- Test - lists.nanog.org

Re: TCP congestion
by Jay Hennigan 13 Jul '07

13 Jul '07

Philip Lavine wrote: > Can someone explain how a TCP conversation could degenerate into congestion av oidance on a long fat pipe if there is no packet/segment loss or out of order se gments? > > Here is the situation: > WAN = 9 Mbps ATM connection between NY and LA (70 ms delay) > LAN = Gig Ethernet > Receiver: LA server = Win2k3 > Sender: NY server = Linux 2.4 > Data transmission typical = bursty but never more that 50% of CIR > Segment sizes = 64k to 1460k but mostly less than 100k > > Typical Problem Scenario: Data transmission is humming along consistently at 2 Mbps, all of a sudden transmission rates drop to nothing then pickup again afte r 15-20 seconds. Prior to the drop off (based on packet capture) there is usuall y a DUP ACK/SACK coming from the receiver followed by the Retransmits and conges tion avoidence. What is strange is there is nothing prior to the drop off that w ould be an impetus for congestion (no high BW utilization or packet loss). > > Also is there any known TCP issues between linux 2.4 kernel and windows 2003 S P1? Mainly are there issues regarding the handling of SACK, DUP ACK's and Fast R etransmits. > > Of course we all know that this is not a application issue since developers ma ke flawless socket code, but if it is network issue how is caused? Duplex mismatch on an intermediate ethernet segment? -- Jay Hennigan - CCIE #7880 - Network Engineering - jay(a)impulse.net Impulse Internet Service - http://www.impulse.net/ Your local telephone and internet company - 805 884-6323 - WB6RDV

1 0

Re: peter lothberg's mother slashdotted
by Randy Bush 13 Jul '07

13 Jul '07

d.org> <Pine.LNX.4.64.0707122342300.29164(a)uplift.swm.pp.se> <4696AB85.1030407@in oc.net> <Pine.LNX.4.64.0707130039220.29164(a)uplift.swm.pp.se> <4696B668.2080807@i noc.net> <4696C8DA.3070101(a)bogus.com> <4696D017.7010100(a)inoc.net> In-Reply-To: <4696D017.7010100(a)inoc.net> X-Enigmail-Version: 0.95.1 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Sender: owner-nanog(a)merit.edu Precedence: bulk Errors-To: owner-nanog(a)merit.edu X-Loop: nanog X-Junkmail-Status: score=10/50, host=mozart.merit.edu X-Junkmail-SD-Raw: score=unknown, refid=str=0001.0A090204.4696D52A.0095:SCGAP167720,ss=1,fgs=0, ip=198.108.1.26, so=2006-09-22 03:48:54, dmn=5.3.14/2007-05-31 Status: RO X-Status: X-Keywords: X-UID: 47 > you're not going to find any 40GB capable CPE now or in the > foreseeable future that's going to be affordable for the residence. i would agree if we had not once said that about a few meg per sec. randy

1 0

None
by None 13 Jul '07

13 Jul '07

NbwMvYItZ/yhmeizQs5PAXpb76Rzx2GXeOm1BvaVus1mJtLqxcbjpnZwZRpGxMciLmLW9sqcz4rxon6L hU7Bletqm3+UTqGTuzK50=; Received: from [172.17.67.246] (unverified [12.104.158.98]) by mx1-b.inoc.net (build v7.5.2) with ESMTP id 127880137-1941382 for <nanog(a)nanog.org>; Fri, 13 Jul 2007 01:06:32 +0000 (UTC) Message-ID: <4696D017.7010100(a)inoc.net> Date: Thu, 12 Jul 2007 21:06:31 -0400 From: Robert Blayzor <rblayzor(a)inoc.net> Organization: INOC, LLC User-Agent: Thunderbird 2.0.0.4 (Windows/20070604) MIME-Version: 1.0 To: nanog(a)nanog.org Subject: Re: peter lothberg's mother slashdotted References: <12997.1184264272(a)sa.vix.com> <20070712133910.B8475(a)sprockets.gibbar d.org> <Pine.LNX.4.64.0707122342300.29164(a)uplift.swm.pp.se> <4696AB85.1030407@in oc.net> <Pine.LNX.4.64.0707130039220.29164(a)uplift.swm.pp.se> <4696B668.2080807@i noc.net> <4696C8DA.3070101(a)bogus.com> In-Reply-To: <4696C8DA.3070101(a)bogus.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-nanog(a)merit.edu Precedence: bulk Errors-To: owner-nanog(a)merit.edu X-Loop: nanog Joel Jaeggli wrote: >> How practical is it really also that you need CRS-1 at the residence for >> this. I agree with Sean. Since for most people the line card alone >> costs more than the house. :-) > > 40Gb/s per slot routers are not that rare at this point. So the notion > that you need a crs-1 in order to capitalize on 40G is I think > demonstrably false. As I said, "at the residence". I could care less if it's a CRS-1 or Acme router. My point is, you're not going to find any 40GB capable CPE now or in the foreseeable future that's going to be affordable for the residence. I know the intention of the article was to demonstrate the technological advantage of fiber optic networking vs. other technologies, but to mask the article around something that seems to indicate that "40GB to the home" is somehow the current benchmark is incredibly unrealistic. -Robert

