Network Automation and Orchestration for UBS Plc.

Development Case Study: UBS Plc.

Summary

Development of network orchestration and automation solutions

Description

Like many large multinational banks, UBS has a highly complex network, with a high cost of change and management. The New Generation Networks (NGN) programme seeks to progressively introduce new network technologies which will employ extensive automation and orchestration to improve processes and reduce timelines and costs.

I worked with the NGN group’s own development team, primarily in a design and development role. Earlier work had developed a number of fragmented point solutions, and I created a more cohesive solution architecture. This exploits established design patterns to unify request handling and standardise internal data structures, with adapters to handle the multiple disparate endpoint interactions.

I became product owner and lead developer of this architecture’s first use for the DNS solution. This supported a 50% reduction in the DNS team, and will allow a significant element of self-service in what have previously been complex multi-party processes.

Problems and Challenges

Inexperienced development team with immature agile development practices split across multiple locations.
Complex technologies and procedures
Continual high deadline pressure

Outcomes

Successful delivery of a unifying framework for a suite of network management solutions, plus its first application in the DNS domain

Timescale

May – November 2017

Current Status

Completed commercial development, now in use

Tools and Technologies

Technologies: Python/Django, REST APIs, JSON, XML, YAML, Cisco NSO (Tail-F), Linux, CouchDB

Tools: Idea IntelliJ, Jira/Confluence

Abstract

UBS’ New Generation Networks programme aims to simplify management of their highly complex internal network using extensive automation and orchestration.
Early work consisted of fragmented point solutions. I created a cohesive solution architecture, with unified request handling, standardised data structures, and adapters to multiple endpoint interactions. I then developed this architecture’s first use for DNS management, using Python/Django, REST APIs and Cisco NSO (Tail-F), among others.