16 milliseconds delay costs a colorist $120. We obsess over performance metrics like latency and bandwidth, how fast data moves, when latency determines how fast people work.
There are two types of latency in creative production. (1) Infrastructure latency is how quickly shared storage serves data to the application. (2) Operational latency measures how long your team waits to collaborate. Most technologists optimise for the first and ignore the second.
Infrastructure latency matters in a grading suite. If a colourist scrubs through 8K footage, 5 milliseconds feels instant, 50 milliseconds feels broken. This is where local storage and NVMe powered storage earns its premium.
But here’s the rub: that grading suite is 3% of your workflow. The remaining 97% involves editors in London waiting for dailies from Vancouver, artists in Mumbai waiting for approval from LA, producers waiting for renders that finished six hours ago in Sydney.
Operational latency isn’t measured. When it is, it’s in hours. A global team loses 16 hours of productivity each day to handoffs. Assets are duplicated because nobody knows they already exist. This latency costs more than slow storage ever will.
Infrastructure vendors sell a Ferrari driving full speed down the motorway. If your journey is through towns and suburbs, that Fiat 500 is the right choice. A 100 Gbps network means nothing if teams wait hours for the right data.
The media industry spent $13 billion on storage infrastructure last year [link 3]. Did we eliminate the latency between human decisions? Between project start and asset delivery? Between “we need this” and “here it is”?
Optimising infrastructure latency for 3% of the workflow is a marginal gain. Optimize operational latency wins the game. Which one determines whether your make client delivery on time?