On 13/05/2017 10:16, Jens Link wrote:
Jan Zorz - Go6 <jan@go6.si> writes:
Hi,
Hi,
Draft version 2 is now available for reading at https://sinog.si/docs/draft-IPv6pd-BCOP-v2.pdf
I like but I don't see it happening.
1. Stable Addresses - Data protection people will have a hart attack when they read this. As will many customers. Don't get me wrong I *do* want a stable prefix at home but many people don't. Changing addresses gives them some pseudo anonymity and the warm feeling that they are not traceable and secure.
Data protection people will have to learn how technology works and stop breaking IPv6 deployments with enforcing bad practices from IPv4 world. WE dynamically changed IPv4 address because we started running out of them, not to ensure anonimity. That warm fuzzy feeling is made-up collateral damage that was never even a intent ;) As Jordi mentioned, traceability starts on L7 and it doesn't matter how much you change addresses, you'll be trackable. For reference, try it on https://panopticlick.eff.org/ Click, change address, click again.
And stable addresses are a way to make money. sys4 has a office in Munich and VDSL from M-Net. We pay extra for one stable IPv4 address but they wont hand out a stable IPv6 prefix. If you want stable v6 you have to buy their SDSL products which are way more expensive. We don't want to run any service in the office. We just want stable addresses for equipment and some training / lab VMs in the office.
We are documenting best operational practice here, not a bad one.
2. Prefix length. I totally agree: Handout a /48 or /56. But this doesn't happen right now. And I don't think provider who have v6 now wont change their ways of doing things. I'm a customer of Kabel Deutschland an I can get either DS-Lite with a /64 or a public IPv4 Address (I chose the later an tunnel my own IPv6). For some CPEs (provided by KDG) they handout a /62. A friend recently told be about another provider handing out /57. Unfortunately the competitors are not much better.
Again, we are documenting best operational practice here, not a bad one.
Then there are the smalltown providers providing FTTH[1] who think that becoming an ISP is easy. They don't become an LIR, they get a /2x from their upstream and I guess they wont get much more then a /48 *if* they do implement IPv6 (but right now NAT seems to work to well and people are happy that they have faster internet then before).
<shrug> Can't fix those either ;) If operators did wrong things and/or taken wrong decisions when deploying IPv6, then it might be worth sending them this BCOP document (when we reach the stable version), maybe they start thinking about it and make their deployment better. Cheers, Jan