The security trust chain is broken (but we're working on it!)

ShowMeCon Security 2015

Transcript

1. The  Security  Trust  Chain  is   Broken   But  we’re  working  on  it Kenneth White ShowMeCon Security 2015 St Louis June 9, 2015
2. Topics •  Open Crypto Audit Project •  Existing trust chains •  OpenSSL audit •  Emerging •  Final thoughts
3. This  is  me Twitter @kennwhite Talks speakerdeck.com/kwhite OCAP https://opencryptoaudit.org/people
4. Open  Crypto  Audit  Project •  OCAP originally formed to manage community- funded TrueCrypt audit •  Independent technical research public interest organization •  Technical Advisory Board: academic, industry, and legal experts in security •  Mission: Research, analysis & education around technical security in open source software •  Focus: software security, cryptography engineering, public awareness •  Current project: OpenSSL audit
5. The  Software  Security   Trust  Chain 1 year post-Heartbleed • Most serious CVEs are (rarely) about the crypto • But the (most widely deployed) crypto trust chain is fragile • Key pieces of the core Internet network stack are virtually unexamined, and little understood
6. The  Software  Security   Trust  Chain 1 year post-Heartbleed • Most serious CVEs are (rarely) about the crypto • But the (most widely deployed) crypto trust chain is fragile • Key pieces of the core Internet network stack are virtually unexamined, and little understood
7. The  Software  Security   Trust  Chain Questions How well do you know the network stack you’ve deployed? How about your technical staff? Do you/they understand your core dependencies?
8. The  Software  Security   Trust  Chain Questions How well do you know the network stack you’ve deployed? How about your technical staff? Do you/they understand your core dependencies?
9. The  Software  Security   Trust  Chain Are you sure?
12. Mature network hardware
13. A $100K commercial load balancer compromised by a browser ID string
14. Let’s really look at the whole security trust chain…
15. Internet  Core  Trust  Chain For example: o  XML parsers (libxml2, Expat, SimpleXML…) o  Image generators (libpng…) o  Internationalization libraries (libIDN) o  Compression (libzma) o  ASN.1 & x509 (everywhere) o  Middleware core: BouncyCastle, Spring, Struts… o  Deeper: libBFD, libCurl, IPSec netkey, pluto, l2tp
16. Internet  Core  Trust  Chain Time to look really closely, at, say: o  XML parsers (libxml2, Expat, SimpleXML…) o  Image generators (libpng…) o  Internationalization libraries (libIDN) o  Compression (libzma) o  ASN.1 & x509 (everywhere) o  Middleware core: BouncyCastle, Spring, Struts… o  Deeper: libBFD, libCurl, IPSec netkey, pluto, l2tp
18. BFD  is  a  BFD
19. Are you kidding me?!
20. Wait, it gets better. Ever use the shell utility ‘less’?
23. BFD  is  a  BFD.    
24. BFD  is  a  BFD.   But  most  Linux  admins  have   never  even  heard  of  it
25. libcurl
27. Let’s go higher up
28. Basic server certificate deployment is a solved problem, yes?
29. Basic server certificate deployment is a solved problem, yes?
30. Don’t underestimate the impact of applied research
31. Don’t underestimate the impact of applied research
32. Network transport has integrity, yes?
33. Network transport has integrity, yes?
34. Network transport has integrity, yes? https://gist.github.com/kennwhite/1f3bc4d889b02b35d8aa
35. Ad networks are trusted for arbitrary client code, yes?
39. But  trust  is  complicated…
40. But  trust  is  complicated…
41. But  trust  is  complicated…
42. And  this  isn’t  helping
43. The  Security  Trust  Chain   is  Broken  
44. The  Security  Trust  Chain   is  Broken   But  we’re  working  on  it
45. The  OpenSSL  Audit
46. The  OpenSSL  Audit •  Commissioned by Linux Foundation’s Core Infrastructure Initiative (CII) •  Ambitious Scope o Independent review o Coordinating closely with OpenSSL core team o Delayed for v. 1.1 maturity (significant refactor) o Diverse, complex codebase o Linux, BSDs, Windows, OSX, SRV5 (AIX, HP-UX, Solaris) o Intel x86 (incl. AES-NI), ARMv7, MIPS, PowerPC, Alpha… o FIPS module
47. OpenSSL  Audit •  Goals •  Thorough public security analysis of the core code in the next major release of OpenSSL •  Demonstrate viability of a reusable open source test harness framework •  Foster web-scale peer-reviewed public tools & data sets for protocol & negotiation analysis
48. OpenSSL  Audit Rough metrics: 412-494K total SLOC OpenSSL v. 1.1 master (2015-03-14)
49. OpenSSL  Audit •  Phase 1 •  BigNum: multiprecision ints, constant time, blinding •  BIO (focus on composition & file functions) •  ASN.1 & x509 (cert & key parsing, DER/PEM decoding, structs, subordinate chains) •  93M cert corpus, “Frankencert” fuzzing •  Phase 2 •  TLS state machine •  EVP (PKI constructions, H/MACs, envelopes) •  Protocol flows, core engine implementation •  Memory management •  Crypto core (RSA, SHA-2, DH/ECDH, CBC, GGM…)
50. OpenSSL  Audit Caveats •  Schedule, funding, or quality: Pick 2 •  High Priority •  Major architectures •  Modern (TLS 1.3) protocols & primitives •  DH, ECC, signatures, ASN.1 & x509 •  Non-crypto constructions (data structures, memory management, core API/ABI hooks) •  Lower Priority •  AES implementation (finite field tables, matrix transformations, etc. TBD, possibly in phase 3 formal academic analysis) •  RC4 •  S/MIME •  OpenSSL s_server (smtp-aware web server!)
51. Emerging
52. Emerging •  Better primitives and core crypto •  TLS 1.3 •  NaCl/LibSodium, ChaCha20/Poly1305 (OpenSSL soon) •  Marlinspike et al’s work on OTR, axolotl ratchet •  Trevor Perrin’s work on public key pinning & TLS core •  Containers smaller surface (Docker, Rocket, LXC) •  Let’s Encrypt (Mozilla, Akamai, Cisco, EFF) •  USG: All fed websites & services HTTPS-only •  Open threat feeds (AlienVault Open Threat Exchange v2) •  Verizon Data Breach Investigation Report model
53. Parting  Thoughts o  VZ DBIR: 99.9% of successful exploits last year relied on a CVE more than a year old o  Intelligence & defense collaboration & sharing is critical o  Encryption isn’t a magic bullet o  Understand your threat model o  Stronger security chain will require better cooperation, more open exchanges, and trust
54. Parting  Thoughts o  We are very much in the golden age of web security o  We are beginning a serious re-examination of the core stack and fundamental trust chains
55. The  Security  Trust  Chain   is  Broken   But  we’re  working  on  it
56. Be  careful  out  there,  folks
57. Contacts OCAP admin @ opencryptoaudit . org OCAP https://opencryptoaudit.org/people Twitter @kennwhite Talks speakerdeck.com/kwhite