April 2001 - Test - lists.nanog.org

RE: jumbo frames
by Roeland Meyer 26 Apr '01

26 Apr '01

You are correct and my original answer stated such. The original question didn't differentiate. It simply asked who was using jumbo packets. > -----Original Message----- > From: Tony Hain [mailto:alh-ietf@tndh.net] > Sent: Thursday, April 26, 2001 12:55 PM > To: Roeland Meyer; ALH-IETF; John Fraizer; Paul Lantinga > Cc: nanog(a)merit.edu > Subject: RE: jumbo frames > > > > Roeland you are talking about jumbo frames from the end > system lan, while > John is talking about only using the jumbo frames between the > routers. My > point was that in John's environment the packets will all be > 1500 since the > packets are restricted to that size just to get to the router > with the GE > interface. I understand that there are perf gains as long as > the entire path > supports the larger packets, but I don't understand the claim > that having a > bigger pipe in the middle helps. > > Tony > > -----Original Message----- > From: owner-nanog(a)merit.edu [mailto:owner-nanog@merit.edu]On Behalf Of > Roeland Meyer > Sent: Thursday, April 26, 2001 12:13 PM > To: 'alh-ietf(a)tndh.net'; John Fraizer; Paul Lantinga > Cc: nanog(a)merit.edu > Subject: RE: jumbo frames > > > > From: Tony Hain [mailto:alh-ietf@tndh.net] > > Sent: Thursday, April 26, 2001 11:47 AM > > > > April 26, 2001 9:29 AM John Fraizer wrote: > > > We only have jumbo frames enabled on router<->router links. > > The GigE > > > ports facing the aggregation switches runs standard 1500 MTU. > > > > Hence my original question. Packets across the GE will be > > 1500 unless you > > are packing them. > > > > April 25, 2001 8:10 PM John Fraizer wrote: > > > Partially because I can. Partially because there seems to be a > > > performance increase when you start stuffing the pipe. > > > > Assuming you are just passing the packets as received from > > the aggregation > > switch, this would only happen if your router hardware was better at > > managing jumbo buffer allocations than 1500B ones. Clearly it > > will waste > > large chunks of memory, so do you have measurements to show > the actual > > performance increase? > > This depends on the type of traffic. We use it to enhance > performance of the > data tier. We've jiggered the TCP/IP stacks for ~4500 byte > packets and have > knee-capped the slow-start algorithm (which doesn't make > sense in a pure > switched environment anyway). What we then wind up with, is a > dedicated data > LAN between our application servers and the Oracle database > servers. It > comes out to about an order of magnitude increase in > performance and SQL > query responsiveness. At first we went to jumbo packets. We saw such a > radical improvement that we started investigating and found > the slow-start > issue. Jumbo packets are one way around the slow-start > problem if you can't > jigger the stack itself. Most of the queries are reasonably > short so we saw > some serious improvements by killing the slow-start. If you > can modify your > IP stack then it still pays to use jumbo packets because you > reduce the > overhead on the wire. > > We got sufficient performance improvement that we were able > to defer GigE > implementation, at some sites. Those sites are using switched > FDX 100baseTX. > >

1 0

RE: jumbo frames
by Richard A. Steenbergen 26 Apr '01

26 Apr '01

On Thu, Apr 26, 2001 at 12:12:33PM -0700, Roeland Meyer wrote: > > This depends on the type of traffic. We use it to enhance performance > of the data tier. We've jiggered the TCP/IP stacks for ~4500 byte > packets and have knee-capped the slow-start algorithm (which doesn't > make sense in a pure switched environment anyway). What we then wind > up with, is a dedicated data LAN between our application servers and > the Oracle database servers. It comes out to about an order of > magnitude increase in performance and SQL query responsiveness. At > first we went to jumbo packets. We saw such a radical improvement that > we started investigating and found the slow-start issue. Jumbo packets > are one way around the slow-start problem if you can't jigger the > stack itself. Most of the queries are reasonably short so we saw some > serious improvements by killing the slow-start. If you can modify your > IP stack then it still pays to use jumbo packets because you reduce > the overhead on the wire. OS tip of the day: If you run FreeBSD, check out sysctl net.inet.tcp.slowstart_flightsize. This lets you set a multiple for the initial window size and can be very beneficial for servers which do lots of quick start/stop transfers over TCP. I suspect this was much more useful to you then jumbo frames in that application. And on an unrelated note I have a mirror of the Because ICANN swf at http://www.e-gerbil.net/ras/video/icannstage.swf, and I'd like to thank the creators for not including any AYB references... We now return you to your regularly scheduled "can someone from the globix noc in botswana call me about my t1"... :P -- Richard A Steenbergen <ras(a)e-gerbil.net> http://www.e-gerbil.net/ras PGP Key ID: 0x138EA177 (67 29 D7 BC E8 18 3E DA B2 46 B3 D8 14 36 FE B6)