1 0

Re: peter lothberg's mother slashdotted
by Joel Jaeggli 13 Jul '07

13 Jul '07

d.org> <Pine.LNX.4.64.0707122342300.29164(a)uplift.swm.pp.se> <4696AB85.1030407@in oc.net> <Pine.LNX.4.64.0707130039220.29164(a)uplift.swm.pp.se> <4696B668.2080807@i noc.net> In-Reply-To: <4696B668.2080807(a)inoc.net> X-Enigmail-Version: 0.94.1.1 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit X-Virus-Scanned: ClamAV 0.90.2/3653/Thu Jul 12 22:15:16 2007 on nagasaki.bogus.c om X-Virus-Status: Clean Sender: owner-nanog(a)merit.edu Precedence: bulk Errors-To: owner-nanog(a)merit.edu X-Loop: nanog X-Junkmail-Status: score=10/50, host=mozart.merit.edu X-Junkmail-SD-Raw: score=unknown, refid=str=0001.0A090203.4696CA12.0111:SCGAP167720,ss=1,fgs=0, ip=198.108.1.26, so=2006-09-22 03:48:54, dmn=5.3.14/2007-05-31 Status: RO X-Status: X-Keywords: NonJunk X-UID: 45 Robert Blayzor wrote: > How practical is it really also that you need CRS-1 at the residence for > this. I agree with Sean. Since for most people the line card alone > costs more than the house. :-) 40Gb/s per slot routers are not that rare at this point. So the notion that you need a crs-1 in order to capitalize on 40G is I think demonstrably false. The amount of link aggregation being done at 10Gb/s would tend to indicate that need for > 40Gb/s links is becoming acute enough to make it a commercially worthwhile problem to solve if not cheap for the early adopters. As consumers we would all benefit from standardization of 40 and 100Gb/s ethernet phy. > -Robert >

1 0

None
by None 13 Jul '07

13 Jul '07

d.org> <Pine.LNX.4.64.0707122342300.29164(a)uplift.swm.pp.se> <4696AB85.1030407@in oc.net> <Pine.LNX.4.64.0707130039220.29164(a)uplift.swm.pp.se> <4696B668.2080807@i noc.net> Mime-Version: 1.0 (Apple Message framework v752.3) Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed Message-Id: <ACFC7F1E-B1F2-4519-A898-33F787CC3D06(a)multicasttech.com> Cc: nanog(a)nanog.org Content-Transfer-Encoding: 7bit From: Marshall Eubanks <tme(a)multicasttech.com> Subject: Re: peter lothberg's mother slashdotted Date: Thu, 12 Jul 2007 19:51:49 -0400 To: Robert Blayzor <rblayzor(a)inoc.net> X-Mailer: Apple Mail (2.752.3) Sender: owner-nanog(a)merit.edu Precedence: bulk Errors-To: owner-nanog(a)merit.edu X-Loop: nanog On Jul 12, 2007, at 7:16 PM, Robert Blayzor wrote: > > Mikael Abrahamsson wrote: >> I have to disagree, considering the amount of people I've had to >> convice that this really is a single 50GHz wave using 40G per >> second over DWDM system designed for 10G and that it was router >> "LC - optical amplification - LC" > > > Well yes, the technology is quite cool, I'm not debating that. The > fact that it works at his mothers house (residential) doesn't > really have any significance. I don't think you have anyone out > there debating against caping out the upper limits of bandwidth on > what we can get over a single strand of glass. > > I look at this article and say "so what". Now you've gotten the > packets to the house. Now what? Wait 10, 15 or 20 years while > companies try to find ways to get content do you? 1500 HD > channels, thats a lot of TV sets. You can have all the bandwidth > you want, but but until there an actual use for 40Gbps at the home > all you're going to have is the worlds fastest BitTorrent or on net > porn PPV network. Uncompressed HD TV (typically 1.54 Gbps / channel) comes to mind, for which there are applications, albeit typically not residential so far. See, e.g., http://www.gloriad-kr.org/hdtv/ http://ultragrid.east.isi.edu/2005-02-APAN.pdf Regards Marshall > > How practical is it really also that you need CRS-1 at the > residence for this. I agree with Sean. Since for most people the > line card alone costs more than the house. :-) > > -Robert

