Transcript Is IP Going To Take Over The World: Offense

Is IP Going To Take Over The World:
Offense
Isaac Chung
Gary Bramwell
Claim: IP does not dominate global
communications
 VOIP
 Commercially available from Vonage, Skype,
Comcast, etc.
 Video On Demand and IPTV
 Streaming Video
Circuit switching is still big and established, but
IP is rapidly gaining
Claim: IP is not more efficient
Argues bandwith utilization in core
network is lower
 Actually provides less delay for end users
Response time is same for crowded
flows
 IP lets at least some traffic through, circuit
lets none
Claim: IP is not more robust
IP is a best-effort service
 Upper layer protocols like TCP deal with
this fact
 Transparent to application layer
Claim: IP is not simpler
So?
Increased diversity of clients and
sources requires complexity to handle
Trades complexity for flexibility
Claim: IP cannot support real-time
applications
“There is a widely-held assumption that IP network can support
telephony and other real-time applications”
 An assumption made true
 VOIP
 IPTV
 Streaming media/Webcasting
“It is doubtful a new solution…an displace the reliable and high
quality of service (QoS) provided by today’s TDM-based
infrastructre.
 True, but it has no need to replace it to support real-time apps
IP and Circuits living
together…mass hysteria!
Argues that IP could be used on the
fringes of network with a circuit-switched
core
 Don’t support claims that IP isn’t good at
traffic isolation, traffic engineering, fault
isolation, and manageability
Circuit switching has to manually
scale to match IP
“Similarly circuit switching has to respond
to the increases in traffic of packet
switching, by adapting its capacity along
core/edge gateways accordingly”
This won’t work because…
…Things scale too quickly to manually adapt
QuickTime™ and a
TIFF (Uncompressed) decompressor
are needed to see this picture.
Thanks to Geoff Huston. http://bgp.potaroo.net on April 1, 2007
Conclusion
 Paper argues against the plausibility of
services that are already in place
 Claims were based on comparing an
established system with a still-new Internet
 Paper had no actual conclusion or point