Search:  

 
 
   All ForumsHot TopicsGallery






how-to block ads


 
Forums » Don't Fear The Bandwidth Apocalypse » You had me until...
Search Topic:
Uniqs:
317
Share Topic:
RSS topic:
toggle:
flat / full
normal / watch
Post a:
Post a:
Is it that hard »
« lol...  
AuthorAll Replies

jester121
Premium
join:2003-08-09
Lake Zurich, IL
·surpasshosting
·ViaTalk

You had me until...

quote:
"Deep Packet Inspection devices capable of detecting what applications end points are using are both new and intrusive. They do not perform a task of Network management, they are performing tasks of Session and Application management. These are both new and inappropriate roles for Internet Service Providers."
The first part is just plain wrong, upper layer filtering and control have been around for a long time.

The second is merely semantic masturbation -- the phrase network management doesn't mean that all higher levels of the OSI model are ignored, and Topolski knows it.

He'd make a good politician -- make some solid statements at first, get loosened up, and then shove your foot in your mouth trying to hit one out of the park, thus blowing your credibility.


topoloski dumb

@cox.net

The article sites a dude named Topoloski (spelling?). The guy is way way off-base. We can complain all we want, but what Comcast wants to do is fine and ... needed.

What Topo does not understand that if the FCC keeps Comcast, AT&T and other providers from doing what is necessary to move bandwidth from heavy users then all we are going to get is much higher prices. And or poor performance. What Comcast is attempting to do is come closer to guaranteeing quality of service for all customers.

qworster

join:2001-11-25
Los Angeles, CA
·DSL EXTREME
·Brand X Internet
·RoadRunner Cable
·Vonage


4 edits
said by topoloski dumb :

The article sites a dude named Topoloski (spelling?). The guy is way way off-base. We can complain all we want, but what Comcast wants to do is fine and ... needed.

What Topo does not understand that if the FCC keeps Comcast, AT&T and other providers from doing what is necessary to move bandwidth from heavy users then all we are going to get is much higher prices. And or poor performance. What Comcast is attempting to do is come closer to guaranteeing quality of service for all customers.
BUT WE'RE IN THIS (SINKING) BOAT BECAUSE OF COMCAST (and other's) GREED!

Instead of building out their infrastructures to deal with what THEY KNEW WAS COMING, instead they went out shopping-buying cable systems in Los Angeles and elsewhere! They built this billion dollar HUGE phallic symbol of their greed in Philadelphia!

People in the cable business have known since the mid '80s that coaxial cable was a dead end technology. Over twenty years later, what's still on the poles? COAX!

Now their greedy decisions have come back to bite them in the ass, and they reply that the consumer will have to pay more to get LESS???!

AND YOU THINK THIS IS A-OK???????

Are you nuts?


espaeth
Digital Plumber
Premium,MVM
join:2001-04-21
Minneapolis, MN
·voip.ms
·Vitelity VOIP
·Callcentric
·VoiceStick
·ViaTalk
·Comcast
·Embarq

said by qworster See Profile :

BUT WE'RE IN THIS (SINKING) BOAT BECAUSE OF COMCAST (and other's) GREED!

Instead of building out their infrastructures to deal with what THEY KNEW WAS COMING
Last time I checked, Comcast doesn't own CableLabs -- the industry body that defines the DOCSIS protocols that determine what can be deployed on last-mile copper infrastructure. Comcast (and other MSOs) can also only buy product that exists; there isn't a checkbox on the equipment order that simply states "faster modems".

said by qworster See Profile :

They built this billion dollar HUGE phallic symbol of their greed in Philadelphia!
Liberty Property Trust built the tower, Comcast is just the principle leasing tenant so they get their name on the building.

said by qworster See Profile :

People in the cable business have known since the mid '80s that coaxial cable was a dead end technology. Over twenty years later, what's still on the poles? COAX!
Coax is far from dead. Hell, if you order Verizon service the other side of the ONT is coax delivered up to your access router.

For cable plants the islands of copper continue to shrink as nodes are getting moved deeper into the field with vast expanding fiber interconnects. To talk about the coax network of today as being the same coax network of the '80s is ridiculous.

Of course, facts don't help your rants at all, so feel free to ignore all of this.


wifi4milez
Big Russ, 1918 to 2008. Rest in Peace

join:2004-08-07
New York, NY
·Verizon FIOS
·Sprint Mobile Broa..
·RoadRunner Cable
·BroadVoice

reply to qworster
said by qworster See Profile :

People in the cable business have known since the mid '80s that coaxial cable was a dead end technology. Over twenty years later, what's still on the poles? COAX!

Coax is a dead technology?? Where did you get that information, coax is more than adequate for current applications. It was certainly more than adequate 20 years ago, and will be used long into the future.
--
If history teaches us anything, it teaches that simple-minded appeasement or wishful thinking about our adversaries is folly.
-Ronald Reagan-


Pizz
Hi

join:2000-10-27
Astoria, NY
reply to espaeth
Look at the board of directors of CableLabs »www.cablelabs.com/about/board/

every single large MSO is on that board - they do control CableLabs.

qworster

join:2001-11-25
Los Angeles, CA
·DSL EXTREME
·Brand X Internet
·RoadRunner Cable
·Vonage


1 edit
reply to espaeth
said by espaeth See Profile :

said by qworster See Profile :

BUT WE'RE IN THIS (SINKING) BOAT BECAUSE OF COMCAST (and other's) GREED!

Instead of building out their infrastructures to deal with what THEY KNEW WAS COMING
Last time I checked, Comcast doesn't own CableLabs -- the industry body that defines the DOCSIS protocols that determine what can be deployed on last-mile copper infrastructure. Comcast (and other MSOs) can also only buy product that exists; there isn't a checkbox on the equipment order that simply states "faster modems".

DOCSIS is a joke-a set of protocols designed to squeeze every last drop out of a dying technology. Coax is dead-if you don't believe me, then ask Verizon!

said by d://522717 :

They built this billion dollar HUGE phallic symbol of their greed in Philadelphia!
Liberty Property Trust built the tower, Comcast is just the principle leasing tenant so they get their name on the building.

Yeah. With a monitor wall that cost the better part of 2 million dollar$!

said by qworster See Profile :

People in the cable business have known since the mid '80s that coaxial cable was a dead end technology. Over twenty years later, what's still on the poles? COAX!
Coax is far from dead. Hell, if you order Verizon service the other side of the ONT is coax delivered up to your access router.

For cable plants the islands of copper continue to shrink as nodes are getting moved deeper into the field with vast expanding fiber interconnects. To talk about the coax network of today as being the same coax network of the '80s is ridiculous.

Of course, facts don't help your rants at all, so feel free to ignore all of this.
The cable infrastructure of today was designed in the early 1970s. Even your beloved DOCSIS 3 can't get you uploads faster then about 4 mbit. The problem is that there simply isn't enough upstream bandwidth in the coax infrastructure.

Coax is fine-for feeding one or two customers (the way Verizon does).BUT..the cable people still use coax to feed hundreds or thousands of customers per node. For that, it's pitifully obsolete!

I have been a TV/Radio/cable engineer for over 30 years. I KNOW the facts-probably better then YOU do!


funchords
Hello
Premium,MVM
join:2001-03-11
Washington, DC
·Verizon Online DSL
·Skype

reply to jester121
said by jester121 See Profile :

quote:
"Deep Packet Inspection devices capable of detecting what applications end points are using are both new and intrusive. They do not perform a task of Network management, they are performing tasks of Session and Application management. These are both new and inappropriate roles for Internet Service Providers."
The first part is just plain wrong, upper layer filtering and control have been around for a long time.
Yes, but only in very, very limited fashion on the open Internet -- for example, port blocking of the vulnerable Microsoft Netbios ports. And when these were done, even these were controversial at the time, but they were also widely accepted.

said by jester121 See Profile :

The second is merely semantic masturbation -- the phrase network management doesn't mean that all higher levels of the OSI model are ignored, and Topolski knows it.
Keep reading, I said more than that. Technology enabling full-speed full-packet capture of all the data crossing a box has only been available for the past year or two. The advent of DPI does not mean that it is suddenly necessary in order to perform network management. We've had network management for years.

