[tcpm] MSS statistics from MLab data

[Matt Mathis <[e-mail]>]

I tabulated MSS values from one global MLab day (roughly 3M tests).
 The attached .csv shows 587 unique values for <IPv? * MSS> tuples

I also plotted it:
https://docs.google.com/spreadsheets/d/e/2PACX-1vSGi8_TWyjVc50WUYlMEqmKO0mGkRBxqeOIkFCXOSzzVWcshP2oEkrnV1qfvbp0UnFnNZh83Dn6wUeM/pubhtml?gid=1172336962&single=true

Many values are "obvious" combinations of TCP options, tunnels and network
MTU, but most are not.  If somebody else wants to pick this apart, I would
be happy to provide some hints.

You can run the query yourself:

# 2023-07-027 MLab MSS statistics for ietf tcpm

SELECT
  raw.Download.ServerMeasurements[SAFE_OFFSET(0)].TCPInfo.SndMSS AS MSS,
  raw.ServerIP LIKE '%:%' AS Ipv6,
  Count (*) as tests,
FROM
  `measurement-lab.ndt.ndt7`
  WHERE date='2023-07-01'
    AND raw.Download.ServerMeasurements IS NOT NULL
GROUP BY MSS, IPv6
ORDER BY MSS DESC, IPv6

Follow the directions at
https://www.measurementlab.net/data/docs/bq/quickstart/
(tl;dr mlab will pay for the query, but you need to provide a credit card
and a google account subscribed to the MLab discuss list and enrolled in
BigQuery).

Thanks,
--MM--
Evil is defined by mortals who think they know "The Truth" and use force to
apply it to others.