Advanced analytics applications is a separate subject matter area; it is distinct from software development and from Data Science, though advanced analytics application development might require Data Science work and results in software modules. Mitek Analytics has strong experience in analytics applications and analytics architecture for the Industrial IoT. We use this expertise to help our customers who would like to start using advanced analytics.

Analytics Architecture

The architecture of enterprise software system can be described using established architecture frameworks such as TOGAF, which is widely adopted in software industry and promoted by The Open Group Architecture Forum, or DoD Architecture Framework (DODAF), which is used for large scale systems that include both software and hardware components. These architecture frameworks define several architecture layers (views). Systems design flows down through these layers. The main architecture layers are

  • Business Architecture, which describes business processes related to the software
  • Applications Architecture, which describes applications and analytics engines
  • Data Architecture, which describes databases and data management
  • Technology Architecture, which describes both operating software platform and hardware platform (e.g., the IIoT sensors)

So far, much of work in the IIoT area is in the two lower architecture levels that provide the IIoT platform. From the end user viewpoint this is necessary investment, expense. The added value of the IIoT is provided by the two upper architectural levels; Mitek's Analytics Engines enable the value added applications. The IIoT applications are mission critical and require high confidence analytics. These applications require both Predictive Analytics and Prescriptive Analytics. The IIoT Predictive Analytics use the historical data to build predictive models that can be used in simulations for verification and validation (V&V) of the analytics. The Prescriptive Analytics rely on the predictive models for data exploitation. They provide value added functions such as fleet monitoring, reliability analysis, risk analytics, predictive trending, anomaly detection, fault isolation, decision support, optimization, feedback control, fault tolerance and redundancy management, and many others.

Examples

As one example, Mitek Analytics led a 4-year NASA project in establishing analytics architecture for Integrated Vehicle Health Management (IVHM). In IVHM, the vehicle (such as aircraft, launch vehicle, or space vehicle) data are used to evaluate system health and provide optimized response to the detected anomalies. The team was led Mitek Analytics and included Boeing, GE, and Honeywell; the DODAF was used to formulate the high-level application architecture. The IT architecture was prototyped using open-source tools and followed service-oriented architecture design.

Other examples include analytics architecture in an integrated regional grid project for a US electrical utility and analytics application architecture for a high-tech startup. We developed analytics architectures for predictive monitoring of industrial turbomachinery fleet, for advanced large-scale distributed control in process industry, and for space systems.