1 0

RE: liscensed wireless net service delivery
by Thomas Gainer 26 Apr '01

26 Apr '01

I have about 60 such devices deployed in the field from microwave radio to lasers, including 28 Triton IFU devices Jim is talking about. More than happy to answer questions regarding deployment and reliability. Thomas -----Original Message----- From: J Harris [mailto:jim@angrybuddha.com] Sent: Thursday, April 26, 2001 3:07 PM To: nanog(a)merit.edu Cc: airwavemaster(a)hotmail.com Subject: liscensed wireless net service delivery I am using both 100MB and OC-3 systems from Triton Network Systems to connect buildings, with excellent results and reliability. The inhibitor for most companies is acquiring usage rights from spectrum license holders. TNS is beginning to offer a package deal including both the wireless units and license rights for one monthly lease rate, or purchase. Either way, you can't beat the convenience, patching a 100FX into each site. I believe the availability covers much of the U.S. Let me know if you want more info off-line. -Jim -----Original Message----- From: Jayson Mays [mailto:airwavemaster@hotmail.com] Sent: Thursday, April 26, 2001 2:15 PM To: Nanog(a)merit.edu Subject: liscensed wireless net service delivery Is anyone using licensed wireless (100MB or OC-3) for corporate service delivery? Who's system and how satisfied are you? _________________________________________________________________ Get your FREE download of MSN Explorer at http://explorer.msn.com

1 0

RE: Why? Because ICANN
by Cerqua, Toby 26 Apr '01

26 Apr '01

For those of you who can't get through to it, I've provided a temporary alternate locale: http://www.blarf.com/media/icann/icannstage.asp -----Original Message----- From: Mike Batchelor [mailto:mikebat@tmcs.net] Sent: Thursday, April 26, 2001 12:25 PM To: nanog(a)merit.edu Subject: Why? Because ICANN It's funny, so laugh. http://www.paradigm.nu/icann/icannstage.html --- ALL YOUR BASE ARE BELONG TO US SOMEBODY SET UP US THE BOMB

1 0

Tier-2 information sharing
by Sean Donelan 26 Apr '01

26 Apr '01

Instead of trying to get Tier 1 providers to improve their inter-provider communications, why not ignore them? There are thousands of "customers" of all the major networks who notice problems with the major transit networks all the time. How about setting up a way for those customers of various network providers to pass along information about the trouble tickets they are opening with their providers. Rather than re-publishing the trouble tickets, summarize just the facts and inform the tier-2 providers which contributed tickets what the tier-1 providers told other people with the same problem. The customer's names would be sanitized, but the names of the Tier-1 providers (who don't participate) would be included in the summary of the facts. Facts aren't copyrighted. Since the government (i.e. FCC, FBI, etc) are also customers of the Tier-1's, they would also be eligible to participate and receive the summary. Is this a way to break the logjam?

1 0

Planning for community-built networks
by Pete Kruckenberg 26 Apr '01

26 Apr '01

I'm curious how much grass-roots last-mile efforts like community-funded networks are being factored into near-future network planning. In the two months I've been at my new job (network architect for Utah Education Network), I've been amazed at how much infrastructure activity is going on very quietly within communities. A lot of this is motivated by frustration with the lack of commercial infrastructure investment except in the most lucrative business areas, and a realization that there is not (and may never be) a business case to invest beyond those areas. Also relates to http://slashdot.org/articles/01/04/26/0446240.shtml In Utah (which not be representative) I'm aware of 12-15 such projects that are either under construction or will be completed in the next 1-3 years. These projects are running GigE to schools and govt, 10/100Mb Ethernet to homes and businesses. Community networks differ from traditional MANs in at least two ways: they are oriented towards wide-scale connectivity, usually of underserved areas (where MANs typically service dense business areas); and they appear to have more momentum in under-served geographies (ie non-NFL cities, and rural areas). In Utah, only 2 or 3 of the projects are urban, most are in communities on the urban fringe and in rural areas. Beyond driving the need for more bandwidth, especially in areas that don't have large pipes to the Net, these projects have the potential to make fundamental changes in the importance of local infrastructure. The biggest change may be to show that Geography Does Matter more than we think. The applications these networks will be used for are not just eBay and Amazon, it's video conferencing, community meetings, video on-demand, education, telemedicine, and many more important ones unimagined today. At least half the projects I'm aware of are looking for a local interconnect (ie Community Internet Exchange / Regional Exchange Point) to hand off traffic to other community, educational and commercial networks in the area. The national-level networks peer at such a course granularity that the majority of community network applications (video conferencing, video-on-demand, education, telemedicine) would fail for latency and performance reasons unless everyone connected to the same backbone. And there's your day's course in community networks. ;) Again, curious if anyone is seeing this on their radar and considering it in near- or long-term infrastructure planning. Pete.

