News for hackers. Stuff that matters.
Re:GOP net neutrality bill would allow paid fast lanes and p... by wanderingbort on Wednesday December 20, 2017 @09:21AM (#15969675)
That is objectively false. Degrading the "slow" lanes is only one way to produce a fast lane. Direct peering arrangements and co-location of services are two ways to create "fast" lanes without having to degrade anything. They are also not free, so it makes sense that business-to-business deals would be made to establish them (aka paid fast lanes).

The effective bandwidth of the internet looks like an hour glass. As things fan out to individual users you may have capacity for 100+mbps but no way to consume it because upstream there is a bottleneck. In many cases that bottleneck exists at the edge of an ISP or farther upstream of it. Direct peering increases the capacity of the ISP's "edge" but only for the services available on the peer. Co-location puts those services on the internal side of the bottleneck, again increasing effective capacity, but only for the services that are co-located in the ISPs datacenters. Almost all major providers (google, akamai, netflix, facebook) offer some form of co-locatable server however, those cost money to deploy and maintain.

Re:GOP net neutrality bill would allow paid fast lanes and p... by dahart on Wednesday December 20, 2017 @09:47AM (#15969925)
You've missed the point. The very existence of fast lanes means that whatever's left is a degradation. This has been the point NN proponents have been making all along. Proposing to ban slow lanes but allow fast lanes effectively does nothing for NN proponents.

It doesn't have to be "actively" slowing down traffic to be degraded. If ISPs can go faster, and they choose not to unless you pay more, that is a form of degradation. One of many forms the degradation could come in is ISPs failing to build out faster networks until people are paying higher prices. Service can be allowed to degrade over time by simply not doing anything unless you pay more money.

Re:GOP net neutrality bill would allow paid fast lanes and p... by wanderingbort on Wednesday December 20, 2017 @10:15AM (#15970204)
> If ISPs can go faster, and they choose not to unless you pay more, that is a form of degradation.

You've missed my point. There are some ways an ISP can go faster that are not neutral AND not malicious/extortive. If google puts a GGC on your ISPs network, youtube now has an advantage over hulu. There is no equivalent internal-deployment that can give a truly neutral increase in capacity.

To conflate this with active degradation (throttling or blocking) means that you would have an ISP be unable to host any service until they can guarantee _all services_ equivalent access (presumably at the ISPs expense otherwise they would price out everyone but the big players).

Before you immediately chose "neutral hosting" realize that it is untenable for every random small website to be able to legally demand any ISPs host their services for free in order to guarantee a level playing field with Google and Netflix. The only tenable solution is to make it illegal to collocate services which will be a far worse hit to the experience you have on the web than any throttling or blocking has ever been.

This is not as simple as you would like it to be.

> If ISPs can go faster, and they choose not to unless you pay more, that is a form of degradation

This is assuming intent and over-simplifying. Is a small regional ISP without the capital or credit necessary to expand now anti-NN by default? Are they extorting their customers by default? Not all ISPs are Comcast and Verizon. If you cannot escape your circumstance and see the bigger picture of how and what the internet is and who "ISPs" are then you are just being selfish.

Re:GOP net neutrality bill would allow paid fast lanes and p... by wtallis on Wednesday December 20, 2017 @10:40AM (#15970451)
> There are some ways an ISP can go faster that are not neutral AND not malicious/extortive. If google puts a GGC on your ISPs network, youtube now has an advantage over hulu. There is no equivalent internal-deployment that can give a truly neutral increase in capacity.

We need laws or regulations to address the fact that ISPs are actually forcing content providers to deploy in-network CDN nodes with the ISPs through extortive refusal to upgrade the links between the residential ISPs and the ISPs serving Netflix et al. Some deployment of local CDN nodes makes sense on its own and benefits both parties, but it's not like video streaming is actually latency-sensitive or anything.

Re:GOP net neutrality bill would allow paid fast lanes and p... by notyourday on Wednesday December 20, 2017 @09:59AM (#15970046)
It is not.

If Netflix did not have PNIs to Comcast, Comcast internet would have been unusable for everyone.

[Edit: Post more downvotes]

Guess what happens to the Netflix traffic when the best route to Netflix is over Comcast transit?

