- Test - lists.nanog.org

Re: Intra/Inter - was Inet-II
by Vadim Antonov 11 Oct '96

11 Oct '96

alan(a)mindvision.com (Alan Hannan) wrote: >> Nobody cares about AUPs in leaf networks. AUPs in transit backbones >> are evil. Or everybody already forgot NSFNET AUP and the tons of >> related hackery in routing policies all around the world? > The interesting thing to me below is the assumption that there is > an inherent difference between a transit backbone and a private > leaf network. Hm. Hm again. Should i start explaining why AUP-type traffic restrictions (i.e. choice of different paths depending on who communicates with whom, or all-out limited connectivity) are different from peer/transit routing policy restrictions generally found in commercial networks which are not interested in AUPs but rather in delivering bits efficiently? Then, any AUP related to government funding is not going to be traffic path-based; but rather content-based, which makes it unenforceable by means other than financial coercion. Those of us who remember how it worked with NSFNET would rather not have that experience again. (Yes, and if you think the government wouldn't drag politics in it you're mistaken -- i remember that getting traffic from USSR to be allowed to NSFNET wasn't all that trivial; and thanks to Steve G.). >I believe Manning makes a good point that an AUP is inherent to a >network. Sorry, not. As is Internet is mostly AUP-free, meaning that nobody cares about content or what providers do with traffic engineering. > We have seen an increase in the discussions of AUP with respect to > backbones (MCI/SL/UU). The discussions regarding dumping defaults > and forced routing to destinations not advertised all centered > around AUP. Please, do not confuse AUPs and routing policies. Those are very different beasts. AUPs are content-based. Routing policies are purely traffic engineering. Routing policies are not a good tool to implement AUPs, and in fact attempts to do that lead to very kludgy solutions, and leave lots of loopholes anyway besides creating enormous headaches to engineering of commercial carriers who are forced one way or another to peer with AUP-full networks by the same government. We've already got NAP peering requirement mess, thank you. > While the case is there, it is not that strong. I think Sagan calls it a >pseudoscientific argument.... I'm not sure why you want to apply scientific criteria to politics or drag in the ghost of Carl. Politic science is an oxymoron, just like Christian Science. It isn't any worse to tote an ohmmeter as a sacred device for measuring purity of devoted than to look for rigid rules in political world governed by emotions. > Don't fool yourself. The I-2 is not the "faster Internet". It is > a tool to force those pesky free-thinkers to shut up. > Maybe. More likely it's a tool to give Higher Education > institutions a QOS independant from the commercial world (also cheaper). Cheaper? Have you _ever_ seen anything done by the government that is cheaper after you count all hidden costs? If it really can do things better than market why not start from more worthy causes like government supplying the foodstuffs? Oh, sorry, history is already clear on what happens then, right? Few dozen millions dead in Ukrain. I guess people allow goverments to play with networks and stuff just because it is not as important as food. That still does not change the principle. > I don't blindly accept the altruistic guise under which it was presented, > but I do think there are sig. other reasons beyond government > control. Aw, how cute. Sounds like first-grader's taking about Granddad Lenin. Surely smart grown-up people up will show us the Only And True Way, they sure know better how to spend our money. --vadim

2 1

Re: Internet II is coming...
by salo＠msc.edu 11 Oct '96

11 Oct '96

> Date: Thu, 10 Oct 1996 14:16:53 -0700 > From: Vadim Antonov <avg(a)quake.net> > To: kwe(a)6SigmaNets.com, swolff(a)cisco.com > Subject: Re: Internet II is coming... > Cc: mathis(a)psc.edu, nanog(a)merit.edu > [...] > Oh, no, not AUP again! > > What a mess. I don't think so, really. This time around, the AUP will not be at the functional core of the Internet, plus most people will be on the AUP-free side of things. I suspect the Internet II AUP will affect the Internet at large only slightly more than does the vBNS AUP. -tjs

3 3

RE: Internet II is coming...
by Jim Browning 11 Oct '96

11 Oct '96

Please accept my apologies, Robert. I didn't notice your message was not copied to the group, and I would never intentionally post a private message to a public forum without prior permission. -- Jim ---------- From: Robert E. Seastrom[SMTP:rs@bifrost.seastrom.com] Sent: Friday, October 11, 1996 5:49 AM Apparently Jim doesn't appreciate the fact that it's bad form to repost people's private email messages by way of responding to it in a public forum.