1 0

None
by None 12 Jul '07

12 Jul '07

h Microsoft SMTPSVC(6.0.3790.1830); Fri, 13 Jul 2007 00:21:36 +0100 Message-ID: <4696B6EC.3050305(a)ukbroadband.com> Date: Fri, 13 Jul 2007 00:19:08 +0100 From: Leigh Porter <leigh.porter(a)ukbroadband.com> User-Agent: Thunderbird 2.0.0.4 (Macintosh/20070604) MIME-Version: 1.0 To: Brandon Butterworth <brandon(a)rd.bbc.co.uk> CC: nanog(a)nanog.org Subject: Re: peter lothberg's mother slashdotted References: <200707122141.WAA06503(a)sunf10.rd.bbc.co.uk> In-Reply-To: <200707122141.WAA06503(a)sunf10.rd.bbc.co.uk> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-OriginalArrivalTime: 12 Jul 2007 23:21:36.0952 (UTC) FILETIME=[64B08380:01C7C4 DB] Sender: owner-nanog(a)merit.edu Precedence: bulk Errors-To: owner-nanog(a)merit.edu X-Loop: nanog X-Junkmail-Status: score=10/50, host=mozart.merit.edu X-Junkmail-SD-Raw: score=unknown, refid=str=0001.0A090205.4696BF0F.0112:SCGAP167720,ss=1,fgs=0, ip=198.108.1.26, so=2006-09-22 03:48:54, dmn=5.3.14/2007-05-31 Status: RO X-Status: X-Keywords: NonJunk X-UID: 41 I see a global demand for perhaps 5 CRS-1s ;-) Brandon Butterworth wrote: >> Wouldn't residential fiber be >> expected to radiate out from neighborhood break-out boxes, or at the >> longest from a central office in the middle of town, rather than having >> some central point where enough individual strands of fiber converged to >> serve everybody in a 2,000 kilometer radius? >> > > CRS-1 so expensive there'll only be one per country. > > brandon >

1 0

Re: some musings on PI v. PA
by Seth Mattinen 12 Jul '07

12 Jul '07

e.php X-Rollernet-Submit: Submit ID 2c4f.4696bed2.ba6e4.0 Sender: owner-nanog(a)merit.edu Precedence: bulk Errors-To: owner-nanog(a)merit.edu X-Loop: nanog X-Junkmail-Status: score=10/50, host=mozart.merit.edu X-Junkmail-SD-Raw: score=unknown, refid=str=0001.0A090203.4696C70E.0163:SCGAP167720,ss=1,fgs=0, ip=198.108.1.26, so=2006-09-22 03:48:54, dmn=5.3.14/2007-05-31 Status: RO X-Status: X-Keywords: NonJunk X-UID: 44 David Meyer wrote: > OTOH, if I announce PI space, "switching to the new path" > is controlled by the announcement/withdrawal of the PI > prefix, and can happen much closer to the source. So in > this sense aggregation breaks a certain kind of "path > selection". I think we all realize that there is no free > lunch, and that this is a property (such as it is) of the > fact that aggregation throws away information in the > interest of computability (a standard technique). > > So folks are using PI for reasons other than the > well-known standard laundry list. In particular, > advertising PI space can cause the "switch" to an > alternate path during an outage to happen much closer to > the source and some providers find this to be a desirable > property. > > I am interested in hearing about anyone who is relying on > this property of PI, or any other comments on what I've > said above. > I always thought that *was* one of the major reasons to use PI space - no aggregation of your space into another announcement. Not having to renumber is ancillary to me. ~Seth

