3 ms·
> 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 sid
by AnthonyMouse 9y ago
> 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.
- wanderingbort 9y agoyou 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)
- AnthonyMouse 9y agoYou'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.
- wanderingbort 9y ago> 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.
- AnthonyMouse 9y agoThe 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.