- NANOG - lists.nanog.org

RE: Internic address allocation policy (fwd)
by Justin W. Newton 20 Nov '96

20 Nov '96

At 02:01 PM 11/20/96 -0600, Jim Fleming wrote: >P.S. Regarding your comments about the >InterNIC. I have objected to people like yourself >having discussions with InterNIC (or NSI) >employees about other people's business >matters and then running around mailing lists >referring to those discussions and acting like >you are very chummy with the InterNIC. I >think that I have been very clear on those >points..."The InterNIC has no business >discussing other people's business with >you, no matter how buddy buddy you are..." Jim, I would love to see one example where I have made any mention of any knowledge given to me by the internic about anyone's business other than my own. It simply has never happened. If you are going to make accusations back them up. Justin Newton Network Architect Erol's Internet Services

1 0

RE: Internic address allocation policy (fwd)
by Jim Fleming 20 Nov '96

20 Nov '96

On Wednesday, November 20, 1996 2:23 PM, Justin W. Newton[SMTP:justin@erols.com] wrote: @ At 02:01 PM 11/20/96 -0600, Jim Fleming wrote: @ >P.S. Regarding your comments about the @ >InterNIC. I have objected to people like yourself @ >having discussions with InterNIC (or NSI) @ >employees about other people's business @ >matters and then running around mailing lists @ >referring to those discussions and acting like @ >you are very chummy with the InterNIC. I @ >think that I have been very clear on those @ >points..."The InterNIC has no business @ >discussing other people's business with @ >you, no matter how buddy buddy you are..." @ @ Jim, @ I would love to see one example where I have made any mention of any @ knowledge given to me by the internic about anyone's business other than my @ own. It simply has never happened. If you are going to make accusations @ back them up. @ I suggest that you read your postings on the ISP mailing list. Do you post to that list ? -- Jim Fleming UNETY Systems, Inc. Naperville, IL e-mail: JimFleming(a)unety.net JimFleming(a)unety.net.s0.g0 (EDNS/IPv8)

1 0

RE: Internic address allocation policy (fwd)
by Jim Fleming 20 Nov '96

20 Nov '96

On Wednesday, November 20, 1996 1:47 PM, Justin W. Newton[SMTP:justin@erols.com] wrote: @ At 12:06 PM 11/20/96 -0600, Jim Fleming wrote: @ >I have suggested in the past a "neighbor net" approach. @ >It is a simple approach in which people on each side @ >of an allocation, be responsible for publishing a periodic @ >usage report. They would obtain this information from @ >their "neighboring" administrator. @ > @ >This approach can work for /24s, /16s and any size @ >allocations, as long as one is aware of who their binary @ >neighbor is in the IPv4 address space. @ > @ >To illustrate a /8 example, Hewlett Packard and Apple @ >Computer would be responsible for the report on @ >Digital Computer Corporation. @ > @ > Hewlett Packard - 15.0.0.0 @ > Digital Computer - 16.0.0.0 @ > Apple Computer - 17.0.0.0 @ > MIT - 18.0.0.0 @ Jim, @ This lynches it. You are absolutely insane. You whine and moan about @ having to tell the EvilInternic(tm) about your network, potential @ customers, etc etc, yet you believe that DEC would gladly give information @ to Apple about their internal network etc etc. More effort would be sepnt @ in court with lawyers haggeling over what DEC had to give Apple and Apple @ had to give DEC in order to properly be audited than would have originally @ been spent with an audit by an external agency doing the audit. Or wait, @ let me guess DISA wants to be audited by a foriegn company (30.0.0.0 for @ DISA and 32.0.0.0 for the foriegn company, 31 is IANA reserved, so I am @ assuming that it would be skipped). @ @ Yeah, that. @ @ Justin Newton @ Network Architect @ Erol's Internet Services @ @ Justin, As usual you missed the point. The idea is to promote a "self-auditing" system. If Digital does not want to provide any information that is fine. Hewlett Packard and Apple simply have to say, "Our neighbor Digital Equipment, has declined to provide information". That will make the job very easy for HP and Apple. That attitude can ripple through the entire system. If that happens then, no one should complain about IP allocations and misuse. What was that old saying about "glass houses" and "casting the first stone"...? P.S. Regarding your comments about the InterNIC. I have objected to people like yourself having discussions with InterNIC (or NSI) employees about other people's business matters and then running around mailing lists referring to those discussions and acting like you are very chummy with the InterNIC. I think that I have been very clear on those points..."The InterNIC has no business discussing other people's business with you, no matter how buddy buddy you are..." -- Jim Fleming UNETY Systems, Inc. Naperville, IL e-mail: JimFleming(a)unety.net JimFleming(a)unety.net.s0.g0 (EDNS/IPv8)

