A Knowledge Graph Driven Approach for Edge Analytics
Additionally, in a situation where the modification or removal of an edge device is required, that change must be communicated throughout the system due to the inability of a vertically integrated solution to handle disparate heterogeneous deployments because there isn’ t a filtering, normalization and routing, of messages based on ecosystem topology. Even robust horizontally-integrated systems such as the Cisco ® IOx platform require some degree of hardware standardization( networking infrastructure). The knowledge graph driven approach requires solely software level standardization, namely the ability to run containers.
Client Use Case Description
Following the generalized use case are two on-going specific use case examples, in the Oil and Gas( O & G) and Mining sectors, from clients for whom we are currently developing our edge analytics framework.
Large O & G Services Company
Consider the use case of a large multinational O & G field services provider with a complex organizational and operational structure. This client has been digitizing and networking their industrial assets in on-shore and off-shore oil / gas fields for decades and has garnered a mature understanding of and experience with Industrial IoT technologies and tools. In the process of doing so the client has accumulated technological“ debt” over the years ranging from proprietary solutions, heterogeneous approaches and dated hardware involving several business units in pursuit of gathering data and running analytics on the edge. Thus, the client is dependent on and manages legacy technology with redundancies in hardware capabilities and software functionality across the stack. This leads to an opportunity to optimize on operational costs, increase efficiency by standardizing methodologies and adopt a modernized enterprise-wide Edge Analytics platform.
Challenges
The Brownfield nature of this client’ s operational environment means any edge framework must be flexible enough to support existing operations and at the same time provide a seamless path forward toward ecosystem modernization, technology migration and device deployment. The diversity of needs for this organization means it must support multiple edge hardware devices, operating systems( OS’ s), variety of data-processing and analytical requirements, data storage and data format needs.
In-flight Implementation Requirements
The O & G client stipulated the following key solution requirements:
1) Heterogeneous Software( SW) Support: support for applications capabilities ranging from Complex Event Processing( CEP) and rules engine for data processing, as well as employment of Artificial Intelligence( AI) open source library frameworks to reduce licensing costs.
2) Extensibility and Agnostic Support: expansion capability for vendor and other 3 rd party application integration, i. e., support ecosystem of players
- 46- September 2017