First of all I'm not "voting" for reclaiming address space for "real" use. I am far from that. Companies do have money to either buy some spaces from other companies or invest in migrating to IPv6 which will be even better. But there are some countries (like Poland) where IPv6 is a dream. I've been poking my ISP about IPv6 once or twice a year *when* there will be IPv6 because 5 years ago I signed a contract only because they told me that they will have IPv6 in less than a year. In our hackerspace we have enough bandwidth and power to run a small colocation for people - Raspberry Pis, whatever. We can provide them with a IPv6 addresses but this will be useless for them because in their homes they don't have IPv6 and they won't have for a long time. We will provide them with a IPv4 by becoming a LIR together with one other hackerspace but it will cost us money that could (and should) be spent on new or better equipment. What's more - because of the limited amount of money we often buy an old networking equipment that supports only IPv4. So we will have to keep the dual stack for a long time. I'm currently doing an internship in a major technological company and even they have to support dualstack in a various places despite having majority of the network migrated to IPv6 only. I know that this is not perfect but this is what it is and it won't change soon whether we like it or not. Ian McLaughlin wrote: > As the owner of a data centre as my day job, it bothers me to see people talk about all of the ‘wasted’ IPv4 space that should be reclaimed. Spend some time and look at historical allocation rates. Even if we were able to reclaim the entire 44 block for ‘real’ use, it would delay the inevitable by months at the most. > > We have run out of IPv4. No amount of draconian policies will change this fact. Things like NAT helped for a while. Next, we’re going to get carrier-grade NAT so most clients will get RFC1918 space (and, quite frankly, 95% of them won’t even notice as long as they can still get to Facebook on their phones). > > The only solution to IPv4 running out is more widespread adoption of IPv6. > > You want to help the maker community with address space? Help them adopt IPv6. The maker movement and IoT will need orders of magnitude more addresses than IPv4 is ever going to give them. > > Ian > >> On Sep 3, 2017, at 7:20 AM, Brian Kantor &lt;Brian(a)UCSD.Edu&gt; wrote: >> >> You are far from the first person to make these points. The Board >> of Directors is acutely aware of how valuable our address space is >> and is actively considering how best to expand our mission to make >> better use of it. But we have to move slowly; IP address space is >> very much like virginity in that once you lose it, you can never >> get it back, so we must be very careful. Thank you for your input. >> - Brian >> >> On Sun, Sep 03, 2017 at 07:06:38AM -0700, pidpawel wrote: >>> It's important to be aware of the timeline but this anniversary might >>> also be a good time to look back at the history and think about the >>> impact this assignment has made. >>> >>> What *new* technologies has been developed because of this network? >>> Which crises have been mitigated using this network? Have it helped to >>> spread the HAM radio "spirit" to the young people? What other good >>> things have this network done? >>> >>> It's very sad for me to say that the only thing I can see about this >>> network is a bunch of guys trying to stick with old technology (RIP? >>> please.) at all cost and arguing who is more important in a tree of >>> people allocating numbers. >>> >>> A /8 network is a great value nowadays, the IPv4 especially in Europe is >>> in a huge crisis and getting new addresses is nearly impossible. From >>> the other hand most of the address space in this network is unused but >>> when you try to request allocation for yourself you can easily get >>> rejected because of silly reasons. (I didn't even try to request one for >>> myself after my friends showed me the coordinator responses.) >>> >>> There might be some things going on the used parts of the network but I >>> couldn't find any example that could be genuinely useful to the world. >>> >>> Could you please prove me wrong or if I'm right try to consider sharing >>> the address space with a "new" movement of hacking and hackerspaces? HAM >>> radio should be all about hacking [1] but frankly speaking I don't see >>> much of it in HAM radio space these days. There are some exceptions - >>> i.e. "OFDM modem" thread from the last days but there are as rare as >>> freakin' unicorns. >>> >>> This message is not meant to be mean. I'm just trying to pinpoint some >>> things I've seen as an observer of this network and HAM radio (mostly in >>> Poland but also the "worldwide" parts) and share some ideas how the >>> things can be done better and