My commentary is reserved at this point... but, it does make me shudder.
Before you complain... It did not require a subscription when I first saw it.
Comcast has been advertising in press releases it gives priority to its
voice traffic over its network for a while.
http://www.prnewswire.com/cgi-bin/stories.pl?ACCT=104&STORY=/www/story/12-12-2005/0004231957&EDATE=
Unlike traditional Voice over Internet Protocol (VoIP) offerings that
run on the public Internet, Comcast Digital Voice calls originate and
travel over Comcast's advanced, proprietary managed network. Because
Comcast Digital Voice is a managed service, Comcast can make sure that
customer calls get priority handling.
If you install a Vonage terminal adapter/router, Vonage gives priority to
its voice packets over other traffic over a broadband connection.
When the various services were separate, there wasn't an issue. DSL data
and voice service use different frequencies over the same copper pair.
Which meant DSL data bandwidth was limited because the voice frequencies
were always reserved for the voice channel.
Now that the networks are converging, how do you provide traditional
levels of reliability to the different services sharing the same network?
Do you want the picture on the TV to stop because you download a big file
on your PC? Do you want to be able to make phone calls when your PC is
infected with Blaster and consuming your Internet bandwidth?
You coaxial cable can support a Gigabit or more of bandwidth, but the
cable company only sells you a few Megabits for Internet traffic. The
cable company keeps the rest of the bandwidth for other services it sells
such as video and voice. The service providers will probably sell you
a few Megabits of Internet bandwidth on your Coax/FTTH/DSL line, and use
the rest of the bandwidth on the line for other services like video or
voice. They may sell the use of the "extra" bandwidth above the level
you bought for Internet service to other companies. You may have only
bought 5 Mbps service from your cable company, but the cable company may
sell a burstable service to a Video On Demand company which lets them
download movies at 30 Mbps above your normal bandwidth cap. However,
the VOD company may have only scavenger class bandwidth, which means if
you are using the cable bandwidth for something else the VOD download
won't interfere with it.
Does Google treat ICMP packets equally as web packets?
http://www.prnewswire.com/cgi-bin/stories.pl?ACCT=104&STORY=/www/story/12-12-2005/0004231957&EDATE=
Unlike traditional Voice over Internet Protocol (VoIP) offerings that
run on the public Internet, Comcast Digital Voice calls originate and
travel over Comcast's advanced, proprietary managed network. Because
Comcast Digital Voice is a managed service, Comcast can make sure that
customer calls get priority handling.
Comcast doesn't have good public Internet access? That's a shame. I commend their bravery in admitting that only their internal network is advanced.
</sarcasm>
John
Sean,
I think you are skirting the real issue here.
Prioritizing traffic in order to provide reliable transport for isochronous
services is one thing; Using QoS features to de-prioritize traffic from a
competitor or a company who refuses to pay to access your customers is
something completely different.
These are not just paranoid ravings from the tin-foil brigades: two telecom
CEO's have recently floated trial balloons proposing exactly this scenario.
What good is 6Mbit DSL from my ISP (say, SBC for example) if only a small
portion of the net (sites that pay for non-degraded access) loads at a
reasonable speed and everything else sucks?
Joe
There are two possible ways of having a tiered system
- one is to degrade competitors/those who don't pay,
and the other is to offer a premium service to those
who do pay.
Would your perception of those two scenarios be
identical?
-David
-Fully RFC 1925 Compliant-
(speaking only for myself, btw...)
David Barak
Need Geek Rock? Try The Franchise:
http://www.listentothefranchise.com
What good is 6Mbit DSL from my ISP (say, SBC for example) if only a small
portion of the net (sites that pay for non-degraded access) loads at a
reasonable speed and everything else sucks?
One might argue that in such a situation, the end user is getting less value than they
did previously. End users might then either demand a price break or might vote with
their connectivity.
Tony
the last 2 times this has come up I think there was the suggestion that
given other options at reasonably close to the same end cost users might
switch to alternate access methods. That works as long as there are
alternate access methods, and as long as the telecom's don't 'cabal' and
all do the same hideously bad thing...
I do think it'd be funny for SBC or BS to do this sort of thing and get
massive customer loss when their customers defect to cable modem networks.
I know I would.
Regards
Marshall
One might argue that in such a situation, the end user is getting
less value than they
did previously. End users might then either demand a price break or
might vote with
their connectivity.the last 2 times this has come up I think there was the suggestion that
given other options at reasonably close to the same end cost users might
switch to alternate access methods. That works as long as there are
alternate access methods, and as long as the telecom's don't 'cabal' and
all do the same hideously bad thing...
There are a few things that this trend would get involved with.
1) It pushes the "cost" of "peering" to the content providers, essentially bypassing the underlying upstream/transit networks. The upstream/transit networks that are essentially getting disenfranchised might react by not peering with "premium" networks that are trying to pull their customers from using their network.
2) The only way this scenario (prioritization) makes any difference is when there isn't sufficient capacity within the "premium" network. If there is sufficient capacity, this is no real issue. However, for example, assuming this were enabled today, a network would have no incentive whatsoever to upgrade its networks -- provided that the customer pain/deprioritized network traffic is low enough. (A ratio that can be experimentally determined).
In the example where end users get 6Mb/s for $50/month. It is conceivable that as part of this "upgrade/premium" service for end users... they'd get 60Mb/s downstream for $50/month. The network could provide this service at no increased operational cost because it only expects to push (whatever they currently push) of deprioritized traffic. (say 6mb/s assuming no over subscription).
They could then cover the costs (and profits) of this 60mb/s premium service through the fees of the so-called premium content pushers. And thus, they could make the argument that no one is being harmed and in fact the end users gain....
Except that as the non-premium traffic levels of their end users grows... they suffer. The network's answer? Pay for premium access aka paid transit aka level-2 peering..
3) The good news is that the RBOCs haven't learned how to run IP networks cost-effectively. Their costs of implementing this so-called "tier-2" network will far exceed the fees their model tells them they will get from it. Remember when they first got into the Internet Access business? They all tried to create their own premium content-portals, search engines, what-have-you. Then they outsourced/sold that function to networks like MSN. I doubt the majority of their users even care. AOL tried to keep their network proprietary. Didn't keep them swimming either.
They can do anything they want with their own bits on their own network (eventually, the FCC will concede). The problem is that anytime you deprioritize the traffic of others for no other reason than because it isn't yours... well that smells a lot like restraint of trade. The RBOCs become gate keepers not underlying bit pushers.
What about the censorship issues? If they won't accept *insert bad site here [porn, hate, etc]* /premium/ network's payments to push traffic across their network, but they will accept it (as they currently do) on a deprioritized level??
Its a mess and they'll spend billions and it will cause some pain, and it'll eventually be abandoned. [My prediction based on what little is known about this thing today and history].
DJ
Thus spake "Christopher L. Morrow" <christopher.morrow@mci.com>
> What good is 6Mbit DSL from my ISP (say, SBC for example) if
> only a small portion of the net (sites that pay for non-degraded
> access) loads at a reasonable speed and everything else sucks?One might argue that in such a situation, the end user is getting
less value than they did previously. End users might then either
demand a price break or might vote with their connectivity.the last 2 times this has come up I think there was the suggestion
that given other options at reasonably close to the same end cost
users might switch to alternate access methods. That works as
long as there are alternate access methods, and as long as the
telecom's don't 'cabal' and all do the same hideously bad thing...
Congress appears to be working hard to make sure that happens.
I do think it'd be funny for SBC or BS to do this sort of thing and get
massive customer loss when their customers defect to cable
modem networks.
Do you really think the cablecos will be significantly less evil than the telcos? I'm not as optimistic about the result of a legislated duopoly.
S
Stephen Sprunk "God does not play dice." --Albert Einstein
CCIE #3723 "God is an inveterate gambler, and He throws the
K5SSS dice at every possible opportunity." --Stephen Hawking
Thus spake "Christopher L. Morrow" <christopher.morrow@mci.com>
> users might switch to alternate access methods. That works as
> long as there are alternate access methods, and as long as the
> telecom's don't 'cabal' and all do the same hideously bad thing...Congress appears to be working hard to make sure that happens.
I hope that congress doesn't allow all the telcos to decide to do the same
bad thing at the same bad time... I suppose they might though 
They've
been known to do some stupid things with respect to 'Internet' stuff.
> I do think it'd be funny for SBC or BS to do this sort of thing and get
> massive customer loss when their customers defect to cable
> modem networks.Do you really think the cablecos will be significantly less evil than the
telcos? I'm not as optimistic about the result of a legislated duopoly.
So far they seem to be not quite so evil (minus their port blocking for
some services, and rate-shaping for other services)... I used them as an
example though, really so long as there is another game in town
(competition) think the SBC/BS proposals will not last very long. At the
very least I'd bet that they won't garner the profits that the SBC/BS
execs are hoping will arrive.
Now that the networks are converging, how do you provide traditional
levels of reliability to the different services sharing the same
network?
Do you want the picture on the TV to stop because you download a big
file
on your PC? Do you want to be able to make phone calls when your PC is
infected with Blaster and consuming your Internet bandwidth?
Simple. You give the consumer the ability to fiddle with
the QoS settings on the provider's edge router interface.
After all, they are paying for the access link.
--Michael Dillon
There are two possible ways of having a tiered system
- one is to degrade competitors/those who don't pay,
and the other is to offer a premium service to those
who do pay.
The only way I know of to offer a premium service
on the same network as a non-premium service is
to delay non-premium packets. This artificial packet
delay is known as "Quality of Service" or QoS because
it degrades the quality of service to some users in
order to allow other users unobstructed use of the
network.
You see the same thing in road networks when the police
block certain intersections to allow a parade or an
important diplomat to move along the streets with no
obstructions. This type of policing can also be used
in networks.
But there is another way. If you provide enough bandwidth
so that your peak traffic levels can travel through the
network without ever being buffered at any of the core
network interfaces, then everybody is a king. If you charge
your customers a higher fee for such a network than your
competitors do, then we have a tiered Internet. This
unobstructed network was pioneered by Sprint on it's
zero-CIR frame relay network and they carried this forward
into their IP network as well. Other companies have
carried forward this architecture as well.
--Michael Dillon
This
unobstructed network was pioneered by Sprint on it's zero-CIR
frame relay network and they carried this forward into their
IP network as well. Other companies have carried forward this
architecture as well.
If I understand you correctly I highly doubt this is the case. If every
customer suddenly was to use the maximum link speed of their
access pipes I would be very surprised if all the traffic would be carried.
On the subject of tiered Internet you could argue that mobile/celluar
access to the Internet is another tier.
Regards,
Neil
That's the way all serious providers did IP-backbone engineering when
there was no QoS. Local congestion in the access-network would happen
from time to time even back in the 90s, but a network with
congestion-problems in the backbone would soon be a network with no
customers. Even today, it's the superior principle for backbone
engineering. Most QoS-handling (and other traffic-engineering) gizmos,
although some look good on paper, are too complex and too
labour-intensive to offer cost-saving or other operational advantage in
large IP backbones. Bandwith in the form of long-haul dark-fiber or
colors would have to be much more expensive to change that equation.
//per
All providers in your market would have to agree to do the same thing.
Capped services only work for monopoly providers.
//per
eeek! I assume you mean "tell the customer what
DSCP/whatever settings you honor, and let them do the
marking" right? The thought of letting customers
actually make changes to my edge routers would keep me
up at night...
-David
David Barak
Need Geek Rock? Try The Franchise:
http://www.listentothefranchise.com
To let customers decide priorities in your backbone is a bad idea, but I
don't think that's the issue here. Assuming the customer's link to the
network to be the primary bottleneck; there's nothing wrong with giving
customers the ability to prioritise traffic on their link, provided that
your access-equipment is able to handle queueing etc (given fool-proof
mechanisms that enable self-service and keep your NOC out of the loop of
course;).
//per
To me, this seems likely to lead to massive consumer dissatisfaction, and a disaster of the
magnitude of the recent Sony CD root exploit fiasco.
Typical Pareto distribution models for usage mean that no matter
how popular "tier 1" sites are, a substantial part of the user time will be spent on degraded "tier 2" sites.
If these don't work, people will complain. Just imagine for a second that cable providers started
a service that meant that every channel not owned by, say, Disney, had a bad picture and sound. Would this
be good for the cable companies ? Would their customers be happy ?
Of course, based on some recent experience this probably means that this will be adopted enthusiastically.
Regards
Marshall Eubanks