- Test - lists.nanog.org

Re: Updated Ookla Speedtest Server Requirements
by Rich Brown 11 Nov '15

11 Nov '15

> On Nov 10, 2015, at 7:00 AM, Hank Nussbacher wrote: > > The value of Ookla dropped significantly so we just let our license lapse > and did what everyone else was doing and pointed our speedtest to: > http://uk2.testmy.net/SmarTest/combinedAuto > and manage with this free service just fine. You might consider pointing people to the DSLReports Speed Test (www.dslreports.com/speedtest) - It is also free (no cost), and HTML-based. - It's graphically far more attractive than testmy.net - It's easier for people to use. Click one button: it automatically selects the nearest test server and runs all the tests. - It retains test history, so your customers can compare today's tests with earlier runs. - It measures latency *under load*, which really affects a lot of people's perception of "network speed". - I find DSLReports to be more accurate than testmy.net. In several runs today, testmy.net results average 90% of the up/download speeds seen by DSLReports, Ookla, and netperf measurements on my 7mbps/768kbps DSL connection. -Rich

4 4

Global Crossing / Level 3 Contact
by CWO Network Operations 10 Nov '15

10 Nov '15

Greetings, Some of my BGP announcements are suppressed at Global Crossing/Level 3. I have no idea why they would be suppressed, since my announcements (routes) are fine everywhere else and I’m not aware of any issues. Since I’m not peering with Level 3, it’s hard to get to anyone useful at Level 3. Feel free to contact me off list Thanks JB

1 0

Updated Ookla Speedtest Server Requirements
by Lorell Hathcock 10 Nov '15

10 Nov '15

Esteemed Legions of NANOG: Does anyone have better and more modern recommendations for the hardware of an Ookla speedtest server? Here is the link to their recommendations. http://www.ookla.com/support/a26461638/ I want a server that is capable of handlilng a speedtest up to 10Gbps. I plan to do this with an SFP+ port when my network comes along. (As soon as MikroTik comes out with a decent 10G CCR router that is compatible with more SFPs.) In the mean time I will just test 1 Gbps speeds off a copper GE port, but want the SFP+ capability so I don't have to repurchase hardware in the next year. Thanks! Sincerely, Lorell Hathcock Chief Technology Officer SolStar Network, LLC Communications FIBER - VOIP - SECURITY - TV FTTH - Commercial - Residential Burglar - Access Control 956-478-5955 (cell) - 956-316-4090 (main) <mailto:lorell@SolStarNetwork.com> lorell(a)SolStarNetwork.com <http://www.solstarnetwork.com/> www.SolStarNetwork.com TX License #B19998

11 15

RDAP adoption?
by Dave Crocker 10 Nov '15

10 Nov '15

Not sure whether this is a permitted query for the list... I'm trying to discover the field maturity of RDAP. How much deployment of servers supporting it is there, but also how much querying from clients is happening? Any sense of the adoption/use trends? Thanks (and maybe apologies.) d/ -- Dave Crocker Brandenburg InternetWorking bbiw.net

