It seems you are making some false assertions.
1) "If you were a Comcast customer attempting to stream Netflix via this
connection, the movie would be completely unwatchable."
This is a false conclusion. Bandwidth is not allocated in static blocks on a
first come first serve basis. It is shared across all users. So adding a
single user only slows the speed of all connections proportionally. In a
pool roughly 16 million customers (across all regions) a single new
connection would not noticeably effect others.
2) "Comcast claims that a good network maintains a 1:1 "
I have never heard them assert that. I have heard them assert that they have
peering agreements with other providers. Those agreements assert that if the
bandwidth ratio remains the same, or close that neither party will charge
the other. For a end user network like comcast, that will never be 1:1. For
a larger network connecting to Level3 or TaTa, that might be 1:1.
Having a peering agreement does not in any way imply a ratio.
3) You assert that the bandwidth is capped and if Comcast purchased more
bandwidth it would not hit the cap.
With 16 million customers and ~20mbit connections average they would need a
320 terabit connection to ensure that they never hit the cap.
The reality is that most customers do not make uncapped connections. File
servers cap bandwidth per user and certain services, like gaming or
streaming media have a maximum rate. As long as the average data rate
allocated per customer is close to the usage then customers will not notice
the difference. Does it matter if it takes 10 seconds or 15 seconds to
download a 5 minute youtube clip?
Could Comcast purchase more bandwidth and speed up a percentage of their
users? Probably yes.
Would it drive up the cost of monthly internet? Yes.
Is your Comcast internet connection to slow to perform reasonable tasks at a
decent rate? Mine is not.
I am not asserting that Comcast has enough bandwidth, just that some of your
assertions are not valid.
1) "If you were a Comcast customer attempting to stream Netflix via this
connection, the movie would be completely unwatchable."
This is a false conclusion. Bandwidth is not allocated in static blocks on a
first come first serve basis. It is shared across all users. So adding a
single user only slows the speed of all connections proportionally. In a
pool roughly 16 million customers (across all regions) a single new
connection would not noticeably effect others.
When the pipes are full, and the latency is up closer to 1s vs a more reasonable 20-100ms, and you have a few percent packet loss, do movies continue to stream? Where are they streaming from? Over the TATA transit, a non-congested Level3 pipe, or on-net CDN servers?
2) "Comcast claims that a good network maintains a 1:1 "
I have never heard them assert that. I have heard them assert that they have
peering agreements with other providers. Those agreements assert that if the
bandwidth ratio remains the same, or close that neither party will charge
the other. For a end user network like comcast, that will never be 1:1. For
a larger network connecting to Level3 or TaTa, that might be 1:1.
Ratios only make sense between peers. When you're buying transit, you don't get to enforce ratios and tell your transit providers you're not going to pay (or they're going to pay you) because they're sending you too much traffic. Back when I ran a dialup network, and our traffic profile was maybe 5:1 input vs output, UUnet would have laughed at me and shut us off if I told them "you're sending us traffic at a 5:1 ratio...because you're sending us so much more traffic than we send you, you're going to have to pay us to deliver that traffic." Comcast can only get away with that because of their monopoly (captive userbase) and size.
3) You assert that the bandwidth is capped and if Comcast purchased more
bandwidth it would not hit the cap.
With 16 million customers and ~20mbit connections average they would need a
320 terabit connection to ensure that they never hit the cap.
That depends on your definition of 'never'. You can oversell your network capacity...everyone does...and not run with the pipes full 99% or better of the time. A responsible network keeps track of bandwidth trends and plans capacity upgrades as needed. Comcast is allegedly intentionally running their transit at or beyond capacity (and removing bandwith capacity instead of adding it, to keep their transit full) and then bullying content providers into buying access to the Comcast network to avoid the congested transit links.
At max capacity, we'd run roughly double our total transit capacity, yet we rarely exceed 70% of N+1 bandwidth before upgrading. Perhaps my model doesn't scale when people utilize multiple 10G, though, as I know I have several burstable 10G available for special events and such as a just in case option. The price of 100G interfaces currently is still too high for someone to actually want a burstable 100G circuit with only a 30 or 50 gig commit.
