About the idea I described today

[Michael Welzl < >]

Hi,

The idea was that TCP could be forced to transmit data even unreliably, by simply acknowledging packets even when they were lost - by sending the receiver a “dummy” packet. Using byte stuffing, an application could detect that this is actually a hole in the data stream.

I thought I knew a paper that did this, but the paper I had in mind doesn’t *exactly* do this:

http://dkl.cs.arizona.edu/publications/papers/mmcn05.pdf

and a more recent, more popular-to-cite one is much more complex:

http://ieeexplore.ieee.org/document/7497221/

https://csperkins.org/publications/2016/05/mcquistin2016tcp-hollywood.pdf

… I wonder, we could either say that the idea is “in the direction” of these papers, or do something else / something similar… not sure.

Cheers,

Michael