From 266d391b858050217a208d9870a8a681e5b41606 Mon Sep 17 00:00:00 2001 From: zzz Date: Wed, 6 Sep 2017 15:20:57 +0000 Subject: [PATCH] meeting 263 --- i2p2www/meetings/logs/263.log | 89 +++++++++++++++++++++++++++++++++++ i2p2www/meetings/logs/263.rst | 13 +++++ 2 files changed, 102 insertions(+) create mode 100644 i2p2www/meetings/logs/263.log create mode 100644 i2p2www/meetings/logs/263.rst diff --git a/i2p2www/meetings/logs/263.log b/i2p2www/meetings/logs/263.log new file mode 100644 index 00000000..f9b580ef --- /dev/null +++ b/i2p2www/meetings/logs/263.log @@ -0,0 +1,89 @@ +20:00:00 0) Hi +20:00:00 1) 0.9.32 update (zzz) +20:00:00 2) 34C3 funding email reminder (zzz/echelon) +20:00:03 0) Hi +20:00:05 Hi +20:00:44 1) 0.9.32 update (zzz) +20:00:58 Hi +20:01:09 ok, str4d has done some UI updates, and I've started on the prop 141 implementation but haven't checked anything in yet +20:01:37 we're on track for an early october release +20:01:49 [Slack/str4d] Hi +20:02:03 I think str4d wants to prop his benchmark branch, he should do that soon? I've commented in his ticket +20:02:20 ay +20:02:36 [Slack/str4d] I've only pushed a minor UI tweak thus far; I have more sitting locally that addresses a bunch more issues, but I need to go through my git -> mtn process +20:03:09 [Slack/str4d] I'll look at the benchmark comments and finish / push that end of this week +20:03:57 ok I need to talk to you at some point about our release process. We had blocker tickets for .31 that weren't closed, probably want to insist that those are closed before a release +20:04:08 or else what does blocker even mean +20:04:23 [Slack/str4d] Correct +20:04:36 anything elase on 1) ? +20:06:01 2) 34C3 funding email reminder (zzz/echelon) +20:06:11 does this release require removal of hostnames? +20:06:15 in RI +20:06:25 gah lag +20:06:33 see the proposal text for migration discussion +20:06:45 kk +20:07:07 [Slack/str4d] -1 on it being in this release without discussion of zombie mitigations +20:07:08 ok re: 34C3, if you want funding or free ticket you MUST email echelon by Sept. 30 +20:07:43 in addition, echelon did have some server issues, so if you didn't get an ACK from him that he got your email, send it again +20:08:46 we have plenty of funds available for people but you must ask. We won't be funding people who ask after the end of the month +20:09:48 so again make sure that echelon has acknowledged receipt of your request +20:10:03 we will set the budget at next month's meeting +20:10:19 anything else on 2) ? +20:10:36 [Slack/str4d] Not from me. +20:11:26 anything else for the meeting? +20:11:54 i have something +20:12:02 psi go +20:12:03 but it's long and teedious +20:12:09 it's that aligned outbound tunnels idea +20:12:36 originally i sold it to you as a OBEP load reduction technique +20:12:45 that is a nice side effect +20:12:53 but that is not the original intent +20:13:10 the original intent was to reduce packet drop +20:13:59 ok, so what would you like to discuss about it? +20:14:08 my question is: would java i2p implement aligned outbound tunels ? +20:14:22 or is it too experimental for you? +20:14:53 i am not as familiar with java i2p's code as i am i2pd's +20:14:57 can't answer now because I forgot the details. If you write it up and post it somewhere I'll be happy to give you an answer +20:15:09 okay +20:15:15 i guess you can close meet +20:15:26 the idea is OBEP == IBGW +20:15:35 with an extra hop on the OB tunnel +20:15:38 nothing from me so far +20:15:43 such that OBEP == IBGW +20:16:14 to reduce packet drop and OBEP pressure +20:16:30 (at the expense of more tunnels) +20:16:51 ok, since you've already implemented it, any data on the benefits would be very helpful +20:17:10 anything else on aligned outbound tunnels? +20:17:31 my initial observations of it is the initial RTT is the same as after +20:17:44 rather, there is no initial RTT spike +20:17:57 possibly because of the release of pressure on OBEP +20:18:03 but that is just an assumption +20:18:15 i want to test this on a testnet, of which we have with docker. +20:18:25 [Slack/str4d] If there's something we can turn into a performance benchmark, LMK +20:18:25 to collect hard numbers etc +20:19:01 yeah same here, i am at a loss for a good perf benchmark +20:19:18 i have been using icmp ping over openvpn +20:19:23 [Slack/str4d] Actually this would be more like a metric, since it would depend also on network performance, and would likely differ depending on endpoint locations +20:19:27 probably not the best way +20:19:48 [Slack/str4d] But if we *can* make a repeatable benchmark, I'd want to add it to the suite I plan to start collecting +20:20:18 what i use right now is, time to connect via dtls and then the following latency measurement via ping +20:20:31 that isn't porable for java i2p i think +20:20:45 unless socks5 udp works +20:20:49 or i do some SAM stuff +20:21:23 anything else on aligned outbound tunnels? +20:21:31 aligned outbound tunnels is still experimental and idk if the increased tunnel count is worth it or not yet +20:21:49 so more research needed and it's getting scienced rn over at i2pd +20:21:56 i'll let you know +20:22:12 [Slack/str4d] Great, keep me appraised of the science in #i2p-science :slightly_smiling_face: +20:22:20 kk +20:22:21 great, thanks for the update psi +20:22:25 anything else on aligned outbound tunnels? +20:22:53 one final thing: it may be worth it to do something in addition to aligning tunnels, i.e. something like tor's rend spec +20:23:17 as for what that is idk and will think about it outloud in #i2p-science +20:23:20 (please join) +20:23:29 that's it +20:23:41 [Slack/str4d] All from me +20:23:49 anything else for the meeting? +20:24:28 i am good +20:25:15 thanks everybody, see you in 4 weeks, which will be .32 release time +20:26:10 * zzz ***bafffs*** the meeting done diff --git a/i2p2www/meetings/logs/263.rst b/i2p2www/meetings/logs/263.rst new file mode 100644 index 00000000..3c5933ff --- /dev/null +++ b/i2p2www/meetings/logs/263.rst @@ -0,0 +1,13 @@ +I2P dev meeting, September 5, 2017 @ 20:00 UTC +============================================== + +Quick recap +----------- + +* **Present:** + +echelon, +psi, +R4SAS, +str4d, +zzz