On Mon, 23 Aug 2010 17:19:26 -0400 Jared Mauch <ja...@puck.nether.net> wrote:
> > On Aug 23, 2010, at 5:11 PM, Mark Smith wrote: > > > On Mon, 23 Aug 2010 17:24:00 +0200 (CEST) > > sth...@nethelp.no wrote: > > > >>> And all you'll end up with is IPv4 with bigger addresses. You really > >>> should catch up with the useful features of protocols that were > >>> designed in the late 80s / early 90s, like IPX, Appletalk, DECNet and > >>> CLNS. > >> > >> For me "more addresses" is the *only* justification for IPv6. All the > >> other "useful" features are either uninteresting or even *unwanted*. > >> > > > > Just so I know, are you confirming that you've only ever used IPv4, and > > know nothing about any other protocols and how they worked? > > > > Your view seems to me to be a bit like saying, "I'm perfectly happy > > with my 1970s car, it gets me from A to B, and I see no reason to have > > electric windows, anti-lock brakes, electronic fuel injection, or a GPS, > > because my 1970s car doesn't have them". > > If customers were saying they wanted "network based security" more people > would be offering it. Problem is that most providers set the price of > $security_thing (either 2547-style vpn, etc..) as something too costly > or external auditors don't sign-off on it, so they say: > > give me $dumb_pipe @ 5G/s and toss their own IPSec/VPN/whatnot > device/appliance on-top of it. > > The biggest feedback I hear from people about IPv6 (besides the extra > bits for addressses) is "Security", but they generally don't know what > that is outside marketing speak. Saying that your 1970's car doesn't > have an onboard GPS is correct, but that doesn't mean the customer won't > just get a $200 aftermarket at walmart. Well, GPS was only one of the examples I used, and I was envisioning one that is built into the dash. To continue with the analogy, how many people would buy and install after-market electric windows, anti-lock brakes, electronic fuel injection etc.? > Is that ideal for the $car_mfg? > Surely it is a hit on the profit margin, but overall customers are just > asking for $dumb_pipe @ 5Gb/s @ $low_low_price_per_meg. > I think you haven't really realised who the customers are of IPv6 or more generally Internet technologies. It isn't the people who pay ISPs or indirectly pay to access the Internet via corporate networks. They're customers of Internet services. More specifically they're customers of the function of allowing applications or processes to intercommunicate internationally. They don't and shouldn't have to care about the technology that makes it work. That's what operators are for. Thats why the "cloud" analogy for networks is a good one - it represents a concept of "don't worry about that magic inside, just assume your data will get from one side to the other". (no I'm not talking about "cloud computing") The people who buy, deploy and operate the technology of the Internet are the customers of IPv6. That's not the general public, that's people like you and I. Our job is to use Internet technologies to deliver Internet services to our customers. So when I say I think this is what I'd like IPv6 to do, or it shouldn't work like IPv4 because there are potentially better ways to do it, I'm the customer. So when I ask for simplicity, I'm asking for it so that my job is easier and quicker, and therefore cheaper in time or money for my customers - the end-users of Internet services. I'm even asking for it so that they don't even have to employ me to get basic connectivity going on a small scale, even without understanding the technology. (I'm also asking for it so that I won't have relatives ask me to help setup their Internet connections at home - as if I want to do my job unpaid on the weekend as well ...) Protocols like IPX and Appletalk demonstrated how easily deployable protocols could be made on a small to medium scale, and they did it in the 1990s. They define the minimum of what any protocol designed since then should be able to achieve in user friendliness. IPv4 doesn't. To paraphrase Soylent Green, "IPv6 customers is operators!" Regards, Mark. -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------