-
Notifications
You must be signed in to change notification settings - Fork 314
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix expander parents generation #827
Comments
Studying new analysis in Stacked-DRGs & ZigZag Security parameters, section 2.1, the conclusion for implementation fixes until now is:
@irenegia @porcuquine Could either of you confirm point 2 please? Reading |
@schomatis Your general reading is correct, but let me make a few notes and help prioritize.
There will probably be changes to how parents are handled in the future.
It's not clear to me that the PDF is updated to reflect this, btw. cc: @irenegia |
@schomatis Elsewhere (will try to comment there also), I think you mentioned wanting to defer this until after caching is moved to disk. I think it's important that we make the logic changes as soon as possible so that they don't block or complicate the many other changes which will be coming soon. If we really want to avoid caches growing in the short term, you could change the parameters to be 'half-strength' (i.e. lower But please prioritize this over restructuring the cache. It's very high priority that this be taken care of before we start in on new ZigZag work. [UPDATE: after catching up, it's not clear to me that my comments above address any plan you actually had — so ignore whatever is inapplicable. That this is a high priority change that should be taken care of as early as possible remains true, though.] |
@porcuquine I'm sadly missing most of what you wrote in the last two messages, sorry. Let me ask one question at a time to narrow down the conversation as much as possible. Regarding point 1, the ZigZag graph |
@schomatis The ZigZag graph Per #744 (which seems is still not yet implemented), the base graph's degree should be 1 + its My suggestion above is that we leave So, with current parameters the ZigZag graph's degree should be 5 + 1 + 2*8 = 22. This is the actual number of parents (including padding) that a node will have. |
Thanks for the clarification, will correct PR #860 in the coming days. |
@porcuquine Note that there is a (draft) implementation in place already linked to that issue, see #843, @dignifiedquire had some thoughts about it, I'll need you two to work them out in #744. |
Will be fixed in #864. |
Description
See the work done in filecoin-project/research#144, the newest spec for the chung construction lives in https://www.overleaf.com/3456316943pfxsqxkdghsy%E2%80%A9
Acceptance criteria
Risks + pitfalls
Where to begin
The text was updated successfully, but these errors were encountered: