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;).
Precisely!
In today's world, lots of router configuration is not
done manually by anybody. There is an OSS system that
applies rules to what changes will and will not be
done and when they will be done.
Since QoS works by degrading the quality of service
for some streams of packets in a congestion scenario
and since congestion scenarios are most common on
end customer links, it makes sense to let the end
customers fiddle with the QoS settings in both
directions on their link.
Of course, any incoming packet markings should be
discarded or ignored once the packets pass the
provider's edge router.
This is possible today without any special support
from router vendors. It relies entirely on operational
support systems such as web servers, databases and
remote control servers.
Actually, the cable providers have an alternative. Since the cable network really is "broadband" in the meaning from before it was coopted to mean "high speed", cable operators are able to utilize many channels in parallel. If they want their voice traffic to be unimpeded, they could certainly pick up an IP address on a private network space on a different cable channel (i.e. frequency pair) and make use of that. The consumer's Internet service, being on other channels, is unaffected. Yes, the backhaul fiber network would need to be using multiple paths as well to make that work. I have no idea to what extent present cable plants make use of the ability to use multiple channels for data service. Clearly they use it for video carriers, and where there is/was telephone over cable before the present VOIP-based offerings, those also appear to have used separate channels.
So, there is a method possible other than packet prioritization. Just tossing a fatter pipe at the customer isn't a solution, however. It'd still get clogged with p2p traffic pushing pirated music and videos among residential users.
So where would the payback be for this for the last-mile provider? Compared to the pain of setting this up and supporting it, what percentage of customers would actually use something like this? Just trying to educate users on this would be quite challenging. "Well, sir, the service allows you to select which of your traffic is important and should get priority..." "But all my traffic is important!"
It gets more fun when the medium you use to get to the end customer is a shared medium, with some normal amount of oversubscription.
since Internet is "best-effort" ... any overt attempt
to reduce this best effort service to explictly degraded
service (perhaps due to intentional overprovisioning, causing
degraded service) ... -is NOT the Internet- ... its some
propriatary, substandard networking technology to get me
to the Internet. So i suspect that marketing folks be very
clear on what is being sold.
Actually, the cable providers have an alternative. Since the cable network really is "broadband" in the meaning from before it was coopted to mean "high speed", cable operators are able to utilize many channels in parallel. If they want their voice traffic to be unimpeded, they could certainly pick up an IP address on a private network space on a different cable channel (i.e. frequency pair) and make use of that. The consumer's Internet service, being on other channels, is unaffected. Yes, the backhaul fiber network would need to be using multiple paths as well to make that work. I have no idea to what extent present cable plants make use of the ability to use multiple channels for data service. Clearly they use it for video carriers, and where there is/was telephone over cable before the present VOIP-based offerings, those also appear to have used separate channels.
Allocating those separate channels for different services means that that bandwidth blocks they consume are off-limits to provide customer IP service. Would it be better to have a smaller amount of bandwidth that's isolated from all other services for normal customer IP service, or would it be better to have a bigger pipe with priority when there's congestion going to services other than normal customer IP service?
The answer depends on how much traffic you expect to be prioritized. VoIP traffic at 80kbps probably isn't going to be a huge concern. Tiered services could be, but seperate channels could actually make the problem worse, since bandwidth that had been allocated to the standard services could be permanently allocated to the higher-tiered service to resolve peak load issues, reducing the bandwidth available to the standard service at all times.
How do you know your MSO or ISP hasn't been doing this for years?
Do ISPs have different levels of congestion on peering links between
different networks?
Do ISPs have different levels of congestion between peering links and
internal links?
Do ISPs use different circuits or queues for different types of traffic?
Whether your use time division multiplexing, frequency division
multiplexing or packet division multiplexing, the effect on congestion is
similar. You are creating multiple queues.
You don't "delay" or "slow down" packets. Routers don't have big enough
buffers to "delay" a packet for milliseconds. Whether you have two queues
on one physical interface or two physical interfaces with one queue each,
the end effect is the same if you forward different traffic through
different paths.
The reality is converged or partially converged networks have been using
some level of QOS at some layer (MPLS, IP, ATM) for years. CIR, PIR, UBR,
CBR, DSCP, TOS, choose your style. Even the PSTN has multiple classes of
service for different calls ranging from choke numbers used for radio
stations and call-in contests to GETS numbers used by emergency responders.
Yes, MSOs and other ISPs are already doing this. The primary difference
seems to be the telco's discuss more of their network engineering practices
in public.
I'm seeing a lot of comments here that appear to be looking at this as a very binary issue -- either it's ok, or it will cause the customers to defect en masse to the competition. This seems to ignore questions of how it would be implemented, and what the competition's offering would be.
If I've got a choice between two providers, both of which are offering a 3 Mb/s pipe, but one of them restricts services from other networks to half of that pipe, that's going to effectively be a situation where one provider is only offering half the Internet bandwidth the other offers.
On the other hand, there could be a scenario in which one network offered a 3 Mb/s unrestricted pipe, while the other offered a 6 Mb/s pipe, with prioritized traffic potentially eating 2 Mb/s of it. That would still be 4 Mb/s of unrestricted traffic vs. the other provider's 3 Mb/s.
In other words, a provider with sufficiently better last mile technology than the competition should be able to do lots of stuff like this and still come out ahead. Providers in markets that are technologically more even might have more trouble.
