Caspian Comes Out

COMMENTS

Newest First | Oldest First | Threaded View

<< < Page 8 / 9 > >>

rjmcmahon 12/5/2012 | 12:14:05 AM

re: Caspian Comes Out I have given up on the "political" process to drive down the local loop cost.

Then you have given up.

specifically by advances that do not require a truck roll to each home. Current contenders include 802.11g.

Sorry to be so blunt, but these are distractions which have little to do with solving our industry's problems.

Reply | Post Message | MESSAGES LIST | START A BOARD

arch_1 12/5/2012 | 12:14:04 AM

re: Caspian Comes Out [I said]
I have given up on the "political" process to drive down the local loop cost.

[rjmcmahon replied]
Then you have given up.

I respectfully disagree. We've tried to change the way monopolies deliver local-loop service for the last 7 years or so, and failed miserably. I see no reason to hope that the next 7 years will be any different. Fortunately, wireless technologies are improving very rapidly, which at least gives us some hope that we can render the monopolies irrelevant. I have much greater faith in Moore's law than I do in the FCC.

In my neighborhood, I waited for 10 years and we've never gotten DSL. There is NO technical reason for this. We finally got cable modems in the neighborhood, but it costs >$40/mo. As soon as it is technically feasible I'll start a neighborhood ISP using an 802.16 trunk and 802.11g local loops.

(reminder: this related to "Caspian Comes Out" via the observation that flow-based QoS is relevant only on the local loop.)

Reply | Post Message | MESSAGES LIST | START A BOARD

arch_1 12/5/2012 | 12:14:04 AM

re: Caspian Comes Out [I said]
>> But since fabric bandwidth is far cheaper than the external bandwidth it supports, it's cheaper and easier just to grossly overprovision. <<

[Mark Seery replied]
Waste not want not applies here as well IMO. There are good incentives for silicon efficiency still as overall system power is still something to be concerned about.

[I reply]
There are costs associated with gross overprovisioning in the fabric, and there are costs associated with clever ways to achieve the same result. I can choose to spend my money/gates/cycles/power on either or on both. A system with a 2x-speed fabric and crude feedback may perform as well as a system with a 1.2x-speed fabric and sophisticated QoS-aware feedback, and it may be cheaper.

Actually, I used the 2x-speed system more as an example. With this overspeed, it should be obvoius to all that even with a simple design the fabric is not the bottleneck. Note, however, that even the most sophisticated system in the world will need more than 1.0x-spped to be non-blocking, so the best the "sophisticated" approach can possibly do is double the fabric efficiency.

Reply | Post Message | MESSAGES LIST | START A BOARD

IP_freely 12/5/2012 | 12:14:03 AM

re: Caspian Comes Out Sounds like somebody needs a hug... chill bobby

Reply | Post Message | MESSAGES LIST | START A BOARD

BobbyMax 12/5/2012 | 12:14:03 AM

re: Caspian Comes Out Dear "Dr." Joe Average:

First all I must tell you my contributions to the world is much more than what you and your comminity, and your associates would ever be able to achieve would ever achieve. Youy are one of the lowest creature that the human kind has produced. Please go and perform community service. It might bring some goodness in you.

Reply | Post Message | MESSAGES LIST | START A BOARD

TheNet 12/5/2012 | 12:13:59 AM

re: Caspian Comes Out [Mark Seery wrote]
We do not make reservatiote on the road system because a) there is no incetive to provide better service, i.e. it is a best effort service, with some quite good differential service for those that travel in the "CBR" lane by traveling with a companion; but like many differential service algorithms, the system is broken for every one else some of the time, and the system works exceptionally well for every one some of the time.

[TheNet] Right on. Agreed.

There are some activities in life that are well suited to a reservation paradigm, and then there are many that are not. The notion that one networking mode, and perhaps even one network, will solve all problems may be a more broken thought than anything that is wrong with the individual network modes that we as an industry already understand.

[TheNet] Right on.
I think there is a perception that the evolution of networking ultimately involves one architectural paradigm to solve all problems. I think that's why people have difficulties with "flow based QOS with resource reservations" because they map every applications & services they know to this architecture. Sorry, our world doesn't count on "silver bullets". Like Mark Seery pointed out keenly, there is room from different solutions to different problems.

In the case of the current day Internet, all applications must use "Best Effort" paradigm. Many of those applications (such as email, instant messaging etc) are very suited to this BE traffic class. Going forward, I see a packet based networking paradigm where a sizable portion of the bandwidth is allocated to BE and the rest is allocated to CE (Committed Effort ie guarantees). Both traffic classes can share layer 1 containers (transport links), both can share switch fabric etc. It's a matter of proper scheduling and control.

So, don't worry guys, the world is not about to "micro manage email flows" :-)

[[email protected]]

Reply | Post Message | MESSAGES LIST | START A BOARD

crapshooter 12/5/2012 | 12:13:57 AM

re: Caspian Comes Out Boobie:

Did you play hockey before the mandatory helmet rule?

CS

Reply | Post Message | MESSAGES LIST | START A BOARD

gea 12/5/2012 | 12:13:56 AM

re: Caspian Comes Out BoobyMax wrote...

"First all I must tell you my contributions to the world"

By "contributions to the world" I assume you mean your Lightreading posts, no?

And in case you don't understand the irony, let's just say that a few minutes ago I flushed down a greater "contribution to the world" than you've ever made, Booby.

Reply | Post Message | MESSAGES LIST | START A BOARD

gladysnight 12/5/2012 | 12:13:46 AM

re: Caspian Comes Out Mark Seery wrote:
"The notion that one networking mode, and perhaps even one network, will solve all problems may be a more broken thought than anything that is wrong with the individual network modes that we as an industry already understand."
===============================

Thank You, Mark, for that eminently sensible insight. It has rarity value, if nothing else.

As another poster had pointed out, we DO have planes and trains as well as automobiles, and for reasons that are quite relevant to your comment.

Reply | Post Message | MESSAGES LIST | START A BOARD

whyiswhy 12/5/2012 | 12:13:46 AM

re: Caspian Comes Out Then again, why switch and/or route at all?

Trying to optimize switching topologies / algorithms without detailed knowledge of the specific traffic at each user for a specific network and set of assumptions / constraints is guaranteed to result in endless (and useless)debate.

JMHO

-Why

Reply | Post Message | MESSAGES LIST | START A BOARD

<< < Page 8 / 9 > >>