1 0

Re: Why doesn't BGP... -Reply
by Vadim Antonov 20 Nov '96

20 Nov '96

Curtis Villamizar <curtis(a)ans.net> wrote: >IP routing has an scaling advantage over call setup. Call setup can >take advantage of the lowest loaded path at the time of setup, >providing a better chance to load balance (ala PNNI). The Pluris load-balanding approach allows gradual shifting of load between pipes while doing native IP routing. >The VC table >has a speed forwarding advantage over the IP radix tree. The upper levels of the radix tree fit nicely in L1 cache, so the real advantage of VC table is far smaller. >The radix >tree may have a space advantage due to aggregation. Radix tree is O(log n) from size of the network. (n is the number of hosts). VC table is O(n), at least at some switches. >That in a >nutshell is the IP vs ATM flamewar (did I miss anything - ommisions >private mail unless it was major). Nah, it is much funnier. I plan to publish my comments on the issue at Pluris web site shortly. >The interest in IP switching >(CSRs, tag switching, IP switching proposal de jour) comes from a >desire to combine the best of the two. What happened to K.I.S.S.? :) --vadim

1 0

Re: Why doesn't BGP... -Reply
by Vadim Antonov 20 Nov '96

20 Nov '96

Larry J. Plato <ljp(a)ans.net> wrote: >Connectionless and Connection oriented both refer to packet switched >technologies, whereas the phone company uses circuit switched technology. Yes, this is absolutely correct. >Circuit switched means that the same wires/timeslots are dedicated to >a call from the time it starts until the time it finishes.If you do >not speak, the wires are idle/wasted. I am sure you understand packet >switching. In a packet switched network, connectionless means that each >packet has no state information, and stands alone, in the IP world we call >this UDP. Connection oriented would be the equivalent of telnet or >some other TCP service. This is a terminological question. Here the talk was about connectionless and connection-oriented _network_ layer; not the transport layer. The connection-oriented packet routing network is a generalized case of circuit switching -- you can multiplex connections differently. The fundamental difference between connectionless and connection-oriented networks is the amount of state necessarily kept by gateways in order to perform forwarding of the bits. A gateway of a connectionless network has only to keep the traffic-independent topological information (this definition is more generic than just next-hop routing; it includes source-based routing, semi-flows, etc). A gateway of a connection-oriented network has to keep the topological information (in order to be able to route connections) and the traffic-dependent table of connections. BTW, by that definition an IP network which supports RSVP _is_ a connection oriented network. If gateway has state which is modified by traffic, that state must necessarily grow with the traffic; in Internet case, exponentially and at the rate far exceeding Moore law's. >SS7 (Signalling System 7) is a connectionless packet switched technology >used to control the setup and teardown of circuit switched calls. >Originally is was used as a database query technology to make 800 >numbers portable across carriers. If this did not make sense I can descibe >it in a little mnore detail offline. At least that telcos got right. --vadim

3 3

RE: IP Allocation
by Jim Fleming 20 Nov '96

20 Nov '96

On Wednesday, November 20, 1996 8:30 AM, Randy Bush[SMTP:randy@psg.com] wrote: @ The use of address to differentiate between multiple web sites on one host @ was only a hack to get around a gap in the http protocol. This too shall @ pass. @ Some might also claim that there is a limitation in DNS system because "port" numbers can not easily be returned with an IP address. If that had been possible, then one IP address could have been used and different TCP port numbers. With the introduction of AAAA records into the DNS system, there is now an opportunity to send 128 bit binary quantities. Part of that quantity can be interpreted as a 32 bit IP address and part can be interpreted as a port number. -- Jim Fleming UNETY Systems, Inc. Naperville, IL e-mail: JimFleming(a)unety.net JimFleming(a)unety.net.s0.g0 (EDNS/IPv8)

