Case study · Router Protocol
Cross-Chain Forwarder at Scale
Reduced critical transaction execution from 16 seconds to under 1 second while handling 2M+ transactions and $1B+ in volume.
All consulting services
Consulting · Blockchain & Web3
Build custody and cross-chain systems that hold up when the value moving through them is real.
Blockchain infrastructure fails differently from ordinary backend work: mistakes are irreversible, the adversary is financially motivated, and the slow path is often the one holding customer funds. I have built exchange-grade cold custody and a cross-chain forwarder carrying over $1B in volume, and the lesson from both is that security and throughput are design problems you solve together or not at all.
01Who this is for
Where this work fits
You might recognize
- Exchanges and custodians who need institutional-grade key management rather than a hot wallet with extra steps.
- Teams running cross-chain or bridge infrastructure where latency on the critical path costs revenue.
- Protocols whose indexing pipeline cannot keep up with burst transaction load.
- Teams whose custody policies are hard-coded, so every rule change ships as application code.
What it costs you
- Key material reachable from an online system is a single compromise away from irreversible loss.
- Slow critical-path execution caps throughput and degrades the product precisely when volume spikes.
- Policy logic embedded in application code means a deploy for every limit change, and no audit trail.
- Without structured audit events, incident response and compliance review start from raw logs.
02Approach
How the work runs
Security and throughput get designed together, because fixing one after the other usually means redoing both.
Draw the trust boundary first
Decide what must never leave hardware, then design around that constraint. HSM-backed keys and multi-party signing follow from it rather than being bolted on.
Separate hot operational paths from cold signing
Deposits, balances, and monitoring run online; anything that moves funds crosses into an isolated path, which shrinks the blast radius of any single compromise.
Make policy data, not code
A rule engine evaluating allowlists, limits, and thresholds at request time means policy changes without a deploy, and every decision leaves an audit record.
Profile the critical path before optimizing it
On the Router forwarder the win came from finding where the 16 seconds actually went — indexing, validation, and coordination — rather than optimizing by intuition.
Scale indexing and execution independently
Event-driven services over Kafka and Redis, so burst load on one stage does not stall the others.
03Tradeoffs
What each choice costs
Every one of these has a downside. Here is what you are accepting and what you get for it.
Cold custody over hot-wallet convenience
Withdrawals get slower and operationally heavier, requiring a signing ceremony. That friction is the control; removing it removes the protection.
A rule engine over hard-coded policy
More upfront design and a system that must itself be correct. In return, policy changes stop being code deploys and become auditable data changes.
Event-driven over synchronous pipelines
Harder to reason about and harder to debug end to end. Necessary when indexing and execution need to scale on different curves.
04Deliverables
What you get from a review
Scoped as a focused assessment. Anything beyond this is agreed before it starts, not assumed.
- A custody architecture review covering key storage, signing flow, and the trust boundary.
- A threat-informed assessment of where funds or key material are currently exposed.
- A performance analysis of the critical transaction path, with the measured bottlenecks named.
- A policy and audit model that supports compliance review without bespoke engineering per rule.
- A prioritized roadmap separating security-critical fixes from throughput work.
05Evidence
This is not theoretical
The approach above describes work I led, not a framework I read about.
Case study · CoinDCX
HSM-Backed Cold Custody Wallet Platform
Designed and delivered a cold-custody wallet platform with HSM-protected keys, multi-signature authorization, and a sub-30ms configurable policy engine. Served 100K customers and 10K daily transactions with full auditability and no per-rule code deploys.
Technical write-up
Two Servers, One Secret: Envelope Encryption Inside Nitro Enclaves with KMS
Attestation-bound decryption with KMS — the same trust-boundary thinking that custody key management depends on.
Next step
Tell me what you are securing, and what it has to keep up with.
Custody model, throughput targets, and where you think the weak point is. If the honest answer is that your current setup is fine, I will say so.