Future Internet

Context
- original design: everyone shared a common goal
- today: many players with conflicting/competing goals

Some (near) failures in enhancing the core Internet and IP
- Early Congestion Notification (ECN)
- QoS
- mobile IP
- IP multicast
- BGP security (and, to a lesser extent, DNS security)
- IPv6

Some successes "below" IP
- MPLS and VPNs
- WiFi and cellular

History (can research fields be in "a mood"?)
- starting in early 2000s (during dot.com boom)
- "looking over the fence at networking" report
- researchers concerned about "ossification" and how to best do networking research
     * intellectual ossification: pressure for compatibility with existing Internet
     * infrastructure ossification: limited ability to effect change on the Internet
- theme #1: measurement, modeling, management research (existing protocols and behaviors)
- theme #2: fresh areas (overlay networks, P2P like Chord, wireless, data centers)
- theme #3: strategies for incremental deployment (economic, overlay, etc.)
- theme #4: disruptive prototypes on testbeds/platforms (planetlab, emulab, orbit, 
       and later GENI, OpenFlow/SDN)
- theme #5 (closely related): clean-slate networking research (vs. evolutionary)

Players
- users
- commercial ISPs
- governments (enforce laws, protect consumers, regulate commerce, ...)
- intellectual property rights holders
- content providers

Tussle
- ongoing contention among parties with conflicting interests
- timescale: design time, configuration time, and run time
- design for variation in outcome
- modularize along tussle boundaries
- design for choice

Examples
- provider lock-in from IP addressing (consumers benefit from DNS, DHCP)

Discussion on enabling choice and evolution
- network virtualization 
- overlays
- programmability (e.g., SDN, white-box switches)
- extensible header formats (e.g., IP/TCP options) and types (DNS records)
- client and server code written by the same party

Discontinuities that can lead to change
- rise of the "hyper giants" (increasing footprint, close to the edge)
- IoT (e.g., massive increase in devices)
- developing world (a chance to do something different)
- privacy (e.g., most traffic encrypted, desire to control routing)
- SDN (refactoring who gets to innovate)