Skip to content

Commit 39ff6a9

Browse files
ch1boch1bo
committed
Consistent language across threat categories
1 parent b9f7b29 commit 39ff6a9

File tree

1 file changed

+32
-32
lines changed

1 file changed

+32
-32
lines changed

‎docs/threat-model.md‎

Lines changed: 32 additions & 32 deletions
Original file line numberDiff line numberDiff line change
@@ -135,12 +135,12 @@ In Leios, successful grinding would allow attackers to increase their probabilit
135135
- Amplify costs of VRF grinding Ouroboros Phalanx ([CIP-0161](https://github.com/nhenin/CIPs/tree/CIP-Ouroboros-Phalanx/CIP-0161)), which introduces computational cost amplification to make grinding attacks economically infeasible by increasing grinding costs by approximately 10^10 while maintaining lightweight computation for honest participants.
136136
- Standard key management practices protect against VRF key compromise.
137137

138-
| # | Method | Effect | Resources | Mitigation |
139-
|---|------------------------------------|-------------------------------------------|----------------------------------|-----------------------|
140-
| 1 | Any threat to Praos | Leios is only as secure as Praos | - | Dependency on Praos |
141-
| 2 | VRF grinding on EB eligibility | Increased probability of EB creation | CPU & stake (>20%) | Ouroboros Phalanx R&D |
142-
| 3 | VRF grinding on voting eligibility | Increased probability of voting selection | CPU & stake (>20%) | Ouroboros Phalanx R&D |
143-
| 4 | VRF key compromise | Unfair advantage in eligibility | Very high - cryptographic attack | Strong key management |
138+
| # | Method | Effect | Resources | Mitigation |
139+
|---|------------------------------------|-------------------------------------------|--------------------------------|-----------------------|
140+
| 1 | Any threat to Praos | Leios is only as secure as Praos | - | Dependency on Praos |
141+
| 2 | VRF grinding on EB eligibility | Increased probability of EB creation | CPU, stake (>20%) | Ouroboros Phalanx R&D |
142+
| 3 | VRF grinding on voting eligibility | Increased probability of voting selection | CPU, stake (>20%) | Ouroboros Phalanx R&D |
143+
| 4 | VRF key compromise | Unfair advantage in eligibility | Very high cryptographic attack | Strong key management |
144144

145145
### Equivocation
146146

@@ -154,12 +154,12 @@ A particularly interesting case involves BLS key compromise for voting. When a B
154154

155155
**Mitigation**: The Leios protocol specification includes explicit equivocation detection mechanisms that identify misbehaving nodes and equivocation proofs are forwarded through the network. For BLS key compromise, key rotation procedures enable recovery while defensive equivocation provides interim protection. Double voting has limited safety impact since multiple certificates can exist but only RB inclusion determines chain progression.
156156

157-
| # | Method | Effect | Resources | Mitigation |
158-
|---|--------------------|---------------------------------------|--------------------------------------------|------------------------------------------|
159-
| 5 | EB equivocation | Lower throughput, resource waste | Stake for block production | Equivocation detection per Leios design |
160-
| 6 | Vote equivocation | Interferes with certificate creation | Stake for voting eligibility | Equivocation detection per Leios design |
161-
| 7 | Double voting | Multiple certificates, resource waste | Stake for voting eligibility | Chain selection prioritizes RB inclusion |
162-
| 8 | BLS key compromise | Unauthorized vote creation | Cryptographic attack or social engineering | Key rotation + defensive equivocation |
157+
| # | Method | Effect | Resources | Mitigation |
158+
|---|--------------------|---------------------------------------|------------------------------|------------------------------------------|
159+
| 5 | EB equivocation | Lower throughput, resource waste | Stake for block production | Equivocation detection per Leios design |
160+
| 6 | Vote equivocation | Interfere with certificate creation | Stake for voting eligibility | Equivocation detection per Leios design |
161+
| 7 | Double voting | Multiple certificates, resource waste | Stake for voting eligibility | Chain selection prioritizes RB inclusion |
162+
| 8 | BLS key compromise | Unauthorized vote creation | Cryptographic attack | Key rotation, defensive equivocation |
163163

164164
### Inaction and Nuisance
165165

@@ -177,15 +177,15 @@ The incentive structure of Leios is designed such that most inaction attacks are
177177
>
178178
> TODO: Move certificate forging somewhere else as it is quite different?
179179
180-
| # | Method | Effect | Resources | Mitigation |
181-
|----|--------------------------------------|----------------------------------|------------------------------|------------------------------------------|
180+
| # | Method | Effect | Resources | Mitigation |
181+
|----|--------------------------------------|-----------------------------------|------------------------------|------------------------------------------|
182182
| 9 | Decline to create EB | Lower throughput | Stake for block production | Reduced rewards |
183183
| 10 | Decline to vote | Lower throughput | Stake for voting eligibility | Reduced rewards |
184-
| 11 | Create invalid EB | Lower throughput, resource waste | Stake for block production | Reduced rewards, Validate before forward |
185-
| 12 | Create invalid vote | Lower throughput, resource waste | Stake for voting eligibility | Reduced rewards, Validate before forward |
186-
| 13 | Reference invalid transactions in EB | Lower throughput, resource waste | Stake for block production | Reduced rewards, Validate before forward |
184+
| 11 | Create invalid EB | Lower throughput, resource waste | Stake for block production | Reduced rewards, validate before forward |
185+
| 12 | Create invalid vote | Lower throughput, resource waste | Stake for voting eligibility | Reduced rewards, validate before forward |
186+
| 13 | Reference invalid transactions in EB | Lower throughput, resource waste | Stake for block production | Reduced rewards, validate before forward |
187187
| 14 | Include invalid certificate in RB | Lower throughput, resource waste | Stake for block production | Certificate verification |
188-
| 15 | Forge certificate without quorum | Manipulate transaction inclusion | Cryptographic attack | Strong BLS cryptography |
188+
| 15 | Forge certificate without quorum | Manipulate transaction inclusion | Cryptographic attack | Strong BLS cryptography |
189189

