4 ms·
> maybe using static compression could help. Maybe but that's not a trivial question to answer: the linked article is 236K, gzipping yields 88K (84 at -9, brot
by masklinn 5y ago
> maybe using static compression could help.
Maybe but that's not a trivial question to answer: the linked article is 236K, gzipping yields 88K (84 at -9, brotly results in 68), so the questions become:
* how much energy does the extra 148K (/ 152 / 168) cost, aka once you've factored in the initial handshake what is the efficiency per byte of the transfer
* how much does decompression cost on the client side
* what is the mix of clients accepting each compression algorithm (and what is the overlap)