1 0

RE: Internet II is coming...
by Jim Browning 11 Oct '96

11 Oct '96

>From: Robert E. Seastrom[SMTP:rs@bifrost.seastrom.com] >Sent: Thursday, October 10, 1996 6:21 AM > >2) Copper pairs to the telco (both in the POTS scenario and the >D4-framed HDSL DS1 with a channel bank scenario) suit my purposes for >switched voice just fine. The concept of getting phone service from >an organization like a cable company or an ISP kinda leaves me cold -- >I want my phones to be more reliabe than that. What is your motive in >proposing to fix something that ain't broke? Consolidating a customer's investment in local loop circuits enables the purchase of a larger pipe ($$ leverage), and enables that circuit to be used more effectively by taking advantage of the fact that each service is not used 100% of the time. The effect of this is to encourage broader use of higher-capacity local loops, thereby cranking up the average speed of 'last mile' connections to the Internet. This makes the introduction of new applications more viable, which in turn helps to feed the continued growth of the Internet, which I consider A Good Thing. But I suspect you've heard this before... :-) >3) If you try to make a vehicle that will both haul a few tons of logs >and handle like a Lotus in the turns, you will end up with something >that does neither well. This is what you have when you try to run >voice and data across *any* network, not just ATM, but IP as well. And given enough resources, one can choose to build separate networks for each type of service. I believe MCI has done just this, although their data network now does contain ATM. However, At least one large carrier, just announced plans to migrate both their voice _and_ data networks to ATM, including support of LANE. Go figure... >4) The most often touted advantage of voice over ATM is bandwidth >savings, not having to have dedicated bandwidth per circuit, etc. But >take a look at the decay curves of (a) the cost of bandwidth, and (b) >the cost of ATM switches, and tell me how much longer it's gonna make >any sense to shell out all that extra dough. These are important factors, however it will always be an economic advantage to effectively utilize purchased bandwidth, regardless of how much is available at what cost. What will happen is that integrated services will become affordable to smaller and smaller businesses. >5) Decay in the cost of bandwidth notwithstanding, in an Internet >network given the current explosive growth, even a 15% cell tax is >unacceptable. Not to mention that native IP networks have a much more >graceful performance degradation curve when they get congested than IP >over ATM networks. Not to mention, saturation comes 17% later. However, many operators have realized that the best way to avoid congestion is to get the fattest pipe they can, and that remains an ATM pipe. If bandwidth becomes as cheap as is necessary to support your arguments, then you can stay ahead of congestion with _either_ approach. I prefer to do this as while maintaining the ability to provide everything IP has to offer, with the added benefits (raw speed, voice) that ATM brings to the table. >Well, I really have to run, but I don't think the above comments boil >down to "ATM sucks". I agree. This is now just good old fashioned argument... -- Jim