190190
### Omission and Manipulation
191191

@@ -201,12 +201,12 @@ SPOs concerned about front-running competition may choose to bypass the EB mecha
201201

202202
**Mitigation**: The primary defense is the memory pool design - omitted transactions remain available for inclusion in subsequent honest blocks, limiting censorship effectiveness. The distributed nature of block production means no single actor can permanently censor transactions. Detection of MEV extraction is challenging since legitimate transaction selection and ordering can appear similar to value extraction. Mitigation options are limited since EB opportunities are coupled to RB opportunities and cannot be parameterized separately.
203203

204-
| # | Method | Effect | Resources | Mitigation |
205-
|----|---------------------------------------------|---------------------------------------------|----------------------------|--------------------------------------------------------------|
206-
| 16 | Omit transactions from EB | Reduces throughput, temporary censorship | Stake for block production | Memory pool persistence |
207-
| 17 | Reorder transactions in EB | MEV, market manipulation | Stake for block production | Limited detection capability |
208-
| 18 | Insert or replace transactions in EB | MEV, market manipulation | Stake for block production | Limited detection capability |
209-
| 19 | Ignore certificates, include txs in RB only | Reduces EB throughput, avoids front-running | Stake for block production | Reduced rewards, Self-limiting when load exceeds RB capacity |
204+
| # | Method | Effect | Resources | Mitigation |
205+
|----|---------------------------------------------|-----------------------------------------|----------------------------|--------------------------------------------------------------|
206+
| 16 | Omit transactions from EB | Lower throughput, temporary censorship | Stake for block production | Memory pool persistence |
207+
| 17 | Reorder transactions in EB | MEV, market manipulation | Stake for block production | Limited detection capability |
208+
| 18 | Insert or replace transactions in EB | MEV, market manipulation | Stake for block production | Limited detection capability |
209+
| 19 | Ignore certificates, include txs in RB only | Lower throughput, avoid front-running | Stake for block production | Reduced rewards, self-limiting when load exceeds RB capacity |
210210

211211
### Data withholding
212212

@@ -231,11 +231,11 @@ A particularly dangerous and sophisticated variant targets blockchain safety its
231231
>
232232
> TODO: Should this be also about network delays?
233233
234-
| # | Method | Effect | Resources | Mitigation |
235-
|----|-------------------------------------------------|-----------------------------------------------------|-----------------------------------------|----------------------------------------------------------|
236-
| 20 | Withhold announced EB or endorsed transactions | Reduces throughput | Stake for block production | Connection timeouts, peer pruning |
237-
| 21 | Selectively withhold data from voting committee | Prevent honest EB certification, reduces throughput | Network position control | Redundant peer connections, diffusion monitoring |
238-
| 22 | Selectively withhold data from honest nodes | Allow certification, but delay block propagation | Network position control + modest stake | L_diff parameter sizing, worst-case diffusion validation |
234+
| # | Method | Effect | Resources | Mitigation |
235+
|----|-------------------------------------------------|----------------------------------------------------|----------------------------------------|----------------------------------------------------------|
236+
| 20 | Withhold announced EB or endorsed transactions | Lower throughput | Stake for block production | Connection timeouts, peer pruning |
237+
| 21 | Selectively withhold data from voting committee | Prevent honest EB certification, reduce throughput | Network position control | Redundant peer connections, diffusion monitoring |
238+
| 22 | Selectively withhold data from honest nodes | Allow certification, delay block propagation | Network position control, modest stake | L_diff parameter sizing, worst-case diffusion validation |
239239

240240
### Protocol bursts
241241

@@ -252,9 +252,9 @@ The attack magnitude depends on the adversary's stake proportion and EB size par
252252

253253
**Mitigation**: The primary defense is traffic prioritization implementing freshest-first delivery semantics - Praos traffic must be preferred over Leios traffic, and fresh Leios traffic over stale traffic. However, some infrastructural resources cannot be prioritized perfectly, including CPU, memory, disk bandwidth, and network router buffers. The attack's magnitude is bounded by the adversary's stake proportion, but ultimately requires engineering solutions for effective prioritization during burst conditions.
254254

255-
| # | Method | Effect | Resources | Mitigation |
256-
|----|-------------------------------------------|----------------------------------------------------------------------------|-----------------------------------|------------------------------------------------|
257-
| 23 | Withhold then release large number of EBs | Bandwidth saturation, processing delays, potential Praos timing disruption | Proportional to adversarial stake | Freshest-first delivery, Traffic prioritization |
255+
| # | Method | Effect | Resources | Mitigation |
256+
|----|-------------------------------------------|----------------------------------------------------------------------------|--------------------------------|-------------------------------------------------|
257+
| 23 | Withhold then release large number of EBs | Bandwidth saturation, processing delays, potential Praos timing disruption | Stake (proportional magnitude) | Freshest-first delivery, traffic prioritization |
258258

259259
### Legacy threats
260260

0 commit comments

Comments
 (0)