We've also had much smaller DPI boxes, and these have done session and application management, but only for the enterprise, not on the public Internet.

said by jester121 See Profile :

He'd make a good politician -- make some solid statements at first, get loosened up, and then shove your foot in your mouth trying to hit one out of the park, thus blowing your credibility.
No I wouldn't. For one, I learn. Two, I admit when I'm wrong. And secondly, I can do math.
--
Robb Topolski -= funchords.com =- Hillsboro, Oregon
HTTP is the new Bandwidth Hog...

qworster

join:2001-11-25
Los Angeles, CA
·DSL EXTREME
·Brand X Internet
·RoadRunner Cable
·Vonage


1 edit
reply to wifi4milez
said by wifi4milez See Profile :

said by qworster See Profile :

People in the cable business have known since the mid '80s that coaxial cable was a dead end technology. Over twenty years later, what's still on the poles? COAX!

Coax is a dead technology?? Where did you get that information, coax is more than adequate for current applications. It was certainly more than adequate 20 years ago, and will be used long into the future.
AGAIN I SAY....THE ONLY REASONABLE WAY TO GET LOTS OF BANDWIDTH FROM POINT A TO POINT B IS VIA FIBER!!!!

The cable TV plants of today ARE OBSOLETE!!

Coax is a DEAD END technology, while fiber's capabilities are only beginning to be exploited!

jester121
Premium
join:2003-08-09
Lake Zurich, IL
·surpasshosting
·ViaTalk

reply to funchords
Errr... Netbios port blocking occurs at layer 4, friend. You know, where all the UDP and TCP port number are used? Session and Application layers are 5 and 7, respectively. Which you of course know.

But if these were being blocked on "the open internet" (whatever that means) then why do I still see traffic on these ports in my firewall logs? The answer is, they aren't being blocked on "the open internet". Many ISPs choose to block them at the edge of their subscriber networks, or on their own transport networks, just as some block outbound 25/tcp to addresses other than their own SMTP hosts, or inbound 80/tcp for subscriber lines where they don't want people hosting webservers.

You're right, at the enterprise level we've had DPI to manage (i.e. block, inspect, or prioritize) traffic based on what's inside the packets for a long time. Why? Because processing power hadn't caught up with the speeds that ISPs deal with. At the provider level, I certainly don't think this capability is going to go away any time soon, so the focus should be on making sure they get it right.

I'm not that familiar with AIMD but once again your comment about "complete conjestion" sounds like semantic masturbation, not to mention nit-picking. When I'm using all 1 Mbps of my upload speed for P2P, anyone else in my house will most certainly experience degraded performance due to congestion, to the point that web pages may time out (just for example). I think we're all able to figure out what happens when a network is moving around all the traffic that it can move around. For the pea brains at the FCC gridlock is probably as good an analogy as any, even if it offends the sensibilities of geeks like us who know better.

Finally, your apparent view that capacity upgrades are simple and infinitely effective (and practical) don't help you make your case, unless you know of anyone giving out carrier grade routers for free, guaranteed unlimited peering for a set cost, and indentured technicians to manage the entire thing. ISPs and carriers have CFOs who need to make sure that they have money to pay bills, including salaries so the smart people can keep things running.

(Lest this sound like I'm against what you're talking about, or bashing you, I'm not. My main point is that the way to win this fight isn't to beat them over the head with engineering knowledge, and it's going to involve some give and take on both sides. At least I think that was my point, when I started this post about an hour ago).

jester121
Premium
join:2003-08-09
Lake Zurich, IL
·surpasshosting
·ViaTalk

reply to qworster
said by qworster See Profile :

Even your beloved DOCSIS 3 can't get you uploads faster then about 4 mbit. ....
I have been a TV/Radio/cable engineer for over 30 years. I KNOW the facts-probably better then YOU do!
You should have stuck with radio, because you don't know much about cable.

