Re: [core] Éric Vyncke's No Objection on draft-ietf-core-new-block-11: (with COMMENT)

["Eric Vyncke (evyncke)" <[e-mail]>]

Jon and Med,

Thank you, I feel way more comfortable with that amount of testing ;-)

Regards

-éric

-----Original Message-----
From: "[e-mail]" <[e-mail]>
Date: Tuesday, 4 May 2021 at 10:05
To: Eric Vyncke <[e-mail]>, 'The IESG' <[e-mail]>, "[e-mail]" <[e-mail]>
Cc: "[e-mail]" <[e-mail]>, "[e-mail]" <[e-mail]>, "[e-mail]" <[e-mail]>, "[e-mail]" <[e-mail]>, "[e-mail]" <[e-mail]>, "[e-mail]" <[e-mail]>
Subject: RE: Éric Vyncke's No Objection on draft-ietf-core-new-block-11: (with COMMENT)

    Hi Eric,

    > >
    > > EV> indeed, thank you. Now, whether this code has been tested in
    > > several setups / test cases (bandwidth, packet loss, reboots) would
    > > be interesting.
    > >
    > 
    > [Med] It was tested in the context of DOTS telemetry. The list above shows
    > how the testing was helpful to refine/assess the design.
    > 

    Libcoap provides the testing capability of one or more specific packets dropped, or randomly dropped. Testing was done with one or more packets missing from the start, around the MAX_PAYLOADS pause point, at the end as well as in the middle of a MAX_PAYLOAD sequence for both CON and NON.  Lessons learnt about the recoveries etc. updated the draft.

    Logic tested with complete packet loss in one direction (for NON, CON fails here) confirming data still gets through.

    Multiple concurrent sessions were run in parallel - including with enforced packet loss.

    Applications running on the peers were restarted frequently (other peer remaining up) to checkout out recovery etc.

    The block size (SZX) was deliberately set to a small size, thus considerably increasing the number of blocks needed to transfer large payloads to exercise that logic.  This brought about the addition of the 'Continue' logic.

    Regards

    Jon