dslreports logo
 
    All Forums Hot Topics Gallery
spc
Search similar:


uniqs
211

kapil
The Kapil
join:2000-04-26
Chicago, IL

1 recommendation

kapil

Member

LOL

Knowing AT&T, this doesn't surprise me at all....given two options, AT&T will always pick whichever one is more idiotic.

Although I thought what is being called "carrier grade NAT" actually had a whole new IPv4 network set aside for it by IANA pending an IETF RFC on the subject...precisely so that existing CPE already using NAT doesn't' conflict with the private IPs being use by the provider.
Skippy25
join:2000-09-13
Hazelwood, MO

1 recommendation

Skippy25

Member

They do, it is 100.64.x.x but of course AT&T being the morons they are think they can do whatever they want and not follow along with what they should.

Quite possibly the worst company ever!

jseymour
join:2009-12-11
Waterford, MI

jseymour to kapil

Member

to kapil
said by kapil:

Knowing AT&T, this doesn't surprise me at all....given two options, AT&T will always pick whichever one is more idiotic.

And cheaper (for them). Witness their solution for broadband services delivery: Essentially DSL on (weak) steroids.
ccjunk
join:2006-06-29
Austin, TX

1 recommendation

ccjunk to Skippy25

Member

to Skippy25
Actually the IETF is part of the problem. rfc6598 was only published *last* month with the 100.64.0.0/10 recommended usage. Very late; after these CPE notices started going out. ISPs have been working on CGN solutions for several years now. ISPs in Japan (APNIC or Asia Pac has already run out of public IPv4 addresses) asked for this very thing several years ago of IANA/IETF and it was shot down by various factions. They are already using 10/8 in their access networks. If this had been done when originally asked then probably CPE firmware updates would not be removing use of 10/8. Although that /10 is 4 times smaller making its use a little more problematic (Comcast received a /10 alone at the end of 2010 just for their 2011 growth allocation); it requires a more complicated plan to be reused multiple times across the footprint at a large ISP (as pointed in the RFC; which by the way is not a "standard" but a only practice recommendation by its authors).