Tags: python sql geospatial
Developing a user friendly mapping service for telecom infrastructure.
The goal is to create a service analogous to Google Maps for telecommunication infrastructure, optimized for the best use of existing and new telecoms equipment. The service should be capable of balancing the cost of using current equipment against the necessity of new installations. We aim to provide two solutions: a highly customizable client-facing platform for connecting one to a couple hundred premises and data centers with varied options, and a 'bulk' solution for connecting all UK premises within reasonable resource usage and time constraints. The 'bulk' solution would require a significant speed up in calculations and the ability to process terabytes of data.
Our approach tackles this as a graph problem, wherein the task is to find the shortest path in a network. We incorporate the necessary data, business, and costing logic into the graph and the routing algorithms. Comprehensive understanding of the business logic, the data, and all components of the fiber optic networks connecting premises and data centers are essential to cater to different settings and products. For the 'bulk' solution, we developed new custom algorithms, shifted some ad-hoc processing into batch preprocessing and orchestration, and optimized queries, despite data quality issues, complex business logic, rules and exceptions, and computational resource limitations due to the large amount of data to be processed and represented in a graph.
Graph algorithms, graph analysis, linear optimisation, custom algorithms, geo-spatial data, SQL optimisation, OS Mastermaps.