1 0

liscensed wireless net service delivery
by J Harris 26 Apr '01

26 Apr '01

I am using both 100MB and OC-3 systems from Triton Network Systems to connect buildings, with excellent results and reliability. The inhibitor for most companies is acquiring usage rights from spectrum license holders. TNS is beginning to offer a package deal including both the wireless units and license rights for one monthly lease rate, or purchase. Either way, you can't beat the convenience, patching a 100FX into each site. I believe the availability covers much of the U.S. Let me know if you want more info off-line. -Jim -----Original Message----- From: Jayson Mays [mailto:airwavemaster@hotmail.com] Sent: Thursday, April 26, 2001 2:15 PM To: Nanog(a)merit.edu Subject: liscensed wireless net service delivery Is anyone using licensed wireless (100MB or OC-3) for corporate service delivery? Who's system and how satisfied are you? _________________________________________________________________ Get your FREE download of MSN Explorer at http://explorer.msn.com

1 0

RE: jumbo frames
by John Fraizer 26 Apr '01

26 Apr '01

We only have jumbo frames enabled on router<->router links. The GigE ports facing the aggregation switches runs standard 1500 MTU. --- John Fraizer EnterZone, Inc On Wed, 25 Apr 2001, Paul Lantinga wrote: > well, actually, I haven't had any for, oh, 5 years. Anyhow, what I was > getting at was what kind of price you pay performancewise for mtuX <--> > mtuY? Isn't there some penalty you pay in the conversion of jumbo to > normal? > > Years ago I ran a network that was all fddi with 4k mtu in the core and to > all the servers Everthing else ran ethernet and thus didn't get the > advantage of the big packets. > > -Paul. >

2 1

Re: Why? Because ICANN
by John Fraizer 26 Apr '01

26 Apr '01

I think it's experiencing the /. effect. [nanog@Overkill nanog]$ traceroute www.paradigm.nu 1 CORE-0-GE-3-1000M.CMH.ENTERZONE.NET (66.35.65.1) 2.847 ms 0.818 ms 0.887 ms 2 core1-FE1.Columbus.fnsi.net (209.115.93.129) [6259] 2.110 ms 1.774 ms 1.902 ms 3 border1-atm1-1-0-2.Chicago.fnsi.net (216.29.101.46) [6259] 40.817 ms 24.073 ms 12.248 ms 4 aads.cw.net (206.220.243.12) 21.258 ms 14.528 ms 13.564 ms 5 corerouter1.WillowSprings.cw.net (204.70.9.135) [3561] 14.598 ms 14.494 ms 16.626 ms 6 acr1-loopback.Chicagochd.cw.net (208.172.2.61) [3561] 14.461 ms 13.466 ms 14.597 ms 7 acr2-loopback.NewYorknyr.cw.net (206.24.194.62) [3561] 30.372 ms 30.098 ms 30.307 ms 8 savvis.NewYorknyr.cw.net (206.24.193.146) [3561] 26.389 ms 28.131 ms 31.131 ms 9 connectcom-1.usphla.savvis.net (64.241.148.62) [6347] 389.497 ms * 460.616 ms 10 sdweaver.com (64.242.242.56) [6347] 421.626 ms 390.159 ms * Note below that the DNS servers listed as authoritative weren't even serving up the zone previously. > >Funny, it doesn't work. > > > > > >Domain: paradigm.nu > >ID: 35184 > >Status: Active > >Created: 04-Mar-1999 > >Modified: 14-Jan-2001 > >Expires: 04-Mar-2004 > > > >Name servers: > > DNS.CCWEBHOST.COM 209.207.168.30 > > DNS2.CCWEBHOST.COM 208.155.67.49 > > > > > > > > >www.paradigm.nu > >Server: dns2.ccwebhost.com > >Address: 64.242.243.2 > > > >*** dns2.ccwebhost.com can't find www.paradigm.nu: Non-existent host/domain > > > server 209.207.168.30 > >Default Server: [209.207.168.30] > >Address: 209.207.168.30 > > > > > www.paradigm.nu > >Server: [209.207.168.30] > >Address: 209.207.168.30 > > > >*** [209.207.168.30] can't find www.paradigm.nu: No response from server > > > > I did finally get through and it is definately worth the wait. --- John Fraizer EnterZone, Inc

2 1

liscensed wireless net service delivery
by Jayson Mays 26 Apr '01

26 Apr '01

Is anyone using licensed wireless (100MB or OC-3) for corporate service delivery? Whos system and how satisfied are you? _________________________________________________________________ Get your FREE download of MSN Explorer at http://explorer.msn.com

1 0