Read on Twitter
Have been obsessed with data+edge-computing recently. @Cloudflare Workers are giving me speed at a bargain, which unlocks much better UX
Did an experiment for my pollution tracking dashboard. Version with Workers on the left, optimized Django+PostgreSQL on the right
Did an experiment for my pollution tracking dashboard. Version with Workers on the left, optimized Django+PostgreSQL on the right
Notice the difference in speeds? It only gets worse with longer timeranges and more traffic. The SQL server often crashes when there are multiple users logging in with complex queries, unless I pay much more for an expensive machine. Workers are incredibly scalable [2/]
I had to rearchitect my entire workflow to use workers. A generic serverless architecture using a document store would’ve also resulted in better performance than what I’m currently using. But still doesn’t beat Workers ($5/mth for 10Mreads + more!) was hard to compete with [3/]
Speed makes a difference. In my experiments launching dashboards, speed of interaction is key to whether or not users explore them
Dashboards at the speed of thought inspire creative questions. Waiting for your data to load creates frustration [4/]
Dashboards at the speed of thought inspire creative questions. Waiting for your data to load creates frustration [4/]
This is also true in personalization. Am trying to create an architecture to power media personalization that updates recommendations quickly using workers. If it works, will be a super fast, cost effective solution. Have used slower, more expensive ones in the past [5/]
The promise of Cloudflare (and others, like Fastly) is the idea of the network as a computer. Opens up so many use cases
Workers are limited in what they offer, but are so much cheaper and faster than the competition. Textbook case of Clay Christensen’s Disruption theory!
Workers are limited in what they offer, but are so much cheaper and faster than the competition. Textbook case of Clay Christensen’s Disruption theory!
