Posts

Last week I noticed that F root was showing poor connectivity with Indian RIPE Atlas probes for F-root. The graph looked really terrible.

 

I traced to it from one of RIPE Atlas probes and saw this trace:

Probe #6107
  1 2401:7500:fff0:1::1                      0.838 ms     0.747 ms     0.632 ms
  2 2400:5200:1c00:d::1                      1.755 ms     1.745 ms     1.726 ms
  3 2403:0:100::2be                          2.089 ms     2.054 ms     2.049 ms
  4 2404:a800:2a00::13d                     45.589 ms    26.274 ms     33.64 ms
  5 2404:a800::178                          26.376 ms    25.406 ms    25.276 ms
  6 2001:de8:1:2::3                         25.363 ms    25.232 ms    25.223 ms
  7 *                                           *            *            *
  8 *                                           *            *            *
  9 *                                           *            *            *
 10 *                                           *            *            *
 11 *                                           *            *            *

Here the last hop before timeout i.e hop 6 is of NIXI Chennai peering subnet 2001:de8:1:2::/64. As soon as I saw it, it reminded me older issue which happened and broke IPv4 connectivity to root DNS servers. I blogged about it here, here and here. So the problem remains that NIXI is broken cost wise due to charge on in - out policy. This leads to people accepting routes at all NIXI’s but they do not announce their routes. Thus return path is broken and essentially traffic is being blackholed. Earlier this issue was fixed by adding IP transit support to these root DNS servers so that a default route stays in case of all other failures. It seems like same is missing in IPv4 world and routes are not being announced. During this time, I saw two BGP sessions at NIXI Chennai for F root:

A few weeks back I got in touch with Marc from Meghalaya. He offered to host RIPE Atlas probe at Shillong and that’s an excellent location which isn’t there on RIPE Atlas coverage network yet. It took around 5 days for the probe to reach Shillong from Haryana. I think probably this probe is the one at the most beautiful place in India. :) Now that probe is connected, I thought to look into routing which is super exciting for far from places like Shillong. Marc has a BSNL FTTH connection & mentioned about not-so-good latency. Let’s trace to 1st IP of the corresponding /24 pool on which probe is hosted:

Seems like NPIX traffic has started hitting 1Gbps levels and that is just so amazing. Came across this post by Niranjan.

NPIX or Nepal Internet Exchange is located in Kathmandu and operates out of its own datacenter. There’s a huge amount of (overhead) dark fibre availability in Kathmandu and as of the writing of this post I see 30 members at NPIX. 1Gbps might seem low from Western IX’es standards but that’s a quite good amount of traffic for an IX in South Asian region. If you see the member list there are just smaller ISPs and not many content players over there. Comparing numbers by NIXI traffic,

I recently came across an excellent draft at IETF by Job Snijders & friends. This is to address scenarios where a network might miss communication about a maintenance activity when BGP shutdown happens. Once implemented, this can potentially offer to send peer a message with up to 128 bytes with info about shutdown like “Ticket XXX: We are upgrading the router, will be back live in 1hr” etc.

It depends by appending such data to the sys notification which is part of BGP protocol. This is one which sends a message just before the shutdown of the session. So it similar to the way you see session tearing down due to prefix limits etc. This has already been implemented in some of the open source routing implementations like OpenBGPd, GoBGP, PMacct, Exabgp etc.   Here is the latest draft of this change.

Today I was having a chat with my friend Hari Haran. He mentioned that Facebook has started its PoP in Mumbai. This seems true and Facebook has mentioned GPX Mumbai as their private peering PoP in their peeringdb record.

I triggered a quick test trace to “www.facebook.com” on IPv4 from all Indian RIPE Atlas probes and resolved “www.facebook.com” on the probe itself. The lowest latency is from Airtel Karnataka and that’s still hitting Facebook in Singapore. I do not see any of networks with probe coverage hitting Facebook node locally.