I was directed here from the #spin channel in the CNCF Slack:
https://cloud-native.slack.com/archives/C089NJ9G1V0/p1778007085277469
and opened an issue to propose adding FIPS support to Wasmtime:
https://github.com/bytecodealliance/wasmtime/issues/13293
I'd be interested to hear what you think.
Thanks for filing that. I don't know much about FIPS other than that it's a quite heavyweight certification process. Am I right in reading the issue that Wasmtime itself wouldn't have to undergo any certification or whatnot, simply "using the right backend" is enough?
If it's as "simple" as configuring rustls+hyper correctly, and perhaps if you can show us a diff of what it would take, this seems plausible to me but @Pat Hickey would almost certainly have more detailed thoughts...
Yeah, that's correct. As long as you're correctly using the module, you should be good.
In terms of showing you a diff, unfortunately I don't know rust well enough to propose the necessary PR. I'm hoping it's as simple as following these rustls docs:
https://docs.rs/rustls/latest/rustls/manual/_06_fips/index.html
(I'll add a link to those in the issue)
This is the sort of feature where we'd probably say "PRs welcome": we don't have the right combination of free time and highest-priority motivation to allocate our time, but we would welcome contributions
Fair enough.
Getting up to speed enough to be able to contribute this myself could be a fun advent-of-code alternative at some point during the year.
it would. The difficulty of the certification process is relative to the value it would provide companies using it. How that weighs on you is entirely up to you. it would be a great thing, yes. Most of us do not have FIPS needs.... yet.
I don't think any certification of Wasmtime itself would be needed, since it would just be inheriting the NIST certification of the existing crypto module if it's done right.
For companies putting a product into FedRAMP, they'd still need to do the normal audit with ATO etc for their service as a whole, but this would remove a blocker for Wasmtime being used as a component of the implementation, when trying to satisfy the "data in transit" part of the SC-12 family of controls.
Last updated: May 26 2026 at 09:09 UTC