1 0

None
by None 12 Jul '07

12 Jul '07

f Date: Thu, 12 Jul 2007 16:17:32 -0700 From: David Meyer <dmm(a)1-4-5.net> To: nanog(a)nanog.org Cc: dino(a)cisco.com Subject: some musings on PI v. PA Message-ID: <20070712231732.GA17524(a)1-4-5.net> MIME-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="IS0zKkzwUGydFO0o" Content-Disposition: inline X-public-key: http://www.1-4-5.net/~dmm/public-key.asc X-gpg-fingerprint: 2409 8B50 B389 A307 BA5C 2A16 3918 03D6 A099 D8A7 X-philosophy: "I laugh and I cry, and I'm haunted by...Things I never meant nor wished to say" -- Bob Dylan, "When The Deal Goes Down" User-Agent: Mutt/1.5.13 (2006-08-11) Sender: owner-nanog(a)merit.edu Precedence: bulk Errors-To: owner-nanog(a)merit.edu X-Loop: nanog X-Junkmail-Status: score=10/50, host=mozart.merit.edu X-Junkmail-SD-Raw: score=unknown, refid=str=0001.0A090209.4696BCA8.0083:SCGAP167720,ss=1,fgs=0, ip=198.108.1.26, so=2006-09-22 03:48:54, dmn=5.3.14/2007-05-31 Status: RO X-Status: X-Keywords: NonJunk X-UID: 40 --IS0zKkzwUGydFO0o Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable I've been thinking about a benefit of PI addressing that I have not seen discussed on this list or others (at least recently). In particular, PI addressing enables a certain kind of "path selection" that might not be easy=20 (or possibly desirable) to retain in any of the the LOC/ID split schemes we have been discussing. This is in contrast to all of the standard reasons for wanting PI space (e.g., I don't want to renumber, ...). =20 Consider the following simple scenario: I'm a multihomed stub (I don't transit packets between my two upstreams). Further, I have PA delegations from each of my upstreams. Now, I'm corresponding with a remote site using addressing out of one of the PA blocks, call it X. Now, my link to the ISP aggregating X breaks. A packet destined for X will then travel very close to my site before learning that the link is down, possibly too far to be rerouted. And BTW, if I advertise X to my other upstream, then my advertisement of X has the same cost to the routing system as a PI advertisement (X is effectively PI). This problem is common to all (I think) of the schemes that seek to improve/optimize aggregatability in the core. For example: (a). In the 8+8/GSE case, the problem is that the packet will follow the RG in the src address all the way to the "end" of the path, that is, to the ISP that can forward it to the site. You don't learn that the site can't receive the packet until that point, and there is no way to reroute it. =20 (b). The situation is similar with PA space, since the fact that the link at the "end" of the path might be down is hidden in the aggregate. You don't learn this until you are close to the "end" of the path, and there may be no way to reroute the packet.=20 (c). The situation is similar for the map/encap schemes (e.g., LISP), since one of primary goals of map/encap is to enable better topological aggregation. =20 OTOH, if I announce PI space, "switching to the new path" is controlled by the announcement/withdrawal of the PI prefix, and can happen much closer to the source. So in this sense aggregation breaks a certain kind of "path selection". I think we all realize that there is no free lunch, and that this is a property (such as it is) of the fact that aggregation throws away information in the interest of computability (a standard technique).=20 So folks are using PI for reasons other than the well-known standard laundry list. In particular, advertising PI space can cause the "switch" to an alternate path during an outage to happen much closer to the source and some providers find this to be a desirable property. I am interested in hearing about anyone who is relying on this property of PI, or any other comments on what I've said above. Oh, and BTW, none of this to say that we shouldn't be moving forward with the various alternatives we've been discussing (e.g., 8+8/GSE, LISP, ...); quite the contrary. Rather, my question is really about revisiting assumptions, requirements, and tradeoffs.=20 Thnx, Dave --IS0zKkzwUGydFO0o Content-Type: application/pgp-signature; name="signature.asc" Content-Description: Digital signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.3 (GNU/Linux) iD8DBQFGlraMORgD1qCZ2KcRApWxAJ4thXUsuh4TdnXy2LytbzK4mRCEmwCdEE8A fZicCEufrqixFXp0F1ZbxbA= =0Puq -----END PGP SIGNATURE----- --IS0zKkzwUGydFO0o--

