We all have looked at the IoT value matrix, device and their sensor capabilities which enhances productive yield where organizations and individuals are gearing up to tap in the potential to enhance and transform their business model.
However, during the discovery process to transform your business using IoT, the aspects which should be considered are, your devices should have right sensors, they should seamlessly integrate into your current ecosystem, and they should focus on making/ deriving decisions for you faster.
The IoT use case should focus on.
1. Business transformation
2. Operational Efficiency and Cost-benefit analysis
3. GxP guidelines
4. Future proofing the solution by Adaptive or progressive improvisation via analytics
To realize these strategic ideations of the use case, have them timeboxed to develop an MVP, where you evaluate the feasibility, integrity and the return on investment to consider it. These act as input variables towards evolution of your architecture by improvisation which melds into your current enterprise architecture. The IoT paradigm reflects the building blocks of an enterprise IoT solution.
Once your foundation is laid, take a stepped approach of IoT and Analytics path. This starts with devices and sensors getting connected to your cloud ecosystems, and gradually building data paths to drive monitoring for devices explore business insights.
Primary Path (Hot Path) should co-relate to descriptive analytics telling you the current state of devices and raw information needed, it is crucial to have a minimal time-window delay to process your telemetry information which is relayed to other notification or workflow management systems.
Secondary Path (Warm Path) is responsible for deterministic measures aka diagnostic analytics where a snapshot of information is required to analyze what caused this event.
Ternary Path (Cold Path) should be spilt across supervised, unsupervised learning and reinforcement learning methods, ideally cleansed via streaming units or post-processing jobs in the solution.
Choosing the right mathematical model and adapting them to your use cases can be reflected via just 2 basic question.
- Do I need to classify it to an either-or case?
- Do I need to tag this result set to a bucket?
That’s easier said than done, but as business users with a strategic view initial sketched boundary, would have more fruitful results and can be extended gradually.
The primary purpose of IoT solutions is business insights. The nature of information and its relevance to the user is of utmost importance.
Your dashboard should be segmented into
- Monitor and Control of your devices
- Real time Telemetry Information
- Notification and actionable workflow insights
- Historical Time series information
- Preventive Maintenance depicting device life-span/ burn reports.
- Tailored Prediction metrics as to your business need.
These all dashboard, should be enabled with role-based access (RBAC) polices, with a granular magnification window over the reports/dashboards spanning from a Worm’s eye view to an Bird’s eye view, segregating them according to the relevant business units and its designated users avoiding clutter, providing with right information to the right user.
For a promised IoT solution, it should never affect the business continuity and needs to be ingrained into your core business by realized quantifiable business benefits.