Imagine spending 30 years doing something and still not knowing anything about it?


Anon123

@comcast.net
reply to qworster
I think you need to re-read the DOCSIS specs. With a 64QAM channel you can get ~27mb/s. And channel bonding with D3.0 on the upstream should scale nicely.


espaeth
Digital Plumber
Premium,MVM
join:2001-04-21
Minneapolis, MN
·voip.ms
·Vitelity VOIP
·Callcentric
·VoiceStick
·ViaTalk
·Comcast
·Embarq

reply to funchords
said by funchords See Profile :

Keep reading, I said more than that. Technology enabling full-speed full-packet capture of all the data crossing a box has only been available for the past year or two.
That's very much not correct. I was using NBAR to classify random-port Outlook/Exchange traffic for QoS mapping purposes in 2000/2001 on the Cisco 7206VXR/NPE400 platform. Around the same time my colleagues that worked for the University of Minnesota were using Packeteer appliances to identify and throttle P2P traffic based on fixed signature matching.

This technology is far from new. Sandvine was just one of the first players to get to 10GigE interfaces on these appliances.


funchords
Hello
Premium,MVM
join:2001-03-11
Washington, DC
·Verizon Online DSL
·Skype

reply to jester121
said by jester121 See Profile :

Errr... Netbios port blocking occurs at layer 4, friend. You know, where all the UDP and TCP port number are used? Session and Application layers are 5 and 7, respectively. Which you of course know.
IP devices work at level 2/3. ISPs don't have to have layer 4 devices to do their job. Port blocking was the first use of DPI on the Internet. (although it wasn't full-capture, and some don't call this DPI they call it shallow which creates funny concepts like shallow deeper packet inspection).

Please stop trying to vilify me. I'd much rather just have a conversation. As often as you recognize what I know, I recognize that you know what you're talkign about.

But if these were being blocked on "the open internet" (whatever that means)
meaning not at an end point or within a private enterprise
then why do I still see traffic on these ports in my firewall logs? The answer is, they aren't being blocked on "the open internet". Many ISPs choose to block them at the edge of their subscriber networks, or on their own transport networks, just as some block outbound 25/tcp to addresses other than their own SMTP hosts, or inbound 80/tcp for subscriber lines where they don't want people hosting webservers.
The ISP's network is 'the open internet' -- at least how I have painted the picture of with my description. Your take is accurate as well.

You're right, at the enterprise level we've had DPI to manage (i.e. block, inspect, or prioritize) traffic based on what's inside the packets for a long time. Why? Because processing power hadn't caught up with the speeds that ISPs deal with. At the provider level, I certainly don't think this capability is going to go away any time soon, so the focus should be on making sure they get it right.

I'm not that familiar with AIMD but once again your comment about "complete conjestion" sounds like semantic masturbation, not to mention nit-picking. When I'm using all 1 Mbps of my upload speed for P2P, anyone else in my house will most certainly experience degraded performance due to congestion, to the point that web pages may time out (just for example). I think we're all able to figure out what happens when a network is moving around all the traffic that it can move around. For the pea brains at the FCC gridlock is probably as good an analogy as any, even if it offends the sensibilities of geeks like us who know better.
Except that I've met these guys (and have been privileged in doing so). They don't know better. When they hear "complete congestion" they draw up a mental picture of the only kind of traffic congestion that they understand -- the one involving cars. They see gridlock -- where nothing is moving.

AIMD prevents that from happening for applications that use TCP -- as all the p2p applications do.

NCTA and Insight were scaremongering.

Finally, your apparent view that capacity upgrades are simple and infinitely effective (and practical) don't help you make your case, unless you know of anyone giving out carrier grade routers for free, guaranteed unlimited peering for a set cost, and indentured technicians to manage the entire thing. ISPs and carriers have CFOs who need to make sure that they have money to pay bills, including salaries so the smart people can keep things running.
YIKES! Where did I say that? I didn't say that they were free.

