2020-08-28
2807
#rust
Andre Bogus
23972
Aug 28, 2020 â‹… 10 min read

Rust serialization: What’s ready for production today?

Andre Bogus Andre "llogiq" Bogus is a Rust contributor and Clippy maintainer. A musician-turned-programmer, he has worked in many fields, from voice acting and teaching, to programming and managing software projects. He enjoys learning new things and telling others about them.

Recent posts:

how to use the Interest Invoker API for better, more accessible UX

How to use the Interest Invoker API for better, more accessible UX

Explore how to use Google’s new experimental Interest Invoker API for delays, popovers, and smarter hover UX.

Emmanuel John
Jun 19, 2025 â‹… 7 min read
How To Build And Deploy A Web App With Bolt.new

How to build and deploy a web app with Bolt

Bolt.new revolutionizes how you build and deploy web apps with no-code development and seamless AI integration.

Isaac Okoro
Jun 19, 2025 â‹… 8 min read
how to get the most out of Cursor

Frontend devs: Here’s how to get the most out of Cursor

Learn how to get the most out of Cursor AI — one of the hottest tools in AI-assisted coding, with practical workflows and underrated features.

Isaac Okoro
Jun 18, 2025 â‹… 5 min read
When to Use OpenAI vs. Open Source LLMs LogRocket

When to use OpenAI vs. open source LLMs in production

Learn about OpenAI vs open source LLMs for frontend devs, with an integration guide, costs, performance comparison, and implementation tips.

Clara Ekekenta
Jun 17, 2025 â‹… 6 min read
View all posts

4 Replies to "Rust serialization: What’s ready for production today?"

  1. Hi Ander,

    Thanks for the interesting article.
    I try to understand the background of the serialization/deserialization methods on the Serde. For example, how the BinCode serialization method encodes data and which metadata store after encoding.

    Best,
    Saeed

  2. This is great. I really appreciate this concise overview which is exactly what I needed to start shopping. One thing though that would help immensely is if, at least in the comparison table, the units were all standard or if there was an option to make them standard or to sort them. For today, I will calculate them in a spreadsheet (happy to share, obviously) but I’m sure the next person who wants to use this the same way will appreciate the lack of an extra step.

    Thanks for all the great work!

    1. Thanks for the nice words. The units *are* all standard SI units. But I think you may actually mean they should be the same unit (e.g. nanoseconds). The downside is that this would yield some pretty big numbers which aren’t that easy to parse either. I’ll try and see if a nanosecond-based comparison is any easier to read.

  3. MessagePack is a self describing protocol and therefore has to encode the field names (‘variant’,’opt_bool’,’vec_strs’,’range’). I really can’t see how this would fit into 24 bytes total, only the field names would take 28 bytes (so without any protocol overhead and values) if I’m counting correctly…

Leave a Reply