- NANOG - lists.nanog.org

Re: QoS/CoS interest
by garyz＠savvis.com 22 May '97

22 May '97

Would be nice to start using these hooks, if everyone would get on board. We would support any movement in this direction, it would help us with some of our QoS/CoS issues. We have talked to a few of the larger providers without much luck. I like the sanity of using the precedence bits. Gary Zimmerman V.P. of Network Engineering Savvis Communications Corp. email: garyz(a)savvis.com http://www.savvis.com Office: 314.719.2423 Address: 7777 Bonhomme Suite 1000 St. Louis, MO 63105 ---------- > From: Paul Ferguson <pferguso(a)cisco.com> > To: peter(a)injersey.com > Cc: nanog(a)merit.edu > Subject: Re: QoS/CoS interest > Date: Wednesday, May 21, 1997 8:16 PM > > At 06:06 PM 05/21/97 -0400, Peter wrote: > > >Seems this would be easy to do in netflow switching. Also sounds a lot > >like FECN/BECN in frame relay. > > > > Sounds alot to me like sanity -- using the existing hooks > in IPv4 to deliver differentiated classes of service by using > the IP precedence bits in the TOS portion of the IP header. > > - paul

1 0

Re: QoS/CoS interest
by Scott Bradner 22 May '97

22 May '97

Bill, -- However, I always alike to be a bit contrarian and point out that QoS or Multicast may never be needed because of the explosive growth of fiber bandwidth. I believe, in the future, it will be a lot easier and cheaper to deploy bandwidth rather than manage complex router/switch technology to support QoS/CoS. -- I think this is a myth, at least for now - the production of fiber in the world was 1.25 million fiber miles short of demand last year and is expeced to be about the same this year. IN addition rights of way are getting rather hard to obtain. WDM will be a great help but it does not cause fiber to be run into Seattle. The cost of putting down fiber is still very high (particularally across the ponds) and even with WDM the cost of bandwidth will continue to be high. Yes the cost should drop (assuming that it is not kept high by other things like telco or government policies) but we ain't going to see "bandwidth too cheap to measure" (as someone put it on comm-priv 2 years ago) in any timeframe that will let the Internet community avoid looking at QoS as a very real issue. Scott .

2 1

Re: QoS/CoS interest
by pkavi＠pcmail.casc.com 22 May '97

22 May '97

Paul, Even if QoS/CoS is desirable, is its widescale deployment possible given the current Internet? Here are some of the issues stopping QoS: 1. No billing available: If there is no billing system which can determine when people use QoS capabilities, than how can it be deployed widely and effectively. 2. No support on routing infrastructure: Routers today have several problems: * No policing at ingress: You can't have QoS unless you can limit how much traffic enters the network, and discard, or at least mark the excess traffic. * No effective BW reservation mechanism: RSVP may solve this in the future, but it's not deployed now. * No effective Class of Service mechanism: * Packet/sec processing limits: Even if the Internet infrastructure knows how to do all of the above, (reserve, police, mark, and queue according to QoS parameters), it will still fail if the box cannot satisfy the packet rates coming to it. Both your company and mine are working on ways to fix these limitations. But widescale QoS capabilities requires that every router between a source and destination has these capabilities. QoS guarantees maybe possible within a single AS that uses a router infrastructure with Frame Relay or ATM switching capabilities (which tend to have implicit support for all of the above listed features). Of course, this leads to the next problem, namely: 3. No support across ASes. First of all, BGP provides no QoS metrics. So there is no way to determine if a particular AS should even be considered in setting up a QoS path. Second, while a single AS could be upgraded to QoS-capable equipment, a forklift upgrade across the Internet to QoS-capable equipment won't happen anytime soon. I think the Internet will stay at best-effort for quite a while, with small pockets of QoS in special situations. Prabhu ______________________________ Reply Separator _________________________________ Subject: QoS/CoS interest Author: Paul Ferguson <pferguso(a)cisco.com> at SMTPLINK Date: 5/21/97 3:04 PM I'd like to get a feel for what the temperature level is in the ISP community for this issue -- I have a vested [personal] interest in understanding what your understanding & implementation plans are in this arena. Please send your thoughts, requirements, bitches, etc., to me personally, or preferably, to the NANOG list; it is a greater audience than myself that wishes to understand these issues. Thanks! - paul

1 0

Re: Root Name Server Confederations
by Sean M. Doran 22 May '97

22 May '97

You don't want my code. It has too many parentheses in it. However, I can send you a crisis (well, one seems to be coming without my help at all, actually), which might prompt more talented programmers to come up with something that fixes the current silly politics and improves the utility of the DNS for what it's actually potentially good for. Sean.

1 0

Re: Root Name Server Confederations
by smd＠clock.org 22 May '97

22 May '97

Paul A Vixie <paul(a)vix.com> writes: > makes one long for the days of HOSTS.TXT. which was at least reliable and > coherent. maybe it'll make a comeback in the form of everybody-run-their-own > "." server. Hey, I take it back, a PVM-must-die thread on NANOG hasn't happened in a long time. Maybe NANOG isn't _utterly_ useless (only mostly)... > i guess i've said this about 900 times now. DNS is a coherent, distributed > database. don't do anything that will make it less > coherent. Kill it outright. Build something that doesn't use the unreliable datagram protocol with all of its wonderful deficiencies, perhaps leaving the description of the database itself to the implementors (concentrating on protocol instead), and ideally avoiding the in-addr.arpa botch and the difficulties of scaling really large zones. Denninger and company are wonderful because they are forcing some issues which go right back to the days of HOSTS.TXT and the ARPA zone. I seriously hope they are able to forestall the insufficient bandaid approaches being proposed by more moderate technical people. > start back in on things that will make the same name mean something different > (or become meaningless) depending on where you're standing when you resolve > it. But this is a really neat idea; have the DNS or its follow-on(s) represent _services_ and the utility of having this feature is fairly obvious. This is to some extent how NATs work, after all. A complaint about how '<port>.<zone>.<zone>...' can lead to confusion is a complaint about the lack of directory services through one can find a service in the first place. Sean.

