Twitxy Saved

Prateek Reddy

@PrateekReddyy

Researcher @MarlinProtocol | Formerly @ethereum
Joined February 2018
148 Following 347 Followers 62 Posts

Latest posts 5

Prateek Reddy @PrateekReddyy
Documented the expanded version of the monitoring protocol for Marlin Oyster enclaves along with FAQs based on the questions below

research.marlin.org/t/monitoring-p…
research.marlin.org Monitoring protocol for Oyster Instances Some FAQs based on https://x.com/saxenism/status/2002311606752215253 Do you assume auditor-operator separation? If an operator can also register as an auditor, how do you prevent self-auditing or...
Rahul Saxena Verified account @saxenism Dec 20, 2025
This is a pretty cool attempt by @MarlinProtocol to address the availability concerns in any TEE-heavy protocol.

But I had some questions around it, mainly arpund edge cases found by a little threat modelling:

1. Do you assume auditor-operator separation? If an operator can also register as an auditor, how do you prevent self-auditing or “friendly auditor” capture?

2. What’s the Sybil resistance model for auditors? What stops an attacker from spinning many auditor identities to bias assignments or local majorities?

3. For incorrect audit bits, delayed reveal makes challenges possible once R_{T_i}​​ is posted. But for “offline” reports, does the chain get a cryptographic proof that the audit request was delivered and the response arrived?

Absence claims are asymmetric. If an auditor says “no response”, what evidence exists either way?
Have you considered receipt-style proofs (enclave-signed challenge response or transcript hash) so “I responded” is provable on-chain?

4. The host/provider controls scheduling and networking. Even with encrypted payloads, they can shape traffic and selectively drop. How do you model:
+ healthy-to-auditors but degraded-to-users
+ blocked auditor traffic but normal user traffic
+ or the inverse (users blocked, auditors allowed)

5. Do you rely on audit traffic being indistinguishable from user traffic? In practice, hosts can still use metadata, routing and timing.

6. Monitoring is periodic (per Age/slot). Does that create “between-check” windows for strategic micro-downtime, censorship, or MEV-timed unavailability? Do you randomize challenges to reduce predictability?

7. What happens if the instance crashes near epoch end and cannot reveal R_{T_i}​​? Is there a commitment/receipt mechanism earlier in the epoch that avoids ambiguous states?

8. Slashing prices unavailability, but does not prevent it if the payoff is higher than the penalty. How do you think about strategic unavailability where the attacker’s profit dominates slashing?

9. This protocol seems great for detection and penalties, but what’s the story for safe recovery when downtime happens anyway? In privacy/TEE-heavy systems, graceful fallback often means leaking what you were protecting. How do you handle continuity without unsafe fail-open paths?

10. Beyond incentives, do you enforce redundancy requirements or recommend operational patterns (replicas, failover, multi-provider) to reduce correlated downtime?

----

Genuinely curious how you’ve reasoned about these edge cases. If there are parts of the spec that already address some of this, I’d love pointers

cc: @PrateekReddyy
View quoted post on X
0 replies 4 reposts 11 likes 2.2K views View on X
Prateek Reddy @PrateekReddyy
Looking forward to the discussion today!
Join us tomorrow for a panel with @mpiekarska8 (@Hyperledger) @ChayanKeshari (@Cognizant) @daviderutter1 (@inside_r3) and @PrateekReddyy (Marlin) and @SilviaAtta (ABI)!

Brought to you by @labosnomades!

To listen to them click here: https://t.co/5osdPcs4pb
View quoted post on X