- NZNOG - lists.nznog.org

[Fwd: Re: Charging for IPv6]
by Gerard Creamer 29 Nov '06

29 Nov '06

On 30/11/2006 11:36 a.m., Andy Linton wrote: > For us having this address space from APNIC pushes us into the next tier of > membership and costs us an additional US$2500 a year. We have had a very > serious discussion here about returning this to APNIC and getting a reduction > in our charges. We recently got our first allocation from APNIC, and I was surprised that we had to choose 4 or 6 and had to pay more for both - as Andy says, are we being encouraged to do this or not? We are still small, so the additional cost is prohibitive. I would have thought there would be an associated IP6 allocation with every IP4 allocation, but I guess I was naive. Gerard

1 0

Re: [nznog] Charging for IPv6
by jdmwoolley＠ihug.co.nz 29 Nov '06

29 Nov '06

> A rough eyeball of the DSC stats for the last day or so > seems to indicate a little over 50% A queries and about > 10% AAAA. > Now THAT is an interesting statistic. Correct me if I'm wrong, but it means that a fair proportion of the queries coming in are TRYING to use v6, but presumably either aren't given a v6 address back or are unable to access it once obtained. Given that Vista will default to using v6 and fall back to v4 and the above statistic, one could surmise that if v6 was deployed, it would almost immediately get a fair portion of traffic. Indeed, when I tried pinging between two boxes at home that I had put Vista RC1 on I got a reply from a v6 address. Even though I'd specified the v4 address in the ping command Jonathan

1 0

Re: [nznog] THE SKY IS FALLING ( was Re: IPv4 Exhaustion)
by Scott Weeks 29 Nov '06

29 Nov '06

--- mjl(a)luckie.org.nz wrote:----------------------- From: Matthew Luckie <mjl(a)luckie.org.nz> > This was discussed earlier in the recent flurry of email on the > subject. Also, this is why I was asking about VPRNs. However, I > missed the part on why STBs need unique IPs or why they need 2 per > box. : http://www.nanog.org/mtg-0606/real/100million.ram : : the answer to the second question is about a minute in. ---------------------------------------------------- Geez, I'm trying to listen, but the quality is so bad I can hardly understand it. Does anyone know where a 'clean' copy of the speech is? scott

2 1

Re: [nznog] IPv4 Exhaustion
by Scott Weeks 29 Nov '06

29 Nov '06

> The major driver will be customer demand (read: $$$$), or the network : Funnily enough, I was reading this only yesterday: : : http://www.microsoft.com/technet/windowsvista/network/ipv6wv.mspx : : Have a look a the "Business Drivers" section especially. It's just funny numbers: "A recent report from NIST3 estimates that IPv6 could provide a US$10 billion per year benefit to the U.S. and the Japanese government estimates that IPv6 could generate a value of US$1.55 trillion." They give a reference to footnote 3. I searched footnote 3 (www.microsoft.com/technet/windowsvista/network/ipv6wv.mspx) for the word trillion and only get one match: "The 128-bit address header in IPv6, in contrast, provides approximately 3.4x1038 addresses, enough to assign trillions of addresses to each person now on earth or even to every square inch of the earth’s surface." So, to segway into pet peeve of mine, this is the reason to allocate very large (for example: allocation of IPv4 /8 blocks to everyone that asked in the not so distant past) blocks of v6 space. Heck, no one will EVER need that much IP space! <8-) Show me the money, not funny numbers... scott Oops, I forgot: beer :-)

18 27

Another Topic
by Peter Youngquest 29 Nov '06

29 Nov '06

Hey All I am personally bored with the IPV4-6 discussion, but while everyone is up and talking. What is everyones opinion in regard to PBT. Does this have a place, can it work happily as a feeder into an MPLS cloud or will it remove the need for MPLS. Rgds Peter

4 5

Re: [nznog] THE SKY IS FALLING ( was Re: IPv4 Exhaustion)
by Scott Weeks 29 Nov '06

29 Nov '06

--- richardn(a)cs.waikato.ac.nz wrote:---------------------- From: Richard Nelson <richardn(a)cs.waikato.ac.nz> While not wanting to spoil a good discussion with facts, there is a presentation on Comcast's plans for IPv6 from a recent NANOG here: http://www.nanog.org/mtg-0606/durand.html The summary is 20 million customers * 2.5 set-tops/customer * 2 IPs/box = 100million IPs. The network design requires unique IPs. The ran out of 10net in 2005 and have got a large part of 73net from ARIN. Their long term solution (new services = more IPs/customer) is IPv6. ----------------------------------------------------------- This was discussed earlier in the recent flurry of email on the subject. Also, this is why I was asking about VPRNs. However, I missed the part on why STBs need unique IPs or why they need 2 per box. scott

3 2

Re: [nznog] THE SKY IS FALLING ( was Re: IPv4 Exhaustion)
by Scott Weeks 29 Nov '06

