| Moderator: Rick Summerhill, Director Network Research, Architecture, and Technologies, Internet2 | |
| Joint Techs | |
| Salt Lake City, UT |
| Early days - 56 kbps | ||
| NSF T1 Network | ||
| NSF T3 Network | ||
| Privatization | ||
| vBNS OC-12 primarily | ||
| Abilene OC-48 | ||
| Current Abilene OC-192 | ||
| Trend to larger bandwidth circuits | ||
| Is this trend about to end? Parallelization? Change in Architecture? | ||
| What does the community want now? | ||
| It is possible to saturate a 10 Gbps backbone with flows between two PC class machines | ||
| Recent flows across the Abilene network have reached 9.5 Gbps using dual bus PCs having dual Ethernets | ||
| Research networks have applications that require large flows across long distances | ||
| On commercial networks, most data flows are small | ||
| Parallelizing may be sufficient to add capacity | ||
| What are the economic effects/factors? | ||
| Technology exists | ||
| 40 Gbps modulation techniques | ||
| Muxing at various levels | ||
| Economics of higher bandwidth circuits | ||
| What are the technological directions, both LAN and WAN – 40 Gbps? 100 Gbps? | ||
| Optics and electronics? | ||
| Parallel and single optics? | ||
| Practical limitations? | ||
| Limitations of parallelization at layers two and three? | ||
| What are the driving applications? | ||
| What are the economic incentives? | ||
| Impacts of Standardization? | ||
| Timeframe? 40 Gbps? 100 Gbps? | ||
| Jeff Cain, Cisco Systems | |
| Ken Gold, Nortel Networks | |
| Hong Liu, Juniper Networks | |
| Jeff Livas, Ciena | |
| Stephan Garrison, Force10 Networks | |
| Mark Stine, Qwest Communications |