DNA 26-01: Trigger’s Broom

For 2026 I am setting myself a goal which I will almost certainly regret. On LinkedIn, I am going to post a meaningful technical update once a week and call it my Diary of a Network Architect (DNA). And no, I didn’t steal the title from a very popular and insightful podcast, that’s a podcast and this is a diary. What’s more, I will not be endorsing any products, as vendor independence is a core Starfish value. I am going to start simple, with an often-overlooked benefit of SDN, lifecycle management.

I have been working on SDN and Cisco ACI for 10 years. After my first major ACI implementation I wrote about the key benefits of Cisco ACI and SDN in general. I cited key features such as programmability, performance, scale, and integration with network functions virtualisation. Over the years all these things have proved very valuable. Today we make good use of APIs to automate configuration, while Clos and Fat Tree are the predominant data centre topology for low latency, high bandwidth, resilient networks. However, following successful SDN upgrades in 2025 by the Starfish team, it became obvious to me that there are significant lifecycle management benefits too.

Traditional networks are decentralised with configuration and policy spread across all switches, routers, and firewalls. This makes hardware upgrades complex and organisation will regularly choose to implement entirely new solutions as part of a multi-year refresh cycle. However, with SDN the configuration and policy are held centrally on a controller and only the hardware itself needs to change. Upgrades only need to be carried out to individual components such as controllers, leaf switches, or spine switches. This reminds me a lot of the classic comedy Only Fools and Horses and Trigger’s Broom, a broom which lasted 20 years but, in that time, had 17 new heads and 14 new handles.  

The lifecycle management benefits are achieved in several ways. Total cost of ownership is reduced, as upgrades are easier and less intrusive, and integration with peripheral infrastructure such as compute, storage and security, does not need to change. Investment protection is achieved as hardware components are only upgraded, when necessary, thus reducing hardware spend and training costs. Finally, the reduced hardware turnover helps with sustainability. As a rule, I advise customers to realise these lifecycle benefits by maintaining their SDN platforms over the long term, unless there is a genuinely compelling reason for change.  

Feel free to reach out to me with any comments, corrections, and questions.

https://www.linkedin.com/pulse/sdn-orchestration-why-fabrix-went-cisco-aci-stephen-hampton/

https://www.youtube.com/shorts/ij-wOtJqsHg