Re:GOP net neutrality bill would allow paid fast lanes and p... by wtallis on Wednesday December 20, 2017 @10:27AM (#15970342)
> If Netflix did not have PNIs to Comcast, Comcast internet would have been unusable for everyone.

Only under the assumption that Comcast makes no attempt to do their job properly and uphold their contractual obligations by implementing reasonable QoS measures and upgrading their capacity. You can't point to Comcast's extortionist refusal to upgrade their links to the outside world and say that it's Netflix's fault; that was just their abusive tactic to force Netflix to deal with Comcast directly on Comcast's terms. Verizon did the same thing.

[dead]
Re:GOP net neutrality bill would allow paid fast lanes and p... by wtallis on Wednesday December 20, 2017 @10:44AM (#15970496)
> QoS does not stand for Quality of Service. It stands for Quantity of Service.

Fuck you. You aren't even trying to have a reasonable conversation, are you?

You're welcome to attempt to argue that there's an unavoidable equivalence between QoS as everybody defines it and the other concept you're trying to equate it to, but you have to actually make that argument, not just declare the definition of QoS to have changed because you don't like what it means. QoS still stands for Quality of Service.

> But hey, yes, NN is telling me I should at best use RED to do QoS as everything else requires lookng at protocol and destination

You can use CoDel without violating any principles of net neutrality. You can even use FQ-CoDel without violating the principles of net neutrality, because it only looks at port numbers to determine which packets are related to each other; each port is subjected to the same set of rules and no port is preferred over another. Net neutrality doesn't require you to ignore the past 20+ years of research into QoS.

It may help your political cause to regard QoS as being strictly a zero-sum game, but that doesn't make it true.

> There's not a single network that can provide this kind of capacity even at 10Mbit/sec, not to mention at 50Mbit/sec or 150Mbit/sec or Gigabit/sec.

> The only way to deal with this is PNIs between those that originate massive amounts of traffic and the networks that consume massive amounts of traffic.

The technical requirements of having large, reasonably direct links between big producers and big consumers does not say anything about whether Netflix has to negotiate with Comcast and Verizon directly, or whether they can outsource those concerns to Level 3. That's purely in the domain of business shenanigans.

Re:GOP net neutrality bill would allow paid fast lanes and p... by notyourday on Wednesday December 20, 2017 @11:06AM (#15970751)
> You're welcome to attempt to argue that there's an unavoidable equivalence between QoS as everybody defines it and the other concept you're trying to equate it to, but you have to actually make that argument, not just declare the definition of QoS to have changed because you don't like what it means. QoS still stands for Quality of Service.

QoS being a quantity of service is a very well known term, pioneered and popularized by Dave Rand/Avi Freedman to describe that QOS establishes which packets are to be dropped on the floor first. QOS might work OK on networks with microspikes but it completely breaks on networks that are > 20% oversubscribed.

If I were made to pick between positions of Dave Rand and Avi Freedman and positions of all people supporting Net Neutrality, I would go with the first two as those people actually built networks.

> Net neutrality doesn't require you to ignore the past 20+ years of research into QoS.

This only works within a single AS. It has been demonstrated when a pair of crazy dudes built a Sprintlink on a top of ANS without ANS' knowledge.

> The technical requirements of having large, reasonably direct links between big producers and big consumers does not say anything about whether Netflix has to negotiate with Comcast and Verizon directly, or whether they can outsource those concerns to Level 3. That's purely in the domain of business shenanigans.

It does. Without that you end up with Enron.

Re:GOP net neutrality bill would allow paid fast lanes and p... by wtallis on Wednesday December 20, 2017 @11:17AM (#15970872)
> QoS being a quantity of service is a very well known term, pioneered and popularized by Dave Rand/Avi Freedman to describe that QOS establishes which packets are to be dropped on the floor first.

So you're still trying to re-define the literal text of what the "QoS" initialism expands to, and now you're also trying to narrow the scope of QoS as defined by everybody else.

There's more to QoS than just deciding which packet to drop. You also have to decide when to drop a packet. This second aspect is what makes QoS not a zero-sum game when faced with real-world traffic.

> QOS might work OK on networks with microspikes but it completely breaks on networks that are > 20% oversubscribed.

