Network Analysis & Statistics

There are a number of issues:

perception of a circuit prpblem has really become a biology problem
little incentive for vendors/service providers to examine larger issues
technology races ahead (in part government-inspired) without sufficient inventigation

Comparision of user perspectives

telephony -- client <-> local <-> long distance <-> local <-> client

Internet more meshed

In the telephone world, the use of bandwidth is in short bursts.

In the Internet world, the use of bandwidth is widely variable.

issues: flow aggregation, high bandwidth*duration products, defeating aggregation capabilities

The Internet expects alot of aggregation. The newer (multi-media) applications expect more continuous flows that can't be aggregated (dedicated bandwidth).

Packet volumes as compared to byte volumes are related, but the number of bytes per packet is directly related to what applications require.

For more information, check http://www.nlanr.net/NA

Pro-Active Evolution

utilizing the evolving multicast technology
nimprove protocol/engine efficiencies
implement intelligent information caching to address redudent transmission of documents
Integration of information discovery tools (e.g. Harvest) that make more efficient use of network bandwidth in building information brokers and gathers that index the net

How can NANOG help?

investigate better understanding of Internet workloads at both component and systemic levels
consider models of an accounting/pricing architechture with multiple service qualities: not only incentives ...