provide a better value to the whole world. >>> >>> [1] https://en.wikipedia.org/wiki/Hacker_culture >>> >>> -- >>> I wish you all the best >>> SQ9PID >>> _________________________________________ >>> 44Net mailing list >>> 44Net(a)hamradio.ucsd.edu >>> http://hamradio.ucsd.edu/mailman/listinfo/44net >> _________________________________________ >> 44Net mailing list >> 44Net(a)hamradio.ucsd.edu <mailto:44Net@hamradio.ucsd.edu> >> http://hamradio.ucsd.edu/mailman/listinfo/44net <http://hamradio.ucsd.edu/mailman/listinfo/44net> >> >> >> --- >> Filter service subscribers can train this email as spam or not-spam here: http://my.email-as.net/spamham/cgi-bin/learn.pl?messageid=1C81EE0A90B311E7B… <http://my.email-as.net/spamham/cgi-bin/learn.pl?messageid=1C81EE0A90B311E7B857E88593ED0201> >

On 4 September 2017 at 07:51, Brian Kantor &lt;Brian(a)ucsd.edu&gt; wrote: CIDR notation is a good way to explain the difference in the address spaces. A /32 of IPv4 is a single IPv4 address which people should be familiar with getting from their ISP as standard. A /32 of IPv6 is the same "percentage of the address space" as it is in IPv4. The difference being that in IPv6 a /32 can number an entire ISP with tens of thousands of customers each with hundreds of subnets containing as many globally addressable services as they like. So, everyone should already get more IPv6 addresses than they "need" from their ISP by default. The general "I need global addresses" use case doesn't really apply to IPv6. 2a06:8187:fe19::/48 is a direct allocation I got from RIPE. I decided to get a "static allocation" from RIPE instead of one of my ISPs so of course I had to pay for the allocation. It costs an entire £5/year through my chosen LIR and this is IPv6 space that is "in my name" and not associated with a specific connectivity provider. I am currently routing this through a free tunnel to my home connection. That £5/year is a single site /48 allocation which contains 65536 *subnets*. For other locations I also have 2a06:8181:ab00::/40 which is 256 of those /48 allocations and costs £60/year. This really is an absolutely massive allocation by IPv4 standards and the only justification I have is that this is the next size up from a single allocation! I am routing this IP space directly to different virtual servers. For reference purposes: TunnelBroker.net - Will give you a static /48 IPv6 allocation on a nearby "VPN server" that can be routed to any IPv4 address for free. They also provide free BGP tunnels if you have your own ASN and IPv6 allocations. RMLH - An IT consultancy offering ASNs for £50 one-off, /48 of IPv6 £5/year, a massive /40 of IPv6 £60/year. It's not their main line of business but I can't comment on anyone else. Mentioned mostly as a reference for pricing. VULTR.com - my "main" provider for routing most of my IP space. Has 15 locations offering basic virtual servers for $5/month with BGP available everywhere. In short: It's cheap and easy enough for you to sort out your own IPv6 addressing and routing rather than asking Brian to do it for you. ;) The above numbers could also be used to make a different argument - if you know several people who want an "amateur radio allocation" then you could always self-fund a local one! It's also a good opportunity to learn about a more obscure side of network routing. One final resource I can link to is https://u4477715.ct.sendgrid.net/wf/click?upn=PkDgZs68dlw2b-2FVF7LiU-2FWpYt… which will give free IPv6 allocations for you to route yourself, but does require some technical know-how with regards to setting up your database objects etc as there is obviously no support - I do have an allocation from them, however I would be cautious about putting production services in IP space I didn't have a contractual agreement for given how cheap it is to get a "proper allocation"! Hopefully that explains why an Amateur Allocation makes less sense with IPv6, whilst avoiding that messy political area around who pays for the cost of a large enough allocation for every amateur in the world at commercial rates. With regard to the future of an Amateur Network as some kind of VPN overlay, I would like to point out 2 relevant hackerspace projects. ChaosVPN is taking a similar approach to the current AMPRNet system, just with less outdated protocols *cough*. However, perhaps https://u4477715.ct.sendgrid.net/wf/click?upn=Ki4chJONuNfM0VomxEE-2BocPAFq6… is a more relevant model to be looking at? I believe someone has already started along the lines of signing public key certificates of licensed amateurs... - Mike

Sad he left the ham radio business happy to hear he win the battle with the illness ... ________________________________ I'm reminded of this because Phil and I had dinner last night (he's doing well, thanks) and got to chatting about the old days of packet and AMPRNet. - Brian