2 1

Re: Intra/Inter - was Inet-II
by Vadim Antonov 11 Oct '96

11 Oct '96

Bill Manning wrote: > Please note that firewalls, Intranets, Market-segment > nets (eg ANX, Inet-II, and the raft of MRNs) all are > there -because- of AUPs. And yes it is a mess and > it creates business opportunities and things would > be much better if we could all get along and do things > -my- way. :) > AUPs are a fact of life. We (as a community) have > to figure out how to deal with support of thousands > of AUPs in a global internet. There seems to be a confusion between private leaf networks (which nobody generally cares about) and the major backbone (as I-2 advocates portray it). Nobody cares about AUPs in leaf networks. AUPs in transit backbones are evil. Or everybody already forgot NSFNET AUP and the tons of related hackery in routing policies all around the world? And, then, we all know who pays for all of that, and where a hearty chunk of our paychecks is going. I do not see why governments should be in business of building networks. If academic users (not people claiming to be their representatives) really have a need for faster networks they have their grant money to pay for it, ok? Like they pay for everything else. For now is looks and smells exactly like pork, with a very bad taste at that, with all those Gore-Clintonese plans to buy support from academia for the next elections, and get a strong hand in controlling the same academia at the same time (it's always a sweet position to sit on a tap -- you don something government doesn't agree with, like hosting an encryption archive, or an anarchist website -- poof, here goes your goverment-paid network connection). Needless to say, that commercial entities are unlikely to allow that kind of information to be stored on their computers, so that academical networks provide most of really interesting and diverse content. Actually, that Clinton's network "initiative" is entirely in line with their other efforts to curb the free flow of information -- particularly at the place where there is a contingent of young people who would be affected most by the information. It is no secret that political views ofmost people who have spent some time with Internet tend to shift to more libertaran, as they get taste for free communication not generally afforded by the "democratic" system. Hence the effective opposition to the encryption policy and CDA. Sure as hell, after such embarrasment the administation does not like intelligentsia to have a voice. Don't fool yourself. The I-2 is not the "faster Internet". It is a tool to force those pesky free-thinkers to shut up. --vadim

3 2

Excellent host SYN-attack fix for BSD hosts
by Avi Freedman 11 Oct '96

11 Oct '96

I've been running Jeff Weisberg's SunOS patches for a day now without trouble on my news and web boxes. He's come up with an implementation of the not-going-into-the-SYN_SENT-or-SYN_RCVD state hack. It appears to be working fine. No state is kept locally; when a SYN is received, an ISS is generated that contains a few bits for reference into a table of MSS values; window size and any initial data is discarded; and the rest of the ISS is the MD5 output of a 32-byte secret and all of the interesting header info. ftp.op.net:/pub/src/syn-prophylactica/ Has sun3 and sun4 patches (the sun4 patches work so far on sun4, sun4c, and sun4m architectures). The hypothetical-this-should-work-on-other-BSD- based-systems source code in the 'net2-src' still hasn't actually been tested, I think. Tremendous thanks to Jeff for implementing what is still my favorite SYN defense. Hopefully Sun will incorporate this into their security announcement, which basically says you're screwed if you run SunOS, though it does describe how to increase the queue and decrease the SYN-holding timeout (if you have source...). Object files that do that are still described at http://www.netaxs.com/~freedman/syn/, though I think the approach implemented by Jeff is much better, and if you use that approach, increasing the queue and decreasing the SYN-holding timeout are as useless as a command-line interface on a Bay router. Again, MANY thanks to Jeff. Avi

1 0

Re: Intra/Inter - was Inet-II
by swolff＠cisco.com 11 Oct '96

11 Oct '96

Vadim is right, in that peering and transit restrictions are *address-based* and hence enforceable by routers, whilst AUPs are *content-based* and so absolutely unenforceable by technical means. In this regard, AUPs are like *most* laws in the non-virtual world - e.g., laws against speeding, unleashed dogs, homicide, fraud, etc. - in that the behavior which they proscribe cannot in even the most elementally free society be prevented; they rely instead for their observance on a (largely unwritten) social contract between netizens and network providers, and infractions can only be punished not prevented. And just as only a fraction of all crimes are noticed and reported, and only a fraction of them in the end punished, so it is with AUP violations. Corporate intranets generally have exceedingly strict AUPs, and some infractions can be punished by dismissal; AUPs are in fact very common. The remarkable feature of the old NSFNET AUP was NOT the number of times it was violated, but rather the lengths to which an extraordinary number of decent corporate and private netizens went to ensure its observance. I promulgated it and I know. *Whenever* there is public funding of *any* activity, an AUP is attached to oulaw the misuse of taxpayer money. Internet II is not exempt. Some say that acceptance of an AUP is too high a price to pay in circumscription of personal liberty for the benefits of public funding. I understand and respect that view, but I don't agree. -s Stephen Wolff Business Development ciscoSystems 202 337 1213

