and receive data while also protecting security. Some of the principles of abstraction can help here, where the location, language and naming convention of each component are irrelevant. The system must be designed for transparency.
- Edge computing: The whole idea behind edge computing is to have some decisions made as close to the point of data creation and/or collection as possible. This helps improve speed and reduce network bandwidth. But edge computing also diverges from a traditional cloud-based system. An agile IoT architecture must be able to account for new advancements, not only in hardware and software, but also with networking and other innovations.
- Data management: Data that requires more complex processing and doesn’t require immediate decision making is often delivered to a data center or the cloud for analysis, insight and storage. The architecture for these systems must be flexible enough to support multiple environments. For example, an IoT device might need to connect and communicate with legacy equipment and share that information with a cloud-based enterprise resource planning (ERP) system. IoT architecture must be adaptable and agile enough to process and manage data from many disparate sources.
IoT architectures must also consider elements beyond the technology itself, including data analysis and insight, both of which are driven by people. In the past, companies would create a specific role with responsibility for ensuring that people and workflows were reviewed and prepared to support new capabilities. As part of their work, these individuals would analyze a current process, make recommendations, test their theories and, if successful, reduce the new process to practice. In the highly fluid world of IoT, this approach is too rigid and too slow. Companies have to make benefiting from change a core competency, and that begins by planning for change in their architecture, enabling rapid and thorough adaptation.
Source by : Forbes
Author by : Insights Team , Forbes Insights