All work
- Benchmark Parsec with malice detection turned onMAID-3305
- Integrate malice detection into the integration tests frameworkMAID-3304
- Unit test malice more thoroughlyMAID-3303
- Handle spamMAID-3302
- Fix failures in the consensus_with_fork test when malice-detection is enabledMAID-3301Resolved issue: MAID-3301Pierre Chevalier
- Make PARSEC work with single nodeMAID-3300Resolved issue: MAID-3300Fraser Hutchison
- Functionally test new Requesting/Request/Response pattern introduced in MAID-3282MAID-3299Resolved issue: MAID-3299Fraser Hutchison
- Don't create sync event when handling invalid messageMAID-3298Resolved issue: MAID-3298Qi Ma
- Decide how to handle invalid accusations patternMAID-3297
- Change detection of interesting events according to the new PARSEC paper (use is_ancestor instead of sees)MAID-3296Resolved issue: MAID-3296Jean-Philippe Dufraigne
- Tests with supermajority mode appear not to replay nicelyMAID-3295
- Simplify tie-breaking rule in PARSEC's implementationMAID-3294
- our_unpolled_observations sometimes incorrectly returns observations that have already been polledMAID-3293Resolved issue: MAID-3293Adam Cigánek
- Make graphs contains fork to be dumped into dot files properlyMAID-3292Resolved issue: MAID-3292Adam Cigánek
- Test the implementation change in the task 3287MAID-3291Resolved issue: MAID-3291Pierre Chevalier
- Test the implementation change in the task 3285MAID-3290Resolved issue: MAID-3290Adam Cigánek
- Implement mock ParsecMAID-3289Resolved issue: MAID-3289Viv Rajkumar
- Test the implementation change in the task 3286MAID-3288Resolved issue: MAID-3288Fraser Hutchison
- Fix a possible bug in the Accomplice accusation of casting accomplice even sees the original accusationMAID-3287Resolved issue: MAID-3287Jean-Philippe Dufraigne
- Fix a possible bug in the Accomplice accusationMAID-3286Resolved issue: MAID-3286Jean-Philippe Dufraigne
- Fix a possible bug in the Fork accusationMAID-3285Resolved issue: MAID-3285Jean-Philippe Dufraigne
- Fix the concrete coin bit sourceMAID-3284Resolved issue: MAID-3284Fraser Hutchison
- The way we detect forks in Parsec is currently incorrectMAID-3283Resolved issue: MAID-3283Adam Cigánek
- Fix possible bias of the concrete coin mechanism with fake Responses sent to fake Requests by malicious nodesMAID-3282Resolved issue: MAID-3282Fraser Hutchison
- Performance: Only retain the current meta-electionMAID-3281Resolved issue: MAID-3281Pierre Chevalier
- Fix duplicate Accomplice accusationsMAID-3280Resolved issue: MAID-3280Jon Haggblad
- Investigate potential benefits of replacing hash-to-index mappings with RcsMAID-3279Fraser Hutchison
- Upgrade Accomplice malice to provableMAID-3278Resolved issue: MAID-3278Jon Haggblad
- Use usize index to identify observations, for further speed improvement and memory reductionMAID-3277Adam Cigánek
- Remove redundant fields from Event to reduce memory usageMAID-3276Resolved issue: MAID-3276Fraser Hutchison
- Investigate ways to reduce map/set lookups by PublicIdMAID-3275Resolved issue: MAID-3275Fraser Hutchison
- Investigate performance degradationMAID-3274Resolved issue: MAID-3274Pierre Chevalier
- Prevent accomplice framing scenarioMAID-3273Resolved issue: MAID-3273Bartlomiej Kaminski
- Ensure sync events are assigned the correct 'other-parent'MAID-3272Resolved issue: MAID-3272Qi Ma
- Check correctness of consensus in newly joining peersMAID-3271Resolved issue: MAID-3271Fraser Hutchison
- Investigate performance of tests containing forks with malice-detection enabledMAID-3270Resolved issue: MAID-3270Qi Ma
- Update behaviour of 'has_unconsensused_observations()'MAID-3269
- Test Parsec::gossip_recipientMAID-3268Resolved issue: MAID-3268Fraser Hutchison
- In Parsec, let clients get a list of peers that are valid gossip targetsMAID-3267Resolved issue: MAID-3267Fraser Hutchison
- In Parsec, also return "excess events" from pollMAID-3266Resolved issue: MAID-3266Adam Cigánek
- Use enum in Parsec constructor to replace interesting_event functorMAID-3265Resolved issue: MAID-3265Jon Haggblad
- Random peer names cause dot files to render incorrectlyMAID-3264Resolved issue: MAID-3264Fraser Hutchison
- Run benchmark suite dailyMAID-3263Resolved issue: MAID-3263Pierre Chevalier
- Handle stalled nodes in Parsec testsMAID-3262
- Fix performance regressions after MAID-3248MAID-3261Resolved issue: MAID-3261Pierre Chevalier
- Fix the test for single-vote interesting eventsMAID-3260Resolved issue: MAID-3260Bartlomiej Kaminski
- Abbreviated hashes in Debug traits should be limited to dump-graphsMAID-3259Resolved issue: MAID-3259Adam Cigánek
- Optimize event lookupMAID-3258Resolved issue: MAID-3258Bartlomiej Kaminski
- Avoid calling Observation::create_hash() as frequentlyMAID-3257Resolved issue: MAID-3257Bartlomiej Kaminski
- Reduce the amount of event reprocessing after consensusMAID-3256Resolved issue: MAID-3256Fraser Hutchison
50 of
Benchmark Parsec with malice detection turned on
Description
Environment
None
Details
Assignee
UnassignedUnassignedReporter
Pierre ChevalierPierre ChevalierStory Points
6Components
Priority
Major
Details
Details
Assignee
Unassigned
UnassignedReporter
Pierre Chevalier
Pierre Chevalier
Story Points
6
Components
Priority
Major
Created April 9, 2019 at 4:23 PM
Updated April 9, 2019 at 4:24 PM
Activity
Pierre ChevalierApril 9, 2019 at 4:24 PM
Keeping in the backlog for now as we are first prioritising the Fleming work on the routing side. This is something we will come back to before testnets, though.
Benchmark the malice detection code to ensure there is no premature pessimisation hiding in there.
We want to benchmark two different cases:
Malice detection is turned on but no node behaves maliciously
Malice detection is turned on and some modes behave maliciously (test various kinds of Malice)