2 1

century link ipv6
by samaul carman 09 Nov '15

09 Nov '15

Hello I am writing to y'all in regards to based on my understanding that does not happen on a ipv6 dsl network however I may be wrong about that I am still learning about ipv6 in my college courses however I have been noticing that these only occur on cable network based on my google searches however like I said I may be wrong I have century link service and I am using an asus rt68p ac1900 router the modem is in transpartent bridging and I use ppoe for my connection protocol please excuse me if I made any errors on the post to this list you may contact me off list if you like Nov 6 kernel: net_ratelimit: 979 callbacks suppressed Nov 6 19:50:55 kernel: net_ratelimit: 307 callbacks suppressed Nov 6 19:51:01 kernel: net_ratelimit: 320 callbacks suppressed Nov 6 19:56:29 kernel: net_ratelimit: 93 callbacks suppressed Nov 6 19:56:34 kernel: net_ratelimit: 418 callbacks suppressed Nov 6 19:56:39 kernel: net_ratelimit: 169 callbacks suppressed Nov 6 19:56:44 kernel: net_ratelimit: 391 callbacks suppressed Nov 6 19:56:49 kernel: net_ratelimit: 389 callbacks suppressed Nov 6 19:56:58 kernel: net_ratelimit: 97 callbacks suppressed Nov 6 19:57:03 kernel: net_ratelimit: 439 callbacks suppressed Nov 6 19:57:17 kernel: net_ratelimit: 231 callbacks suppressed Nov 6 19:57:48 kernel: net_ratelimit: 261 callbacks suppressed Nov 6 20:02:19 kernel: net_ratelimit: 157 callbacks suppressed Nov 6 20:02:24 kernel: net_ratelimit: 75 callbacks suppressed Nov 6 20:02:29 kernel: net_ratelimit: 412 callbacks suppressed Nov 6 20:02:34 kernel: net_ratelimit: 439 callbacks suppressed Nov 6 20:02:39 kernel: net_ratelimit: 360 callbacks suppressed Nov 6 20:02:51 kernel: net_ratelimit: 219 callbacks suppressed Nov 6 20:02:56 kernel: net_ratelimit: 345 callbacks suppressed Nov 6 20:03:06 kernel: net_ratelimit: 89 callbacks suppressed Nov 6 20:03:11 kernel: net_ratelimit: 420 callbacks suppressed Nov 6 20:04:33 kernel: net_ratelimit: 180 callbacks suppressed Nov 6 20:04:46 kernel: net_ratelimit: 166 callbacks suppressed Nov 6 20:04:57 kernel: net_ratelimit: 140 callbacks suppressed Nov 6 20:05:02 kernel: net_ratelimit: 142 callbacks suppressed Nov 6 20:05:19 kernel: net_ratelimit: 159 callbacks suppressed Nov 6 20:05:49 kernel: net_ratelimit: 73 callbacks suppressed Nov 6 20:05:58 kernel: net_ratelimit: 76 callbacks suppressed Nov 6 20:06:13 kernel: net_ratelimit: 278 callbacks suppressed Nov 6 20:06:18 kernel: net_ratelimit: 50 callbacks suppressed Nov 6 20:06:32 kernel: net_ratelimit: 173 callbacks suppressed Nov 6 20:07:01 kernel: net_ratelimit: 173 callbacks suppressed Nov 6 20:08:14 kernel: net_ratelimit: 353 callbacks suppressed Nov 6 20:08:19 kernel: net_ratelimit: 149 callbacks suppressed Nov 6 20:08:29 kernel: net_ratelimit: 53 callbacks suppressed Nov 6 20:08:35 kernel: net_ratelimit: 228 callbacks suppressed Nov 6 20:09:32 kernel: net_ratelimit: 398 callbacks suppressed Nov 6 20:10:20 kernel: net_ratelimit: 258 callbacks suppressed Nov 6 20:10:31 kernel: net_ratelimit: 25 callbacks suppressed Nov 6 20:16:45 kernel: net_ratelimit: 301 callbacks suppressed Nov 6 20:16:50 kernel: net_ratelimit: 423 callbacks suppressed Nov 6 20:16:57 kernel: net_ratelimit: 64 callbacks suppressed Nov 6 20:17:02 kernel: net_ratelimit: 440 callbacks suppressed Nov 6 20:17:09 kernel: net_ratelimit: 298 callbacks suppressed Nov 6 20:17:14 kernel: net_ratelimit: 439 callbacks suppressed Nov 6 20:17:19 kernel: net_ratelimit: 112 callbacks suppressed Nov 6 20:17:24 kernel: net_ratelimit: 146 callbacks suppressed Nov 6 20:17:29 kernel: net_ratelimit: 427 callbacks suppressed Nov 6 20:17:34 kernel: net_ratelimit: 405 callbacks suppressed Nov 6 20:17:46 kernel: net_ratelimit: 93 callbacks suppressed