What exactly do you mean by "microspikes"?

20% oversubscription is laughably small compared to the disparity between my 1GbE LAN and my 50Mbps WAN connection. QoS makes a world of difference for the users on my LAN. Oversubscription just means you'll actually get the opportunity to notice the difference between having QoS and having a dumb FIFO. Nobody's actually asking for a circuit-switched Internet, or anything that would require a circuit-switched Internet.

Re:GOP net neutrality bill would allow paid fast lanes and p... by notyourday on Wednesday December 20, 2017 @01:09PM (#15972126)
Ok, lets do basics.

You have 1Mbit/sec worth of traffic and the max connection speed is 10Mbit/sec. Packet loss is at 0% and latency is speed of light + speed of forwarding.

Do you need QoS in this case?

Re:GOP net neutrality bill would allow paid fast lanes and p... by AnthonyMouse on Wednesday December 20, 2017 @10:14AM (#15970189)
> Direct peering increases the capacity of the ISP's "edge" but only for the services available on the peer. Co-location puts those services on the internal side of the bottleneck, again increasing effective capacity, but only for the services that are co-located in the ISPs datacenters. Almost all major providers (google, akamai, netflix, facebook) offer some form of co-locatable server however, those cost money to deploy and maintain.

This is a textbook lobbyist argument -- something which is technically true without doing the math.

A 10Gbps port on a switch is maybe $100. Say it can service a hundred customers, so $1/customer, and has a lifetime of two years. So $0.04/customer/month.

These are very conservative numbers. In reality it doesn't even cost that much.

The fact that the ports go to different places is irrelevant in the aggregate, because there is no need to add a port when the existing links to that peer aren't at capacity. And it's not as though the fact that Netflix and Google have separate ports would allow a 100Mbps customer to get 100Mbps from each of them at the same time.

Colocation is independent of this. If anyone with traffic could get peering at cost (i.e. ~free), they might still do it or might not, but then the ISP would have to charge market rates for space and power instead of monopoly prices because they're really charging for gatekeeper access to customers.

Re:GOP net neutrality bill would allow paid fast lanes and p... by wanderingbort on Wednesday December 20, 2017 @10:24AM (#15970292)
you are over-simplifying. If we want to be pedantic,

and assume that the two edge routers are just switches

and we would use a consumer level 10gbps port

and that the switchers are already in the same rack in the same facility

and that that rack in that facility has direct access to our distribution PoP

and that there are no firewall rules that need to be in place on that port (they can be in the DMZ sure)

and that we don't need to monitor that connection

and that we don't need that connection to be redundant at all

and that we aren't going to use a routing protocol like BGP on that port (we will just hard code routes)

and that we don't want any guarantee that you and I will operate that link at any SLA

and that we don't want to guarantee that both parties will give notice in a timely manner before termination

and... and... and...

So, which textbook are you copying from? You put out something that is technically true without considering any of the practical realities.

(edited for readability)

Re:GOP net neutrality bill would allow paid fast lanes and p... by AnthonyMouse on Wednesday December 20, 2017 @11:01AM (#15970679)
You're just listing things that add small linear costs that are rounding errors in a ballpark estimate like this, and fixed costs that don't add anything because they're the same for two ports as twenty.

Do the actual math. Show me something with an aggregate cost of more than, say, $1/month/customer.

Especially using realistic numbers rather than the "10Gbps port for a hundred 100Mbps customers" ones I gave to make a point.

Re:GOP net neutrality bill would allow paid fast lanes and p... by wanderingbort on Wednesday December 20, 2017 @11:23AM (#15970930)
> Do the actual math.

You first. You made a claim which was woefully inaccurate, I called you out in an admittedly childish way and now the burden of proving you wrong is on me?

