The Latency/Throughput Tradeoff: Why Fast Services Are Slow And Vice Versa

The Latency/Throughput Tradeoff: Why Fast Services Are Slow And Vice Versa

Go ahead. Tell me DevOps and SRE aren’t the same thing. I dare you.

Ask a software engineer their thoughts on the optimal solution to any given problem and you'll usually get some version of "it depends".

This is because engineering is all about tradeoffs - what separates good from great is knowing what tradeoffs to make in what context.

This article talks about the tradeoffs when engineering a system for low latency vs high throughput and that a system made for one will be, by design, worse off for the other.

There are some parallels here to tools of thought as well. A tool that is made for live collaboration (eg. gdoc) might not be a good durable store of knowledge (eg. version-control). A tool that is optimized for tabular data (eg. excel) might not lend itself well to free-form writing (eg. plaintext notes).

What tradeoffs are acceptable depends on your personal context. What kind of ends are you optimizing for in your tools and what tradeoffs are you making?


Backlinks