The spending mix for IoT-centric analytics and information management (AIM) will increasingly tilt toward in-motion, flexible, and low latency systems to predict conditions that will require increasingly fast responses. The adoption of newer AIM technologies will be required, along with the requisite training to acquire needed skills.
The spending mix for IoT-centric analytics and information management (AIM) will increasingly tilt toward in-motion, flexible, and low latency systems to predict conditions that will require increasingly fast responses. The adoption of newer AIM technologies will be required, along with the requisite training to acquire the needed skills.
Building a middle-tier infrastructure for IoT is the opposite of traditional AIM technology adoption, which involves moving data in batches, then normalizing and loading the data into target systems. Analytical software is used once the data is loaded and at rest, typically to produce reports about what happened or statistical analysis that help in decision making or for on-demand decision automation.
With IoT, the goal of the AIM infrastructure is to provide a continuous infrastructure to correlate and analyze incoming data to predict whether further actions are required. Operational intelligence, machine learning and statistical analysis will all be part of the IoT AIM infrastructure.
IDC’s new TechScape covering IoT AIM looks at 25 technologies that will be adopted or extended to modernize AIM infrastructure. IoT AIM consists of generalized middle tier technologies useful for IoT as well as IoT-specific technologies. These are organized into the following categories:
- Edge data collection
- Data transport
- Data event services
- Data services
- Value-added data services
- Conditions and actions
Over time, AIM technology adopted for IoT will be different from an organization’s existing technology investments that performs a similar, but less time sensitive or data volume intensive function. Enterprises will want to leverage as much of their existing AIM investments as possible, especially initially, but will want to adopt IoT-aligned technology as they operationalize and identify functionality gaps in how data is moved and managed, how analytics are applied, and how actions are defined and triggered at the moment of insight.
Four considerations should dominate IoT AIM technology adoption planning:
- What is the total time window available to deliver business value when a condition is identified that requires a response? Adoption of AIM technology is required when time windows are narrower than the cycle time of the end-to-end IoT system, which begins at the emission of a sensor signal.
- How good is the prediction or insight from your analytics software? Quality problems occur for a variety of reasons, but noisy predictions and wrong or non-actionable predictions are all expensive. Using the best approach to analytics for a particular problem requires an assessment of whether there are data gaps that need to be resolved as well as identifying options and experimenting with them prior to adoption. Different techniques may also be required for different workloads or stages within a workload.
- How much technical debt are you accumulating by re-purposing existing AIM technology or investing in custom development? In the beginning, it makes sense to keep costs low by leveraging existing AIM technology for an IoT project. But technical debt rapidly accumulates when existing technology doesn’t really align with needs and has to be customized or contorted on an ongoing basis to make it work. As IoT initiatives are operationalized, the use of purpose-built tools is almost always a better path once those tools reach the required level of sophistication.
- How do technology choices align with your enterprise’s adoption risk profiles? Different organizations have different approaches to risk. When it becomes clear that there is a need to add new functionality or replace non-performing existing technology, the selection has to align with the skills of the team implementing and using the technology. We assess the adoption risk and speed of adoption for each of the 25 technologies highlighted in this TechScape. Planning should take both of those factors into account. However, if a technology identified in this TechScape has a higher risk than is acceptable to your organization but has a fast rate of market adoption, it is important to begin planning and acquiring skills sooner than later for eventual adoption,