W3C home > Mailing lists > Public > ietf-http-wg@w3.org > April to June 2014

Clarifications on flow control and padding

From: Simone Bordet <simone.bordet@gmail.com>
Date: Thu, 12 Jun 2014 19:29:11 +0200
Message-ID: <CAFWmRJ0AabhWm-gx+shE4p9NRLOcGkSj-1kPvA0ujXTsqzrk4g@mail.gmail.com>
To: HTTP Working Group <ietf-http-wg@w3.org>
Hi,

draft 12, section 6.1 states that "padding is included in flow control".

A) Does this mean that, given 100 bytes of data and 300 bytes of padding:
1) the number that updates the flow control window is 400 (data + padding)
2) the number that updates the flow control window is 402 (pad_high +
pad_low + data + padding)

B) In the same example, let's assume that the stream has a remaining
flow control window of 250.
What should an implementation do ?
1) send a frame composed of: pad_low + 100 bytes of data + 149 bytes of padding
2) do not send a frame because the there is not enough window for the padding ?

C) In the same example, let's assume that the stream has a remaining
flow control window of 350.
What should an implementation do ?
1) send a frame composed of: pad_low + 100 bytes of data + 249 bytes of padding
2) send a frame composed of: pad_high + pad_low + 48 bytes of data +
300 bytes of padding

D) What if the remaining flow control window is exactly 300 (or 302 in
case A2 holds true) ?
1) send a frame composed of: pad_low + 100 bytes of data + 199 bytes of padding
2) do not send a frame because there is enough space for padding but
not for data ?

E) An application wants to write 20 MiB of data with a padding of 300 bytes.
Does this mean:
1) N data frame frames without padding, and a last frame with the
remainder of the data and 300 bytes of padding
2) N data frames each with 300 bytes of padding.

Are there security implications for case D2 (an intermediary could
stall a frame by carefully choosing window updates) and E2 (if the
padding is a constant, is its role useful) ?

Thanks !

-- 
Simone Bordet
http://bordet.blogspot.com
---
Finally, no matter how good the architecture and design are,
to deliver bug-free software with optimal performance and reliability,
the implementation technique must be flawless.   Victoria Livschitz
Received on Thursday, 12 June 2014 17:29:38 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:14:31 UTC