Jack
The problem starts when that the choke point is congested enough that the
question isn't "10 seconds or 15", it's "4 mins 30 or 5 mins 30 for that 5
minute clip". Buffer underruns are incredibly annoying.
Or, from personal experience:
The movie stops because the buffer was exhausted, Netflix informs you "Your network connection has changed", shows a progress bar while it buffers /at a lower bitrate/.
Then you get to watch the rest of the movie like it was 1995.
Jon,
<snip>
Ratios only make sense between peers. When you're buying transit, you
don't get to enforce ratios and tell your transit providers you're not
going to pay (or they're going to pay you) because they're sending you too
much traffic. Back when I ran a dialup network, and our traffic profile
was maybe 5:1 input vs output, UUnet would have laughed at me and shut us
off if I told them "you're sending us traffic at a 5:1 ratio...because
you're sending us so much more traffic than we send you, you're going to
have to pay us to deliver that traffic." Comcast can only get away with
that because of their monopoly (captive userbase) and size.
I agree with most of your reply. In regards to ratios, I firmly believe
they don't make sense between peers as well. When you operate an eyeball
network, you need
to expect that your users are pulling the data from the content providers.
Your ratios will always be unbalanced with most of your peers.
If ratios are really a concern and you really need to maximize your port
capacity, there are ways to balance this; balance your customer base. Start
hosting content. Now, this might not help on private peering interconnects,
but if you peer publicly, this will help you balance (and take advantage of)
your public peering capacity.
Either way, ratios are very 1990s. As has been mentioned before, it was
more of an excuse not to peer by the monopolistic entities that made up a
big portion of the Internet in the 90s and isn't very relevant today.
Randy
Bandwidth is not allocated in static blocks on a
first come first serve basis. It is shared across all users. ...
a single new connection would not noticeably effect others.
I love how people demonstrate how they've failed most of the math classes in their life.
Let's start with a 10G link (10,000M). If a live (real-time) video stream needs a minimum of 5Mbps, then the link can support a maximum of 2000 streams. Add one more stream and you will not have the bandwidth to support the required rate for all of them. In a perfectly fair system, everyone's experience begins to be degraded; *every* additional user robs an incremental amount from all the others. With Comcast's 16mil users, it's a safe bet that tens of thousands of them are streaming at any given point. Why do you think Level3 asked for ~30 10G ports?
(I know netflix streams are less than 5M.)
2) "Comcast claims that a good network maintains a 1:1 "
I have never heard them assert that.
Read their blog posts. Read the peering agreement.
Would it drive up the cost of monthly internet? Yes.
Of course it would. But not because comcast would go broke doing it. They make plenty already. They aren't interested in doing anything that doesn't immediately *increase* their profits. (which is making content sources pay them to get to their millions of customers.)
--Ricky
Jon,
If ratios are really a concern and you really need to maximize your port
capacity, there are ways to balance this; balance your customer base. Start
hosting content. Now, this might not help on private peering interconnects,
but if you peer publicly, this will help you balance (and take advantage of)
your public peering capacity.
To that point, comcast does sell wholesale ip transit and they have a
quite a few unused timeslots for packets in the egress direction...
Ratios shouldn't matter in sense that unblanced ratios do not imply that
mutual benfit is not being derived from the interconnection. If for
example, I get access to your customer's and you no longer have to pay a
transit provider for my transit then we both win even if the flow is
virtually all inbound.
1) Sure, if those streams are only video streams and they can only exist at
5mbps. In reality, a network of 16 million users has lots of types of
streams and some like file downloads, UDP data for game players, video with
user buffers, etc, are capable of getting squeezed a little.
It seemed that the OP was stating that once the network was at capacity, all
further users got zero bandwidth.
Also, adding 1 user to 16 million is very different then adding 1 user to a
pool of 10, 20, or 2000. It is a LOT harder to measure the impact.
2) Fair enough, I have not read either of those. Again, I was not arguing
the point, simply pointing out that you can't imply that all relationships
are 1:1 without some backup or proof.
3) That describes every for profit business on the planet. We pay comcast
more then we should for a service, just like buying a burger.