Podcasts
Subscribe 
Industry Bloggers
Om Malik
Rich Tehrani (TMCNET)
David Isenberg
Mark Evans
Alec Saunders
Martin Geddes
Andy Abramson
Andrew Schmitt
Richard Bennett
Ken Camp
Brough Turner
Dan York
Jules
Infinera blog
Steve's Tech Blog
David Beckemeyer
Garrett Smith
Michael J. Santarcangelo, II
Robert Powell
Dave Rustin
NSP Strategist
Ike Elliott
No Jitter
Kevin Werbach
Recent blog posts
- Public Service Announcement - Identity and Security – Pay attention folks!
- The phases of IPv4 exhaustion
- All I want for Christmas...
- Something better than IPv6
- Virtualization – Part 3 - The Abstraction of Applications
- Gut check
- Top 10 in VoIP
- Surprise me! Is Innovation really just more for less?
- And now you see why broadband is not more ubiquitous…
- No major IPv6 content in the near future
User login
Navigation
Terms
Technorati







End-to-end principle != intelligence at the edge
This might surprise you, but I agree with you.
Embedding "read only" intelligence in the network to be able to help determine what's going on is a good thing. Indeed, it in no way contradicts the design principles that made the Internet itself successful: place functions in the right context, make as few assumptions as possible up front as to how the network will be used, and preserve its option value. It's when you build an architecture like IMS that you start to get into trouble.
The "stupid network" is also somewhat of a simplification, since it is clearly very smart if you look at the right layers. The "layers" model is also just that: a model. It hides a great deal of complexity about the real world, and (over) simplifies some of the ways in which information (and "side channel data" like identity) bleeds between those layers. For example, the geographical location where the user resides is something of considerable importance and interest, but isn't transmitted or modelled in any way in the Internet paradigm.
I recently sat down and brainstormed all the ways in which the IP/Internet model doesn't quite fit reality. I came up with over 50 items! Each is a possible route to making money and avoiding being "dumb piped".
But...
The dumb pipe clearly works most of the time, even for voice. I travel the world, and at a fraction of the cost of what a big enterprise paid for communications services I generally get a richer and better experience as a solo practitioner. Something is clearly wrong. User-facing telcos that launch products that are implementable on a "fat, dumb pipe" and don't take advantage of the special position of the operator or it's non-network assets (cash collection, invoicing, distribution, retail etc.) deserve to get a whipping in the marketplace.
The good news is GX is one of the few telcos that really understands these phenomena well, and is a "carrier's carrier", so all the best! In our Telco 2.0 model, you're definitely "Telco 2.0 compliant" in that you focus on one horizontal activity, and do it well.