2 1

Re: Root Name Server Confederations
by smd＠clock.org 22 May '97

22 May '97

Marc Hurst <mhurst(a)fastlane.ca> writes: > If I lose I WILL eat crow at the next NANOG meeting (bbq flavoured if > possible). So here I am in the terminal room of the RIPE meeting and realizing how completely indescribable my current lack of motivation for going to yet another NANOG meeting is, particularly one in the summertime in Florida. I do hear, however, that alligators can be fun to feed. Despite that, could I humbly suggest that the NANOG organization and mailing list be wound down? I think their work is done. Sean.

1 0

IOPS website is now available
by Michael Dillon 21 May '97

21 May '97

I just finished reading through the http://www.iops.org website. Have a look... Michael Dillon - Internet & ISP Consulting http://www.memra.com - E-mail: michael(a)memra.com The bottom line is track record. Not track tearing. Not track derailing. But pounding the damn dirt around the track with the rest of us worms. -- Randy Bush

1 0

Re: QoS/CoS interest
by sob＠academ.com 21 May '97

21 May '97

As I am involved in deploying a GigaPOP here in Houston and attaching it to SESQUINET as well as the I2 infrastructure, I am working with QoS issues (at an architectural level) regularly. I do expect to do some deployment as this year comes to an end, but there are many issues (some technical and some not) that will affect this roll out. -- Stan | Academ Consulting Services |internet: sob(a)academ.com Olan | For more info on academ, see this |uucp: {mcsun|amdahl}!academ!sob Barber | URL- http://www.academ.com/academ |Opinions expressed are only mine.

1 0

Re: What have you done for me lately?
by Eric Germann 21 May '97

21 May '97

At 12:47 AM 5/21/97 CDT, Joseph T. Klein wrote: >I belive it will take a volunteer effort ... > >--- On Wed, 21 May 1997 00:56:00 -0400 (EDT) Todd Graham Lewis <lists(a)reflections.eng.mindspring.net> wrote: > >Yeah ... the stuff you can hardly see between the flames. We need >a few people to gleen the list and turn it into a web page. > http://www.cctec.com/maillists/nanog I don't douse the flame wars, however, it is searchable and is a complete history, so you can search around the bombs. Updated nightly... Eric ============================================================================ ==== Eric Germann Computer and Communications Technologies ekgermann(a)cctec.com Van Wert, OH 45891 Phone: 419 968 2640 http://www.cctec.com Fax: 419 968 2641 Network Design, Connectivity & System Integration Services A Microsoft Solution Provider

1 0

Re: Root Name Server Confederations
by Paul A Vixie 21 May '97

21 May '97

Danger. I'm still ranting. Don't read this. Someone asked in private e-mail... > What's to keep someone from adding some of those "bogus" root servers into > a cache file? What harm would it do? There seems to be something I'm > missing? If I had users that wanted access to .web, couldn't I add those > into the cache? which .web do you mean? the one iahc is creating, the one amblin tried to create, or one of the dozen that will be created next week if piracy proves to be a useful way to add top level domain names? (did the ISP explosion teach you guys NOTHING about copycatting?) and when one of your customers wants foo.web and the other wants bar.web and you can't satisfy both because they're in different .web's? and when one of your customers wants foo.web and the other wants foo.web except they are different web sites depending on which root servers you point at? ultimately the answer is: there is one "." and anybody who tells you otherwise is trying to sell you something that you do NOT want to pay for. (keep your hand on your wallet.) hell, i may create a .web zone of my own if the money is THAT easy to get. in fact, let's ALL create .web and then let's create usenet:alt.gtlds.web and use it to synchronize entries therein. that way when someone creates foo.web (in exchange for a service fee from the FOO company), everyone else who has a .web domain can add the same delegation (even though they did NOT get a service fee from the FOO company) and there will be no conflicts. what's that you say? impractical? well if it won't work for .web, WHAT MAKES YOU THINK IT WILL WORK FOR "." ??? the issues are clear cut. pirates are loose on the seas you're surfing, and most of them want your money. some of them want power. a few of them just want some personal recognition -- it hurts their feelings that they weren't consulted on whether there should or shouldn't have been an IAHC. none of these pirates give a rat's a** whether your customers all get the same FOO when they ask for FOO.WEB. in fact at least one of them thinks this would be a feature since each FOO _deserves_ a chance to be FOO.WEB. makes one long for the days of HOSTS.TXT. which was at least reliable and coherent. maybe it'll make a comeback in the form of everybody-run-their-own "." server. i guess i've said this about 900 times now. DNS is a coherent, distributed database. don't do anything that will make it less coherent. don't lets start back in on things that will make the same name mean something different (or become meaningless) depending on where you're standing when you resolve it.

3 2