1 0

RE: Internet II is coming...
by Jim Browning 11 Oct '96

11 Oct '96

>From: Sean Doran[SMTP:smd@cesium.clock.org] >Sent: Wednesday, October 09, 1996 5:11 PM > >WRT your list, I don't think that either videoconferencing >or connectivity to the PSTN is a particularly interesting thing >to do with the Internet, in that I don't see either as being >a fundamentally important Internet application. I think that any and all services my customers want are at least interesting, if not extremely important. I also think it is very interesting to leverage those customers' local loop circuit investments by providing as many services over the same infrastructure as possible. >Various people, including a Sprint Canada reseller, will tell >you that ALGs exist that convert various voice-over-Internet >things into signals that can be dealt with by PBXes and other >telephony equipment. There are at least two SS7<->IP gateways >in development with what appears to be short times-to-market ahead. > >One of these is being built along the lines of what PNO like an >RBOC or a PTT or a cable company might want to deploy in their >own facilities. > >These are neat hacks which will do fun things to international >and national long-distance tariffs worldwide, in that you may >suddenly have two QOSes for voice: one which is a side-effect >of IP connectivity and which is of iffy-to-tolerable-to-good >quality, and one which has metered, often distance-and-time-of-day- >sensitive charging and a generally predicatably very-good-to-excellent >quality. And I have in fact seen a demonstration of "Internet Phone" over a cable modem, through a gateway, and into the PSTN. I would of course like my customers to have access to _both_ levels of quality, and all of it through the same infrastructure. I'm one of the evil, greedy, money grubbing capitalists who, while willing to invest millions of dollars in network infrastructure, want to sell its use in a variety of ways. >We're now down to 155Mbps medium speed to the desktop. >I imagine that we can avoid arguing about the small details >of cell tax and the like, and start looking at Gbps ethernet >and other forthcoming LAN issues. They really are LAN issues; >the Internet vs global ATM really isn't a factor here, nor is >really IP vs ATM on a protocol-to-protocol basis. I accepted your 'today' challenge, but agree about the LAN orientation. I don't agree that there must be a "vs" between IP and ATM, cell taxes and the impact of either infinite or perpetually constricted bandwidth notwithstanding. >| Oh, and by the way, given that the local loop provider has OC-48 SONET >| provisioned to this particular location, we could just as easily have >| provisioned the connection to our backbone at OC12 as opposed to OC3. Did >| I miss the Cisco announcement of an OC12 IP-SONET card? > >You may wish to discuss an NDA presentation on the forthcoming >generation of routers from each of Cisco, Juniper and Bay Networks. Again, my post was based upon your assertion that this could be done today. I sincerely hope that a new generation of routers is forthcoming asap that can match ATM speeds. >So, just to clarify things, I don't discount ATM as a technology >entirely, I just think that its time as something between IP and >SONET for big-I Internet applications (other than *maybe* for >customer aggregation and the like), is ticking away... Time will tell, and meanwhile, there is ample room for a variety of network architectures and business models. Clearly the arguments on both sides have resulted in new technology which falls somewhere in between. This can only be A Healthy Thing (TM). -- Jim Browning

3 2

Re: Internet II is coming...
by Paul Ferguson 11 Oct '96

11 Oct '96

Potentially, yes. - paul At 05:20 PM 10/9/96 -0400, Fletcher E Kittredge wrote: > >So a 10Mbps fiber ATM circuit translates, after the cell tax and >packet stredding, to 6Mbps on a real link. Is this not slow enough to >be a bottleneck? > >regards, >fletcher > >

1 0

RE: Internet II is coming...
by Vadim Antonov 10 Oct '96

10 Oct '96

Sean Doran <smd(a)cesium.clock.org> wrote: >Moreover, you will note that lightweight private lines is not >something that can be done only by ATM... It can be done with IPv4 as-is. Build an IP-in-IP tunnel. Have the gateways doing encapsulation to slap higher priority on the packets, so there won't be any loss. Place a rate limitation on traffic entering the tunnel. Much better than CBR SVC in many ways. I know of at least one company which plans to offer that kind of functionality, at wire speed :) --vadim

1 0