3 2

Re: Updated Ookla Speedtest Server Requirements
by alvin nanog 09 Nov '15

09 Nov '15

On 11/09/15 at 05:35pm, Josh Luthman wrote: > You can't get SFP+ PCI cards anyways. I don't think you can very easily > get boards with PCI on them, either. lots of vendors w/ PCIe 10gigE cards w/ copper or SFP+ interface .. very few ( almost none ) for explicitly PCI w/ 10gigE and more importantly i usually get intel chipset based SFP+ PCI-e x8 based dual-10gigE cards similarly, for dual 10gigE copper for testing on gigE infrastructcure not many "tyan/supermicro" motherboards with dual-gigE SFP+ connectors ( our primary requirement is dual-nic 10gigE ports squeezed into 1U chassis ) yes, it could be slow ... and most likely not even run at "bit speeds" quoted by marketing, but not everybody goes around measuring bandwidth speeds similarly, not much marketing anymore for 3.xGhz or 4.xGhz or 6.xGhz Ghz "cpu speeds" and lately, past decade, how many cores can you squeeze into 1sq inch which is trivially easy to count and measure cores and it's performance :-) PCIe-3.x spec'd at 8Gbit/s # x8 would easily give you 40Gbit/s PCIe-2.x spec'd at 4Gbit/s PCIe-1.x spec'd at 2Gbit/s PCIe x1 == 2Gbit/s ( per lane ) PCIe x4 == "combined" 8Gbit/s PCIe x8 == "combined" 16Gbit/s ( common PCIe cards ) PCIe x16 == "combined" 32Gbit/s PCIe x32 == "combined" 64Gbit/s > I expect he was just typing it out and left the "E" =) just being brain-dead lazy w/ assumptions and PCI vs PCIe does make a big difference with network thruput magic pixie dust alvin # Custom 1U chassis w/ zero drives and up to 8 drives ----- End forwarded message -----

1 0

ISPs for US BANK TOWER in LA
by Tri Tran 09 Nov '15

09 Nov '15

Hello everyone, Need to quickly turn up circuit. I've only been able to confirm these are in the bldg: TWC, ATT, Level3, and Cogent. CaInternet is the only who can turn up quickly but now sure how good their ClearFiber is. If anyone know of others, please share or PM me. Thank you --Tri

1 0

[TECH] Pica8 & Cumulus Networks
by Yoann THOMAS 09 Nov '15

09 Nov '15

Hello everyone, Under a Cloud project I ask myself to use equipment based on the Pica8 or Cumulus Networks. All in order to mount a Spine & Leaf architecture - Spine 40Gbps - Leaf in 10Gbps Someone of you there a feedback on this equipment. Regards, Yoann THOMAS CTO - Castle-IT

5 4

IP Addresses
by A MEKKAOUI 07 Nov '15

07 Nov '15

Hi Anyone can help on how to get IP addresses, purchase or lease or any broker who can help. Your help will be appreciated. Thank you KARIM

1 0

Long-haul 100Mbps EPL circuit throughput issue
by Eric Dugas 06 Nov '15

06 Nov '15

Hello NANOG, We've been dealing with an interesting throughput issue with one of our carrier. Specs and topology: 100Mbps EPL, fiber from a national carrier. We do MPLS to the CPE providing a VRF circuit to our customer back to our data center through our MPLS network. Circuit has 75 ms of latency since it's around 5000km. Linux test machine in customer's VRF <-> SRX100 <-> Carrier CPE (Cisco 2960G) <-> Carrier's MPLS network <-> NNI - MX80 <-> Our MPLS network <-> Terminating edge - MX80 <-> Distribution switch - EX3300 <-> Linux test machine in customer's VRF We can full the link in UDP traffic with iperf but with TCP, we can reach 80-90% and then the traffic drops to 50% and slowly increase up to 90%. Any one have dealt with this kind of problem in the past? We've tested by forcing ports to 100-FD at both ends, policing the circuit on our side, called the carrier and escalated to L2/L3 support. They tried to also police the circuit but as far as I know, they didn't modify anything else. I've told our support to make them look for underrun errors on their Cisco switch and they can see some. They're pretty much in the same boat as us and they're not sure where to look at. Thanks Eric

8 9