3 2

RE: Internic address allocation policy (fwd)
by Jim Fleming 20 Nov '96

20 Nov '96

On Tuesday, November 19, 1996 10:42 PM, Ed Jackson[SMTP:ed_jackson@INS.COM] wrote: @ I would like to see more information on this item in particular -> is this @ being done currently -> or are plans underway to do so?... in regards to @ 'un-lending' that address space from those organizations to make them @ available for everyone else? @ @ now that would create some interesting discussions - ;) @ @ ed jackson @ @ At 04:03 PM 11/19/96 -0600, Jim Fleming wrote: @ > 7. A complete review of /8 allocations should be done @ > with a complete audit of how the companies @ > that have those allocations are utilizing those @ > blocks (as well as other blocks). The CEOs and @ > shareholders of those companies should be @ > informed that their past policies may not conform @ > to modern standards of frugal allocations. @ @ @ +--------------------------------------------------+ @ Ed Jackson @ International Network Services @ @ numeric pager: 1-888-352-4117 @ email pager: <mailto:page_ed_jackson@ins.com> @ +--------------------------------------------------+ @ @ I have suggested in the past a "neighbor net" approach. It is a simple approach in which people on each side of an allocation, be responsible for publishing a periodic usage report. They would obtain this information from their "neighboring" administrator. This approach can work for /24s, /16s and any size allocations, as long as one is aware of who their binary neighbor is in the IPv4 address space. To illustrate a /8 example, Hewlett Packard and Apple Computer would be responsible for the report on Digital Computer Corporation. Hewlett Packard - 15.0.0.0 Digital Computer - 16.0.0.0 Apple Computer - 17.0.0.0 MIT - 18.0.0.0 Digital would then be responsible for the report on Apple Computer along with MIT. A web site could easily be constructed with pointers to the various sites where the usage reports are stored. To help each person's "neighbor", the company would provide the base material for the report. The neighbors would act more as auditers, than reporters. @@@@@@@@@ Hewlett-Packard Company (NET-HP-INTERNET) 3000 Hanover Street Palo Alto, CA 94304 Netname: HP-INTERNET Netnumber: 15.0.0.0 Coordinator: Milligan, Michael J. (MM53) milli(a)HP.COM 1-415-424-3706 (FAX) 1-415-424-3632 Domain System inverse mapping provided by: HP.COM 15.255.152.4 HPCUOC.CUP.HP.COM 15.255.208.5 HPFCLA.FC.HP.COM 15.254.48.2 HPLB.HPL.HP.COM 15.255.59.2 RELAY.HP.COM 15.255.152.2 Record last updated on 16-Aug-95. @@@@@ Digital Equipment Corporation (NET-DEC-INTERNET) Netname: DEC-INTERNET Netnumber: 16.0.0.0 Coordinator: Reid, Brian K. (BKR) reid(a)PA.DEC.COM 415) 688-1307 Alternate Contact: Digital Equipment Corporation (DEC-NOC) noc(a)digital.com (415) 688-1380 (800) DIGITAL Domain System inverse mapping provided by: NS.DEC.COM 204.123.2.42 CRL.DEC.COM 192.58.206.2 Record last updated on 17-Oct-94. @@@@@ Apple Computer, Inc. (NET-APPLE-WWNET) 20740 Valley Green Drive, MS: 32-E Cupertino, CA 95014 Netname: APPLE-WWNET Netnumber: 17.0.0.0 Coordinator: Zimmerman, David Paul (DPZ) dpz(a)APPLE.COM +1 (408) 974-2436 (FAX) +1 (408) 974-3103 Alternate Contact: Tsuno, Eugene Y. (EYT) eugene(a)APPLE.COM (408) 974-4310 Domain System inverse mapping provided by: NSERVER.APPLE.COM 17.254.0.50 APPLE.COM 130.43.2.2 COLOSSUS.APPLE.COM 130.43.2.6 Record last updated on 15-Jan-96. @@@@@ Massachusetts Institute of Technology (NET-MIT-TEMP) 1 Amherst Street Cambridge, MA 02139-1986 Netname: MIT Netnumber: 18.0.0.0 Coordinator: Schiller, Jeffrey I (JIS) jis(a)MIT.EDU +1 617 253-8400 (FAX) +1 617 258-8736 Domain System inverse mapping provided by: STRAWB.MIT.EDU 18.71.0.151 W20NS.MIT.EDU 18.70.0.160 BITSY.MIT.EDU 18.72.0.3 Record last updated on 14-Jan-94. @@@@@ Some have challenged that this will not work because, get this, some neighbors are no longer using their IP addresses and would not respond to inquiries. I have suggested that this might be a good way to find these fragments of IP address space. Others have commented that they do not want anyone to know how they are allocating those 16 million addresses, and that it is proprietary information. Maybe that should be all that the "neighbor net" report says. As an example only, MIT could claim that they asked Apple and all they saw were No Trespassing signs. This could be reported to the CEO of Apple and the Shareholders. Apple could then decide what they want to present. At the present time, it appears that the CEOs of these companies and the leaders of these Universities do not have any idea that they have effectively homesteaded massive areas of Internet Land. The impression that their company is giving to the Internet community may be controlled by one or two "geeks" who have no authority to create that impression. In my opinion, the CEOs of those companies need to be contacted as part of a "neighbor net" plan. I have a feeling their PR departments will be interested, now that they have heard of the Internet. -- Jim Fleming UNETY Systems, Inc. Naperville, IL e-mail: JimFleming(a)unety.net JimFleming(a)unety.net.s0.g0 (EDNS/IPv8)