1 0

None
by None 12 Jul '07

12 Jul '07

pj306VStB0wybEWJRlCkn567O0jgwoVlC32nil56rSvaI+PMgHqshilvAA4hMKX5WamEVzKyTXMMpNDQ rl7emMULpFA9bSzgS1e+A=; Received: from [172.17.67.246] (unverified [12.104.158.98]) by mx0-b.inoc.net (build v7.5.2) with ESMTP id 127015704-1941382 for <nanog(a)nanog.org>; Thu, 12 Jul 2007 23:16:57 +0000 (UTC) Message-ID: <4696B668.2080807(a)inoc.net> Date: Thu, 12 Jul 2007 19:16:56 -0400 From: Robert Blayzor <rblayzor(a)inoc.net> Organization: INOC, LLC User-Agent: Thunderbird 2.0.0.4 (Windows/20070604) MIME-Version: 1.0 To: nanog(a)nanog.org Subject: Re: peter lothberg's mother slashdotted References: <12997.1184264272(a)sa.vix.com> <20070712133910.B8475(a)sprockets.gibbar d.org> <Pine.LNX.4.64.0707122342300.29164(a)uplift.swm.pp.se> <4696AB85.1030407@in oc.net> <Pine.LNX.4.64.0707130039220.29164(a)uplift.swm.pp.se> In-Reply-To: <Pine.LNX.4.64.0707130039220.29164(a)uplift.swm.pp.se> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-nanog(a)merit.edu Precedence: bulk Errors-To: owner-nanog(a)merit.edu X-Loop: nanog X-Junkmail-Status: score=10/50, host=mozart.merit.edu X-Junkmail-SD-Raw: score=unknown, refid=str=0001.0A09020A.4696BBE6.0003:SCGAP167720,ss=1,fgs=0, Mikael Abrahamsson wrote: > I have to disagree, considering the amount of people I've had to convice > that this really is a single 50GHz wave using 40G per second over DWDM > system designed for 10G and that it was router > > "LC - optical amplification - LC" Well yes, the technology is quite cool, I'm not debating that. The fact that it works at his mothers house (residential) doesn't really have any significance. I don't think you have anyone out there debating against caping out the upper limits of bandwidth on what we can get over a single strand of glass. I look at this article and say "so what". Now you've gotten the packets to the house. Now what? Wait 10, 15 or 20 years while companies try to find ways to get content do you? 1500 HD channels, thats a lot of TV sets. You can have all the bandwidth you want, but but until there an actual use for 40Gbps at the home all you're going to have is the worlds fastest BitTorrent or on net porn PPV network. How practical is it really also that you need CRS-1 at the residence for this. I agree with Sean. Since for most people the line card alone costs more than the house. :-) -Robert

1 0

None
by None 12 Jul '07

12 Jul '07

) by chopin.merit.edu (MOS 3.8.2-GA) with ESMTP id AFJ88688; Thu, 12 Jul 2007 19:13:53 -0400 (EDT) Received: from [66.117.52.128] (unknown [66.117.52.128]) by srv013.carpathiahost.net (Postfix) with ESMTP id 3509193780 for <nanog(a)merit.edu>; Thu, 12 Jul 2007 19:13:53 -0400 (EDT) Message-ID: <4696B59E.9040900(a)kl.net> Date: Thu, 12 Jul 2007 19:13:34 -0400 From: Kevin Loch <kloch(a)kl.net> User-Agent: Thunderbird 1.5.0.9 (X11/20070127) MIME-Version: 1.0 To: nanog <nanog(a)merit.edu> Subject: Re: TCP congestion References: <796919.42253.qm(a)web30811.mail.mud.yahoo.com> <922B68D8-BCB4-40B8-88 5A-2459BDFEDA34(a)muada.com> In-Reply-To: <922B68D8-BCB4-40B8-885A-2459BDFEDA34(a)muada.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-nanog(a)merit.edu Precedence: bulk Errors-To: owner-nanog(a)merit.edu X-Loop: nanog X-Junkmail-Status: score=10/50, host=mozart.merit.edu X-Junkmail-SD-Raw: score=unknown, Iljitsch van Beijnum wrote: > How exactly are you going to get out-of-order packets over a single link? There is a once popular router that has been known to do that in some configurations due to multiple paths within the device. - Kevin

1 0