Here is my ballpark, Both sides will want redundant fiber connections between two edge routers at 10gbps. SFP+ 10g from cisco is ~60 bucks. We need at least 4 (2 on each side), plus a reserve of 4 for faster recovery from failure. In addition, lets assume that we amortize some cost of the _edge router_ for this service. A refurb 16 port SFP+ module for 10g fiber is around $6000 and we will need 2 of them in 2 separate chassis on EACH SIDE and lease Colo space for it but we will amortize it. So, $6k + $1k for the chassis + $800 for the supervisor card. to lease space for this: $1000k/month per chassis(5u at 200/u in a major market carrier hotel) / 16. We need cabling too but lets assume thats commodity.

up front costs so far:

$60 (sfp modules) * 8 = 480

$7800 edge router * 4 / 16 = 1950

so $2430 for equipement (amortized)

1k/month/chasis = 4k/month / 16 = 250/month recurring

100 customers over 2 years = 1950 + 6000 / 100 / 24 = $3.31/month/customer before we even talk about monitoring/maintenance/legal/etc

I have not even begun to dig into the practical costs, and I'm at 3x your target. Your move.

Re:GOP net neutrality bill would allow paid fast lanes and p... by AnthonyMouse on Wednesday December 20, 2017 @12:11PM (#15971472)
The ISP doesn't have costs on each side, the other side is the peer's equipment. And you're just intentionally setting money on fire. You can get more than 16 ports into five rack units, which would result in less amortization on the chassis and less colo per port. And 5U worth of colo is not $1000/month almost anywhere, but that was >75% of your total cost.

Which is still using the numbers where equipment only lasts two years and you're expecting 100% continuous usage from all customers. And even using all these ridiculous numbers it's still only $3.31/month per customer. If we adopted a 10:1 oversubscription rate (which is still conservative) it would be $0.33/month even with you setting money on fire, having the ISP pay for both sides, and using full double redundancy rather than some kind of N+1 or N+M.

Peering is just not a major cost for ISPs.

Re:GOP net neutrality bill would allow paid fast lanes and p... by wanderingbort on Wednesday December 20, 2017 @12:40PM (#15971818)
You are moving the goal posts. I put my estimate in the same context you put your estimate in (no over subscription, 100% continuous usage etc).

The ISP doesn't pay both sides, but someone pays somewhere. So, if we are talking about the "cost of peering" it doesn't serve to talk about it from only one side.

You can get more than 16 ports in the 5u unit, so rent is probably under-amortized but not by the two order of magnitudes that would make up the difference between my estimate and your original.

Peering is not a major cost for ISPs because ISPs don't have a lot of peers. As more services want direct peering relationships, there are some overhead costs which go up to manage the explosion of peering relationships.

My initial objection was that you over-simplified. I feel pretty fine resting on my argument as is. You may have the last word. I cede all my remaining time to whomever wants it.

Re:GOP net neutrality bill would allow paid fast lanes and p... by AnthonyMouse on Wednesday December 20, 2017 @01:40PM (#15972432)
> The ISP doesn't pay both sides, but someone pays somewhere. So, if we are talking about the "cost of peering" it doesn't serve to talk about it from only one side.

But we're talking about the cost of peering paid by the ISP, because the ISP doesn't have to cover costs paid by somebody else and no one else is objecting to paying their part of it.

> You can get more than 16 ports in the 5u unit, so rent is probably under-amortized but not by the two order of magnitudes that would make up the difference between my estimate and your original.

Juniper MX240 supports 80 10GbE ports in 5U. Cisco's product search is lame but I expect they have something similar too.

> Peering is not a major cost for ISPs because ISPs don't have a lot of peers. As more services want direct peering relationships, there are some overhead costs which go up to manage the explosion of peering relationships.

That isn't necessarily true either. Only the major transit providers and CDNs like Level 3 and Akamai and the major players like Google that are so big they're effectively their own transit provider/CDN have any interesting peering.

All the smaller players get connected through one of the transit providers, because Joe's Website only wants to hook up from wherever Joe's servers are, not at every carrier hotel in the world like Google does. So they pay a transit provider to handle that and the ISP only has to deal with the transit provider.

The problem comes when the ISPs want to charge major transit providers monopoly rents for peering (which they would have to pass on), or refuse to upgrade the links to them when they're saturated in order to force their big customers (like Netflix at one point) to pay monopoly rents for peering directly.

> My initial objection was that you over-simplified.

It's supposed to be a simplification. The point isn't to make a full accounting, it's to compare orders of magnitude. Even using your numbers with unrealistically conservative assumptions about everything, they're still a single digit percentage of the typical customer's bill.