29 Nov '06

--- aj(a)sneep.net wrote: From: Alastair Johnson <aj(a)sneep.net> Scott Weeks wrote: : Finally, with VPRNs addresses could be reused. Just put : the customers in different VPRNs and duplicate the address : scheme. It's 'divide and conquer' methodology. I'm a : noobie to VPRNs, so flame me if this isn't feasible. I have : doused my flameproof underpants with beer, so I'm protected... :-) :: Oh yes. This is a valid methodology. Also, you don't :: really need every subscriber to be uniquely addressed across :: your entire network - to the closest PoP site or Super-PoP site :: is sufficient, depending on your topology. :: :: Of course, from your Ethernet switch/DSLAM/FTTx node/whatever, :: you know where and who the subscriber is, so you're not reliant :: on IP addressing to authorize content access, either.... Thanks for the answer. So, it seems that all customer's STBs could be put in as many VPRNs as necessary to divide them into groups such that RFC1918 space is never exhausted and each customer could get one (or as many as they request) IP address that's publically routable for their internet connectivity. Thus, if the Very Giant ISPs designed their network in this manner, IPv4 would take them as far as they need. Yes? scott

3 2

Re: [nznog] THE SKY IS FALLING ( was Re: IPv4 Exhaustion)
by Scott Weeks 29 Nov '06

29 Nov '06

--- gerard(a)face.co.nz wrote: From: Gerard Creamer <gerard(a)face.co.nz> On 30/11/2006 7:41 a.m., Scott Weeks wrote: > Thanks for the answer. So, it seems that all customer's STBs could be put in as many VPRNs as necessary to divide them into groups such that RFC1918 space is never exhausted and each customer could get one (or as many as they request) IP address that's publically routable for their internet connectivity. Excuse my ignorance, but I'm having trouble finding useful information on what a VPRN (Virtual Private Remote Network) is - I could guess, based on my understanding of VPNs and NAT (as it appears to be used as a NAT replacement in context??), but that might be dangerous - any pointers to useful resources would be appreciated. ---------------------------------------------- My apologies, it stands for Virtual Private Routed Network. It's also known by many other names: RFC2547bis, IP Based Virtual Private Network; BGP/MPLS VPNs, Layer 3 VPNs, etc. Here's a link that looks good, though I haven't read it yet: www.spirentcom.com/documents/760.pdf scott

1 0

VPRN
by Gerard Creamer 29 Nov '06

29 Nov '06

On 30/11/2006 7:41 a.m., Scott Weeks wrote: > Thanks for the answer. So, it seems that all customer's STBs could be > put in as many VPRNs as necessary to divide them into groups such that > RFC1918 space is never exhausted and each customer could get one (or > as many as they request) IP address that's publically routable for > their internet connectivity. Excuse my ignorance again, but I'm having trouble finding useful information on what a VPRN (Virtual Private Remote Network) is - I could guess, based on my understanding of VPNs and NAT (as it appears to be used as a NAT replacement in context??), but that might be dangerous - any pointers to useful resources would be appreciated. Thanks, Gerard

1 0

Re: [nznog] THE SKY IS FALLING ( was Re: IPv4 Exhaustion)
by chris＠chrishellberg.com 29 Nov '06

29 Nov '06

My mail app doesn't do quoting sorry :( Peter, Agree, it's not black and white and you raise many pertinent questions. But one of the governing factors should be that a shortage of addresses[1] shouldn't restrict my network design. Not saying that you need to be wasteful. Just that NAT or some proxy equivalent shouldn't be the the solution to the address squeeze[1]. /Chris [1] maybe there is an address shortage, maybe there isn't..not terribly important for this point. ----- Original Message ---- From: Peter Youngquest <peter(a)mynetworks.co.nz> To: NZNOG <nznog(a)list.waikato.ac.nz> Sent: Wednesday, November 29, 2006 8:50:35 PM Subject: Re: [nznog] THE SKY IS FALLING ( was Re: IPv4 Exhaustion) As much as I find this topic boring :-) I doubt the answer is as black and white as some would like. Exposing the evolving subscriber premise to the "world" as such may reduce some problems but obviously increases others. I think it is fair to say alot of this is a balancing game. If you expose the subscriber premise to the service core (CSCF sort of area) you increase loading at the service core. This is not a good thing considering the stability of these pieces of the network at this stage. No vendor rants to that statement pls. IMS/SIP gateways are being developed that reduce Service Core loading but still how much of the home is exposed and how much isn't. Also there are a number of initiatives working currently to further define the home, how the CPE will interact with upstream - each other, QoS on the home segment etc etc If the subscriber premise is exposed to network and as Dave pointed out earlier there are a number of less than trivial considerations. Address management, device management, Security,introducing new devices to the home (another Dick Smith VoIP phone), point of demarcation etc etc. Also is there a need to expose all the devices globally. STBs would only need visibility to the service providers, VoIP phones to the SBC/SBG, fridge to the mail server :-D Also there are the various styles of network connectivity: - single pipe, MUX UNI that will have impact on the addressing schemes. To state that having "20million subs and requiring v6 for management" as a poster child for v6 is not a good look. There are other ways of approaching this - distribution and hierarchy to me sounds feasible. Last time I looked the sky is above us. What is happening is we are moving into an interesting time where the evolution of the subscriber premise will very much determine how the networks will look. Rgds Peter

3 2