What Makes For a Successful Protocol?

• IETF Internet draft, 1/11/2007
• D. Thaler, B. Aboba, IAB.

Annotations

What is success?

• a successful protocol = both meets its original goals and is widely deployed.

What is failure?

• no mainstream implementations.
• no deployment.
• no use.
• cannot be determined until 5-10 years for an average protocol.

Examples of successful protocols:

• Inter-domain:
  • IPv4 [RFC0791]
  • TCP [RFC0793]
  • HTTP [RFC2616]
  • DNS [RFC1035]
  • BGP [RFC4271]
  • UDP [RFC0768]
  • SMTP [RFC2821]
  • SIP [RFC3261]
• Intra-domain:
  • ARP [RFC0826]
  • PPP [RFC1661]
  • DHCP [RFC2131]
  • RIP [RFC1058]
  • OSPF [RFC2328]
  • Kerberos [RFC4120]
  • NAT [RFC3022]

Effects of wild success:

• both good and bad.
• if a protocol is used for a purpose other than what it was designed for
  • undesirable side effects
  • performance problems
  • work around design limitations without complete understanding of effect on overall protocol behavior.
  • high value targets for attackers (death by success)

How to overcome the initial lack of implementation/deployment?

• address a critical and imminent problem.
• provide a “killer app” with low deployment costs.
• under existing unmodified applications.
• narrow down to remove complexity.
• government incentives/disincentives (IPv6)

Initial Success Factors:

• Positive net value (meet a real need). Benefits of deploying the protocol outweigh the costs.
  • Example, IPsec interferes with netflow, deep packet inspection, etc.
• Incremental Deployability
  • protocols that can be deployed by a single group have a greater chance of success than those that require cooperation across organization.
  • protocols that don’t require changes to infrastructure. (NAT)
  • only one end changes vs both ends to support the protocol.
• Open Code Availability (IPv4 vs IPX)
• Open Specification Availability
• Open Maintenance Processes
• Freedom From Usage Restrictions
  • without legal / financial hindrance (Intellectual Property)
• Good Technical Design

Wild Success Factors:

• Extensible
• No Hard Scalability Bound
• Threats Sufficiently Mitigated
  • good: SSHv1, 802.11 WEP
  • bad: early server-based multiplayer games

Case Studies:

• HTTP/HTML vs. Gopher and FTP
• IPv4 vs. IPX
• SSH
• Inter-domain IP Multicast
• Wireless Application Protocol (WAP)
• Wired Equivalent Privacy (WEP)
• RADIUS vs. TACACS+
• NAT