 DewiPremium
join:2001-09-28
united kingd | A bit unclear what they mean .. is this optional? If not, I am not sure how they might enforce it. It is pretty tricky to originate what is behind a a NAT based piece of hardware. The only way I could see is by using source level packet monitoring, and time of origin on packets. IE two packets leave the customer-site at the same time (or within a time delta) for different destinations. Of course, the user could also claim to have two browser windows open at the same time. Not to mention the intensive labor involved in the monitoring (and latency cost as packets as looked at).
Not sure about that article. If you read it in the worst sense, it could mean a disturbing shift. Whereas you previously paid per connection at a (mostly) predefined usage quote (IE 1500/128) .. AT&T may want you to pay per connection, with an additional cost per computer connected. |
|
|
|
pierce2
join:1999-09-22
Santa Cruz, CA | NAT use can be detected pretty easily by monitoring the patterns of source ports... For instance, Linux 2.2.x IP Masquerade always uses source ports above 60000, which is very unusual. Pretty much all common NAT routers and software packages have a fingerprint similar to this. |
|
 ksw_92
join:2001-05-13
La Verne, CA | said by pierce:
For instance, Linux 2.2.x IP Masquerade always uses source ports above 60000, which is very unusual.
That's not unusual; that's just following IANA specs for dynamic/private port number usage (49152-65535).
I liken this to the old trick that cablecos used to use: charging a "per TV" fee. That failed in the end, too. |
|