1 0

RE: IP Allocation
by Jim Fleming 20 Nov '96

20 Nov '96

On Wednesday, November 20, 1996 6:19 AM, Dima Volodin[SMTP:dvv@sprint.net] wrote: @ <<File: ATT00168.txt>> You can use MX records with a port number in the PREFERENCE field and a single 4-byte label in the EXCHANGE field. Or use TXT RRs of whatever format you bloody will. Dima @@@@@@@ Yes, good points... In general I do not think that people have looked closely at the "hidden" features that can be mined from the cryptic DNS system. In my opinion a boundary needs to be placed around the IPv4 "Core" transport network. We can rely on that network for two things: 1. Low-cost international bit transport, using 32 bit addressed encapsulation headers. 2. A working DNS system for use in mapping symbolic names to binary values. A new Internet can be build around the edges of that core network. How the above capabilities are exploited are up to conventions and agreements made outside of that core transport facility. As an example, we are using TXT records to update the name server information in true root name servers. That information is provided by Top Level Domain name servers which provide it via, what looks like, a standard DNS zone file. The root name server uses the DNS zone transfer capabilities to automatically update its records based on the infromation "pulled" from the TLD name server. There are many other applications for the DNS system. I am sure that more and more people will start to invent new uses, now that the "core" transport network is stable and not evolving from a technical point of view as fast as the technology around the edges. -- Jim Fleming UNETY Systems, Inc. Naperville, IL e-mail: JimFleming(a)unety.net JimFleming(a)unety.net.s0.g0 (EDNS/IPv8)

1 0

Re: Internic address allocation policy (fwd)
by owen＠DeLong.SJ.CA.US 20 Nov '96

20 Nov '96

> > FYI: there will be an Internet Address Registry Evolution (yes, > > we'll be coming up with a new name) working group meeting... > > Oh, I dunno, I kind of like it... "Five years ago, independent > registries did not exist, and now IARE one?" :-) > > -Bill This confirms something I have suspected for years... Bill Woodcock _IS_ a PUNdit. Now back to our regularly scheduled political bashing... Owen

1 0

Re: Internic address allocation policy (fwd)
by Brian Tackett 20 Nov '96

20 Nov '96

On Tue, 19 Nov 1996, Kim Hubbard wrote: > Let's say, hypothetically speaking, that every ISP decides to do this. > What's to stop some ISPs from listing a bunch of companies that are > not their customers and are not really receiving address space from > them for the purpose of getting more address space. Absolutely nothing, and therein lies the crux of the problem. To be honest, I can understand the motivations behind InterNIC's policies. However, comprehension of said policies does not mean that I agree on all counts. It truly shouldn't (IMHO) be as difficult as it sometimes is to get address space, but perhaps it's simply pushing the limitations of the current system. The problem I see most often in smaller organizations as they grow larger is the lack of consistency in internal policies, and the like. i.e, FooNet Inc. has 12 or so Class C's from MCI, and as they pick up extra providers perhaps the dance goes on until they have a hodgepodge of networks. Now the ISP enters the next level, and someone decides they need a larger CIDR block. They go to the NIC, and are told to submit a justification. The problem seems to be that many ISP's simply have not kept up with SWIP or RWhois, and are completely unprepared for a stricter accounting of their address space, even when/if they are efficiently using it. The frustration enters when the ISP begins to see the possibility of losing customers, as they try to sort out their justification. It's quite easy to see how people might feel compelled to scream at Kim, call their lawyer, make themselves sick and generally have fits over the problem. On the other hand, the NIC staff are simply not superhuman beings capable of personally babysitting a thousand customers as they try to sort out their addressing problems and policies. Hence the restrictive policies, which are really a defense mechanism to prevent unscrupulous people from taking advantage of the system and further screwing it up. So, my question is: What can we do to make it better? Not "What can Kim Hubbard do to fix it", and not "What can those stupid ISP's do to fix it", but what can both do collaboratively to work out a consistent method of doing things, and making sure that it's well known to both new existing providers of network solutions.

5 6