Blockstack Talk at Princeton CITP

Muneeb Ali gave a talk on decentralized DNS using blockchains at Princeton University. The talk was hosted by Princeton's Center for IT Policy, which is focused on the interaction of digital technologies with society and is also a powerful voice for Internet security and privacy.

Transcript

1. Decentralized DNS & PKI Muneeb Ali
2. • Problem • Background • Lessons from Namecoin • Design of Blockstack • Future Work Outline
3. Problem
4. Step 1: How do I talk to facebook.com? Step 2: Is this the correct facebook.com? — DNS gives route to facebook.com — PKI verifies authenticity of facebook.com
5. Step 3: How do I login to facebook.com? — Facebook does account management — Facebook gives usernames/passwords
6. DNS PKI Centralized User Data Authentication
7. Problems with the centralized model: — “Golden key” access
8. Problems with the centralized model: — “Golden key” access — Central point of failure
9. Problems with the centralized model: — “Golden key” access — Central point of failure — Data breaches
10. Decentralized User Data Decentralized Auth Decentralized DNS +PKI
11. Problems with DNS + PKI: — Internet censorship
12. Problems with DNS + PKI: — Internet censorship — Central points of trust and failure
13. Problems with DNS + PKI: — Internet censorship — Central points of trust and failure — Ease of use and deployment
14. Problems with DNS + PKI: — Internet censorship — Central points of trust and failure — Ease of use and deployment Goals for a decentralized DNS + PKI: — Extremely hard to censor — No central points of trust or failure — Low cost, keypair ownership by default
15. Background
16. How Blockchains Work: — It’s a file! — Append-only global log — Every node on the network has a consistent copy — Writes are slow, reads are fast Blockchain
17. Blockchain Bitcoin’s P2P Network
18. How Blockchains Work: 1 2 3 ….. 3000 Register hash(name) Update name
19. Lessons from Namecoin
22. Production system on Namecoin: — Used u/ namespace — Live between March 2014 and August 2015 — 33,000 registrations — Over 200,000 transactions
23. Security of blockchain:
24. Network latency:
25. Throughput:
26. Potential Selfish-mining:
27. Lessons from Namecoin: — Storage limitations (blockchain bloat) — Introducing new features (hard fork) — Failure of merged-mining — Security / throughput of blockchain — Other engineering challenges
28. Our Approach
29. General Challenges with Blockchains: — Storage limitations (blockchain bloat) — Introducing new features (hard fork) — Introducing new features (hard fork) — Slow writes — Endless ledger problem
30. Blockstack Network bitcoin node (bitcoind) blockstack server bitcoind peer connection rpc connection to bitcoind
31. name_op, hash name_op, hash name_op, hash name_op, hash name_op, hash Cryptocurrency Blockchain Virtual Blockchain DHT (routing info) Storage System A (data) Storage System B (data) lookup route lookup data lookup data Layer-1 Layer-2 Layer-3 Layer-4 Data Routing
32. Blockchain name_op, hash name_op, hash name_op, hash name_op, hash name_op, hash Bitcoin Blockchain Virtual Blockchain Control Plane Data Plane Storage Drivers S3 Dropbox Linux Unlimited Data Storage:
35. 300,000 200,000 100,000 Blockstack bytes written — Largest non-financial production system on Bitcoin
36. Fast Bootstrapping: (1) Records are organized into a Merkle tree (2) whose root is fed into the consensus hash, (3) along with a geometric series of prior consensus hashes
37. Light Nodes:
38. General Challenges with Blockchains: — Storage limitations (blockchain bloat) —> Unlimited data — Introducing new features (hard fork) — Introducing new features (hard fork) —> Virtualchain — Slow writes —> Get updates off blockchain path — Endless ledger problem —> Fast bootstrapping Goals for a decentralized DNS + PKI: — Extremely hard to censor —> state level actor — No central points of trust or failure —> check — Low cost, keypair ownership by default —> check
39. Blockstack CLI Blockstack gives you fast, secure, and easy-to-use DNS, PKI, identity management, and custom namespaces on the blockchain
41. Future Work
42. Pricing Models for Namespaces:
43. General-purpose State Machines:
44. In the pipeline: — Securing DHTs with Blockstack — Unbound (DNS server) integration — Bootstrapping trust for cloud infrastructure (OpenCloud) — Identity and authentication services (Blockstack Auth) — Stateless servers (blogging, email, etc)
45. plus open-source contributors and 600+ community members
47. Thank You! Web: http://blockstack.org Code: github.com/blockstack