Current existing servers and their versions

[Alex MR]

Hello everybody,

I've been testing this protocol with the python client made by Marcus Dansaire for a few months now. Until this summer the Google server (roughtime.sandbox.google.com or rough-time-service.l.google.com) answered my requests but it doesn't anymore. I've been trying different versions of the protocol (nonce size, new or old packet...) to see if they've changed the version but I can't get it to respond. Could you please confirm if the google server gives you a response and what request format/version it needs?

I have identified 11 servers that I think should be offering the service but I have only managed to get response from 6 of them. I am leaving the list in case we can comment on whether we can get a response from the last 5 and what formats/versions are required.

Confirmed active status for:
-roughtime.cloudflare.com:2002;    nonce_size(64);  newver(false)
-sth2.roughtime.netnod.se:2002;    nonce_size(32);  newver(true)
-roughtime.se:2002;                          nonce_size(32);  newver(true)
-roughtime.dnov.se:2002;                 nonce_size(64);  newver(true)
-roughtime.int08h.com:2002;          nonce_size(64);  newver(false)
-time.0xt.ca:2002;                              nonce_size(64);  newver(false)

Unconfirmed active status for:
-roughtime.sandbox.google.com:2002
-roughtime.engineering.nyu.edu:5333
-roughtime.chainpoint.org:2002
-roughtime.qrdate.org:2002
-ticktock.mixmin.net:5333


I am currently trying to convince a major entity to launch a roughtime server to expand the range of servers available to the community. When I get it I will let you know.

Greetings and thanks for your help.

[Justin Cappos]

roughtime.engineering.nyu.edu:5333 is permanently down.  We weren't seeing much traffic and our IT staff was wanting to reclaim VMs.  As a result, it got wiped...

Justin

--
You received this message because you are subscribed to the Google Groups "proto-roughtime" group.
To unsubscribe from this group and stop receiving emails from it, send an email to proto-roughti...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/proto-roughtime/4df3257f-0397-4a8e-8e0e-4449c216ff69n%40chromium.org.

[Christer Weinigel]

Hi,

I'm the guy who has set up the servers at Netnod.  It's a bit of a mess which servers support which versions right now though.  There are some servers which support the pre-IETF draft version of roughtime.  There are some servers that support draft version 05/06 which has quite a lot of changes compared to the pre-IETF version.    Draft version 07 changed the hash algorithm and a few other minor things in incompatible ways.  Marcus Dansarie has updated his "roughtimed" and "pyroughtime" implementations to support draft version 07.  There are also some servers which seem to partially implement draft 05/06.

sth1.roughtime.netnod.se:2002 and sth2.roughtime.netnod.se:2002 are our production servers and run the older version of Dansarie's roughtimed implementation which support draft 05/06.  Our goal is to keep these servers in production "forever", or at least as long as people are interested in roughtime.  We're also planning to add roughtime servers att all our six sites here in Sweden where we also provide NTP services.

We also run a test server at lab.roughtime.netnod.se:2002 which I just updated to Marcus Dansarie's latest version which supports draft version 07.  I will try to update all our servers to roughtime draft version 07 before the IETF hackathon this weekend.

Note that we are also running a server called "falseticker.roughtime.netnod.se" on ports 2000..2009 which return invalid time.  These servers should not be used in production, but have been useful for testing a selection algorithm which tries to select "good" roughtime servers in the presence of bad ones.  More info on that project here: https://github.com/netnod/vadarklockan .

I'm trying to update the implementation in "vadarklockan" to draft version 07 and I'll try to make patches for the upstream of the roughtime implementations used there.

  /Christer

[Christer Weinigel]

Hi again,

I finally got access to the safe where we store our long term keys so that I could generate delegate certificates using the new hashing algorithm. All of Netnod's servers have now been upgraded to Marcus' latest version of roughtimed and use draft version 07.

If anyone wants to try them out here are our production servers:

    {
      "name": "sth1.roughtime.netnod.se",
      "publicKeyType": "ed25519",
      "publicKey": "9l1JN4HakGnG44yyqyNNCb0HN0XfsysBbnl/kbZoZDc=",
      "addresses": [
        {
          "protocol": "udp",
          "address": "sth1.roughtime.netnod.se:2002"
        }
      ]
    },
    {
      "name": "sth2.roughtime.netnod.se",
      "publicKeyType": "ed25519",
      "publicKey": "T/xxX4ERUBAOpt64Z8phWamKsASZxJ0VWuiPm3GS/8g=",
      "addresses": [
        {
          "protocol": "udp",
          "address": "sth2.roughtime.netnod.se:2002"
        }
      ]
    },

Our lab server is this one, but it is a test environment so it might break every now and then:

    {
      "name": "lab.roughtime.netnod.se",
      "publicKeyType": "ed25519",
      "publicKey": "tAPsQc3D36mJPOX1/LLNbV0M3fuTPjwW54mGvw+V1hE=",
      "addresses": [
        {
          "protocol": "udp",
          "address": "lab.roughtime.netnod.se:2002"
        }
      ]
    },

And a server which gives invalid time is here, do not use this one unless you are testing an algorithm to reject falsetickers:

    {
      "name": "falseticker.roughtime.netnod.se",
      "publicKeyType": "ed25519",
      "publicKey": "iFY9glIn8SHGtkFTdUECYdC37T4PNM2YSFzjbEbmfZI=",
      "addresses": [
        {
          "protocol": "udp",
          "address": "falseticker.roughtime.netnod.se:2002"
        }
      ]
    },

all ports from 2000 to 2009 return time with different errors.

  /Christer