4 ms·

Are you sure about that? It doesn't fit what I understand and I've talked to a friend who works for Comcast and he says the only difference is the customer supp

Are you sure about that? It doesn't fit what I understand and I've talked to a friend who works for Comcast and he says the only difference is the customer support. The rep I talked to and the price I gave $156/mo does not include a static IP. That was something like $10/mo more. Sure, the firmware they tftp to my modem will be different but it's not giving me dedicated RF bandwidth on the exact same RF hardware path that existed before I got the firmware. It's the same uplink channels. This can be confirmed by using a splitter, attenuator, and looking at the RF signals directly in the frequency domain over time with an with a software defined radio (see: http://superkuh.com/pictures/docsis3_cable_ziggo_2017-10-08-full-spectrum.jpg http://superkuh.com/pictures/docsis3_cable_ziggo_2017-10-08-...).

Alupis 7y ago
It's definitely different circuits, and you definitely get dedicated bandwidth. You won't burst anymore either (a thing you get with shared lines), but you'll get 100% of your allocated bandwidth all the time, even if you peg it out 24/7. You can do a traceroute and see the reverse-dns names of the hops you go through too. Also remember, TCP has 8Mb of overhead per 100Mb, so you'll get maximum theoretical speed of 92Mbps on a 100Mbps line.
- Dylan16807 7y ago
  > TCP has 8Mb of overhead per 100Mb So to reach that number, I guess you're talking about 60 bytes of headers for each 1500 byte data packet, then again for an ACK packet? That's partly or entirely wrong in multiple ways. First, the combined IP+TCP header is only that big for IPv6 connections. Second, under general circumstances an ACK is only sent for every other packet, halving that bandwidth. Third, that 100Mb is going to be for one direction, so the ACK traffic going the other way doesn't affect it at all. Typical overhead is generally 3%, not 8%.

Alupis 7y ago

It's definitely different circuits, and you definitely get dedicated bandwidth. You won't burst anymore either (a thing you get with shared lines), but you'll get 100% of your allocated bandwidth all the time, even if you peg it out 24/7. You can do a traceroute and see the reverse-dns names of the hops you go through too. Also remember, TCP has 8Mb of overhead per 100Mb, so you'll get maximum theoretical speed of 92Mbps on a 100Mbps line.

Dylan16807 7y ago
> TCP has 8Mb of overhead per 100Mb So to reach that number, I guess you're talking about 60 bytes of headers for each 1500 byte data packet, then again for an ACK packet? That's partly or entirely wrong in multiple ways. First, the combined IP+TCP header is only that big for IPv6 connections. Second, under general circumstances an ACK is only sent for every other packet, halving that bandwidth. Third, that 100Mb is going to be for one direction, so the ACK traffic going the other way doesn't affect it at all. Typical overhead is generally 3%, not 8%.