That assumes rate limiting in the last mile. If what's instead being talked about is QoS tagging of last mile packets, that should be completely irrelevant to those who don't use the services that are prioritized.
Of course, if they're restricting capacity in the backbone and using QoS there, that may be a different story, but that seems unlikely to be what's being talked about. Backbone congestion doesn't tend to happen much in major American cities these days, but individual DSL lines saturate pretty easily.
*IF* they have a choice. In many areas for consumer grade access, you don't. I fully agree that you're not getting the same value/.worth out of a service that behaves like that. The strategy they're proposiing is very anti-competitive and very monopolistic.
You start with a flawed assumption, you end up with wrong conclusions.
Who said this had anything to do with "the Internet"?
Instead, this is about additional private network services, which cable
companies already do over coax, that telco's want to offer over a
multiservice access line in addition to "the Internet." Coax can carry
over a Gigibit of data, but cable companies usually sell user's less
than 10Mbps for Internet data. Cable companies reserve the rest of
the their network capacity for private services like HBO, video on
demand and voice. Just because part of a physical line is used for
Internet service doesn't mean everything going across the same line
is the Internet.
The telephone companies are asking for the same ability to sell multiple
services over the same physical line. Cable companies didn't make their
Internet service slower when they add more private services, why do
people expect the telephone companies to make their Internet service
worse when the telephone companies add private services to their network?
The telephone companies are asking for the same ability to sell multiple
services over the same physical line. Cable companies didn't make their
Internet service slower when they add more private services, why do
people expect the telephone companies to make their Internet service
worse when the telephone companies add private services to their network?
Because they're telephone companies.
Because they can't manufacture bandwidth that isn't there. Cable co's provide
broadband with a fraction of the loop capacity. For telco's to offer premium
service, they have to take from the aggregate capacity. It's a zero sum game,
and for the telco's to get more, the subscribers get less.
Because they can't manufacture bandwidth that isn't there. Cable
co's provide broadband with a fraction of the loop capacity. For
telco's to offer premium service, they have to take from the aggregate
capacity. It's a zero sum game, and for the telco's to get more, the
subscribers get less.
I guess you missed all those trenches being dug in Verizon land to install
fiber to the home. I guess you missed all the network upgrades in ATT/SBC
and Bellsouth land to shorten their copper loop distances.
Sounds like they are manufacturing more bandwidth and the zero sum game
is getting bigger.
I guess you missed all those trenches being dug in Verizon land to install
fiber to the home. I guess you missed all the network upgrades in ATT/SBC
and Bellsouth land to shorten their copper loop distances.
Sounds like they are manufacturing more bandwidth and the zero sum game
is getting bigger.
I believe it when it gets to my street. So far, the reality is Really Slow DSL, with
service and installation times measured in weeks at costs that aren't
competitive. So yes, I missed all of that.
I guess you missed all those trenches being dug in Verizon land to
install
fiber to the home. I guess you missed all the network upgrades in ATT/SBC
and Bellsouth land to shorten their copper loop distances.
Sounds like they are manufacturing more bandwidth and the zero sum game
is getting bigger.
I believe it when it gets to my street. So far, the reality is Really
Slow DSL, with
service and installation times measured in weeks at costs that aren't
competitive. So yes, I missed all of that.
And, at that, only after extracting regulatory concessions at both the
state and federal levels basically giving them their monopoly back to
give them "incentive" to half-*ssed roll out that DSL that is, itself, a
mere fraction of what is technically possible.
There are currently a couple of million IPTV users worldwide. Imagine how
much more useful the conversation would be if it included people who have
actually used it and could say what their experience has been instead of
people leaping to conlusions based on inaccurate reports.
No matter how many million IPTV users there
are, it's not reaching the area where i live. I'd love Verizon
to come into the chunk of the SBC area where i live
that is adjancent to their existing service area and attempt
to compete with each other.
Date: Wed, 14 Dec 2005 20:45:09 -0500
From: Jeff McAdams
And, at that, only after extracting regulatory concessions at both the
state and federal levels basically giving them their monopoly back to
give them "incentive" to half-*ssed roll out that DSL that is, itself, a
mere fraction of what is technically possible.
Hear, hear.
Interestingly, back in 1997, $local_ilec claimed they were "waiting on
the tariff to be approved" for lower ISDN rates. I suspect such a
tariff requires filing for any chance of approval.
General observation: Both cable and DSL are available, or neither are.
That's empirical; don't ask me for an r-squared calculation.
What I'm interested in is how the two service
providers will build a two tiered Internet.
To our experience, current QoS mechanism ( WRR +
multiple_Queue) could not differentiate service
quality when bandwidth is overprivisioned. If there is
congestion, why should I stay with it while there is
another ISP who says their is no congestion in their
network ? If hard limited bandwidth allocation
mechanism is available, how could they calculate the
bandwidth of each service class ? how could they do
with the complexity of nework management? How could
they do with security problems?
Looking at IPTV, I'm not sure where is millions of
people use such service; but I do know P2P IPTV
application (like ppstream) could provide good quality
and multiple TV programs even bandwidth is limited.
So, IMO this is game between ISPs, new technology,
content providers and internet users. Currently,
content providers are the ONLY winner.
