Presentation is loading. Please wait.

Presentation is loading. Please wait.

DASA Logistics Analytics Hackathon An Architectural Approach

Similar presentations


Presentation on theme: "DASA Logistics Analytics Hackathon An Architectural Approach"— Presentation transcript:

1 DASA Logistics Analytics Hackathon An Architectural Approach
Andrew Gordon Principal Engineer Future Technology – Air September 2019

2 The MoD DASA Logistics Analytics challenge…
“Demonstrate the ability to analyse and share structured and unstructured multi-source data; maintaining its classification and permission based access rules at machine speed” The UK MoD, and partner coalition Nation Forces, want to increase the level of data sharing between themselves and industry to support better interoperability and operational outcomes However, once data is extracted from source systems and applications, access control rules are usually lost. Also data access control rules often vary on the same data depending on context such as location and time The target use case for the Hackathon was logistics data, specifically platform, maintenance and inventory information for the UK, US and French C130J fleets. The goal was the demonstrate how this data, and Analytics (AI / ML) algorithm output could be shared between stakeholders, whilst ensuring security and access control rules

3 What are the key characteristics of the solution we came up with?
Architectural, specifically separating the architectural concerns of Access Control, Data / Analytics Service and Data Consumer Data is not moved (if at all possible) Data Service implementation is technology vendor agnostic Promotes the use of Open Standards and interoperability Approach can be extended to the Platform itself, e.g. on-board Avionics Supports deployment in network disconnected environments Distributed and highly scalable Implementation approach promotes collaboration between MoD, Prime Systems Integrators, Technology Vendors and SMEs

4 What does the outline Architecture look like?
Data Sources FR E&AM US E&AM LITS MJDI DTADS Data Services E&AM Data Service E&AM Data Service E&AM Data Service Inventory Data Service Platform Data Service On-Board Health Data Service Analytics Services Security Service Audit Service E&AM Analytics Service Natural Language Processing Data Cleanse Service Access & Data Redaction Service Rules Database Blockchain Presentation Services

5 Our approach is in alignment with Gov
Our approach is in alignment with Gov.UK Digital and MoD ISS Defence-as-a-Platform (DaaP) strategy and policy

6 The Architecture in more detail – separation of concerns is key to modularity and scalability
Data Source { "TAIL NO.":"Z123", "SORTIE TYPE":"SEARCH & RESCUE", "SORTIE LOCATION":"SOME RESTRICTED LOCATION" } Data Service Analytics Algorithm Access & Data Redaction Service Data Redaction Rule Engine Prescriptive Rules ML/AI Derived Rules Analytics Service { "TAIL NO.":"Z123", "SORTIE TYPE":"SEARCH & RESCUE", "SORTIE LOCATION":"[SORTIE LOCATION]" }

7 What did the Prototype we built at the Hackathon in ~48hrs look like?
We built a number of analytical algorithms in MATLAB, R and Python that ran against the C130J fleet data held in the Oracle Database Cloud enviroment Wrapped both the Oracle Database and Analytical Algorithms in RESTful Microservices using Python Flask Built an Access Control Permission Mask Service that returned business rule expressions to redact data dependent on calling service identity – again using Python Flask Built a demonstration Web Client in JavaScript that invoked the Analytics Services

8 E&AM Data Service

9 E&AM Data Service

10 E&AM Analytics Service

11 E&AM Analytics Service

12 Analytics Service Visualisation via Microsoft Excel

13 Data Redaction

14 Data Access via Microsoft PowerBI

15 Data & Analytics Service access via AngularJS Mobile App

16 In summary… We believe that successful adoption and exploitation of AI, Machine Learning and Advanced Analytics in Logistics is dependent on building an open, interoperable and secure data platform For the UK MoD to fully realise the benefits of capabilities such as predictive maintenance, prognostics, supply chain analytics etc., ‘frictionless’ data interoperability with industry and the extended Defence Enterprise is key The challenge for UK MoD and industry is to transform and modernise the current Defence Logistics data landscape to provide the platform for AI/ML exploitation

17 Thank you Restrictions on use:
The information in this document is copyright © 2019 BAE SYSTEMS. All rights reserved


Download ppt "DASA Logistics Analytics Hackathon An Architectural Approach"

Similar presentations


Ads by Google