You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@tuweni.apache.org by "Antoine Toulme (Jira)" <ji...@apache.org> on 2019/12/16 08:42:00 UTC
[jira] [Updated] (TUWENI-10) Move from using a LRU map to a bloom
filter
[ https://issues.apache.org/jira/browse/TUWENI-10?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Antoine Toulme updated TUWENI-10:
---------------------------------
Fix Version/s: (was: 0.9)
0.10.0
> Move from using a LRU map to a bloom filter
> -------------------------------------------
>
> Key: TUWENI-10
> URL: https://issues.apache.org/jira/browse/TUWENI-10
> Project: Tuweni
> Issue Type: Improvement
> Reporter: Antoine Toulme
> Priority: Major
> Fix For: 0.10.0
>
>
> The current plumtree implementation keeps track of the state of the last 1000000 messages it received.
> This is rather memory intensive - and can be gamed by adversarial behaviors by recycling and running replay attacks with a bigger batch of messages.
> We can use bloom filters to identify messages instead.
> The work consists in:
> -Find a bloom filter in java or implement one. (there is a bloom filter in the eth component that can be generalized under its own library, maybe?)
> -Use it to identify messages as an option to the LRU map
> -Test the tests still pass
> -Allow that bloom filters need to still be rotated over time.
>
> Would be also good to move the existing LRU map to a mechanism external to plumtree state, so it can be backed by a DB or a cache specified by the user.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@tuweni.apache.org
For additional commands, e-mail: issues-help@tuweni.apache.org