It is expensive and labor intensive. The fact that it is so is why ISPs would rather use Sandvine and other DPI-based solutions to "shape" the future of the Internet. For every $1 spent on DPI, they think they save $10 on network upgrades. That's dumbing down the future -- without bandwidth, why invent the next great thing to bring it cheaper?

(Lest this sound like I'm against what you're talking about, or bashing you, I'm not. My main point is that the way to win this fight isn't to beat them over the head with engineering knowledge, and it's going to involve some give and take on both sides. At least I think that was my point, when I started this post about an hour ago).
Welcome aboard, man!
--
Robb Topolski -= funchords.com =- Hillsboro, Oregon
HTTP is the new Bandwidth Hog...


funchords
Hello
Premium,MVM
join:2001-03-11
Washington, DC
reply to espaeth
We're talking about the big ISP level, not the enterprise level.

jester121
Premium
join:2003-08-09
Lake Zurich, IL
·surpasshosting
·ViaTalk

reply to funchords
said by funchords See Profile :

Please stop trying to vilify me. I'd much rather just have a conversation. As often as you recognize what I know, I recognize that you know what you're talkign about.
Absolutely 100% not my intention -- but the thing is, you're choosing to tackle this debate from an extremely technical perspective, but some of your statements contain rather glaring technical errors which your opponents can use to discredit or dismiss you. Honestly when I read Karl's post with your quotes, I was basically ready to blow off your comments as uninformed just due to the inconsistencies. Glad I took the time to do a bit more research; lots of others won't do that.


espaeth
Digital Plumber
Premium,MVM
join:2001-04-21
Minneapolis, MN
·voip.ms
·Vitelity VOIP
·Callcentric
·VoiceStick
·ViaTalk
·Comcast
·Embarq

reply to funchords
said by funchords See Profile :

We're talking about the big ISP level, not the enterprise level.
I consider Resnet on a major college campus to be much closer to a "big ISP" than an enterprise.

Again, Packeteer was doing all of this stuff *in ISPs* nearly a decade ago.


funchords
Hello
Premium,MVM
join:2001-03-11
Washington, DC
reply to jester121
I'm glad you did too, Jester! --Robb

Espaeth, that doesn't match my expectation -- what am I missing? So what happened some 18-24 months ago that brought in this sudden flood of full-capture DPI devices into the ISP space?
-
Forums » Don't Fear The Bandwidth ApocalypseIs it that hard »
« lol...  


Wednesday, 09-Dec 23:08:41 Terms of Use | Privacy Policy | Hosting by www.nac.net - DSL,Hosting & Co-lo | feedback | contact
over 10 years online! © 1999-2009 dslreports.com.
page compression OFF
Most commented news this week
· [200] Sprint Sued For Distracted Driving Death
· [107] AT&T Launching New 24 Mbps U-Verse Tier
· [82] 3G Network Test Says AT&T Is Tops
· [72] Mediacom Unveils 105 Mbps Pricing
· [66] Sprint Poised For A Turnaround?
· [64] WPA Cracker: Test WPA-PSK Networks In 20 Minutes
· [59] AT&T Hints At Usage-Based iPhone Data Pricing
· [51] The Future Of Wi-Fi Is Bright
· [47] Site Leaks Yahoo, Verizon Fed Data Share Pricing
· [44] Microwaving Your Innards Is Not 'Extreme'
Most people now reading
· Adobe Flash Player version 10.0.42.34 [Security]
· ICC Strats??? [World of Warcraft]
· ICC strats [World of Warcraft]
· Hot Girl falls face first down stairs [56k Lookout (Broadband Heavy)]
· Is sleeping similar to being dead? [General Questions]
· Windows 7 boot manager editing questions [Microsoft Help]
· RG Firmware update to VDSL2 this morning [AT&T U-verse]
· Connecting to Google Voice Via SIP [VOIP Tech Chat]
· The aftermath [World of Warcraft]
· Cross Server Dungeon Experience [World of Warcraft]