Manage Learn to apply best practices and optimize your operations.

IoT and the legacy system apocalypse

In today’s fast-paced world, people expect personalized information and services in the moment across almost every aspect of their lives — and especially when it comes to technology. Modern applications need to be fast and smart to stay ahead. Yet, despite this growing need for speed and personalization, many companies are clinging to legacy data infrastructures built on traditional relational database management systems (RDBMSes) that do not provide the necessary scale, data management at the edge, and virtual/cloud deployment capabilities required to keep pace with digital transformation strategies.

The internet of things is one of the biggest market drivers for retiring these legacy systems. Most legacy data infrastructures simply cannot meet the stringent requirements to power IoT applications and respond to the explosive growth in connected devices. In Gartner’s most recent IoT report, it predicts that the number of connected devices in use will hit 14.2 billion in 2019 and grow to 25 billion by 2021. Organizations are dealing with millions, if not billions, of connected assets that are streaming vast amounts of data into the enterprise every second.

Simply capturing, processing and analyzing all this data poses a huge challenge on its own — but more importantly, organizations need to then make that information useful by applying the insights to their business. The most powerful IoT applications need to be able to ingest the data and then make decisions on that data in real time. Smart meters that monitor usage and conditions, such as weather, for example, must be able to adjust the price of services according to usage and conditions that can change in an instant. Another example would be if a consumer forgets to turn off his oven at home and that drives up his utility costs unnecessarily; the subscriber must be informed in real time to turn off the oven so he can save that money. In order to get the most out of these meters and IoT applications, utilities need real-time streaming, monitoring and decision-making capabilities.

Looking specifically at industrial IoT, organizations are deploying sensor technology that provides automation and drives long-term cost efficiencies. However, many companies are sticking to their legacy data infrastructures built on traditional RDBMSes that simply don’t provide enough bandwidth to bring all the data back to the legacy systems, analyze it and send the results to the edge. In those cases, it’s often too late to make important decisions as the moment of opportunity has passed. Some organizations try to temporarily “fix” the problem by adding NoSQL caches to mask the challenges. However, these legacy systems simply don’t address the additional data volume and velocity issues every digital organization is now facing.

When it comes down to it, legacy systems do not have enough bandwidth to push all the data from every single sensor back to the central database, especially in a timely manner. This is not just cost prohibitive, but also legacy RDBMSes cannot grow to handle such astronomical scale and simply are not fast enough for real-time decision-making when an event like a natural disaster occurs. For these reasons and more, data must be aggregated at the fog layer — closer to the edge in a distributed, in-memory database.

It is clear that legacy systems have no place in the growing IoT ecosystem. Complementing them with real-time alternatives allows organizations to use the most current data when making critical decisions around operations and the customer experience. Manufacturers and other organizations must be able to stream data from their connected devices and equipment to help improve operational efficiencies and productivity, and reduce costs, while also gaining important insights to help shape business decisions and assess risks. As the amount of data generated by IoT devices continues to climb, having an agile, end-to-end IoT architecture that can both support real-time streaming analytics and scale to handle large volumes of data will only become more vital. In the end, all of the insights being driven by IoT data mean nothing unless organizations can apply them to their businesses before opportunities are lost.

All IoT Agenda network contributors are responsible for the content and accuracy of their posts. Opinions are of the writers and do not necessarily convey the thoughts of IoT Agenda.

Data Center
Data Management