I just thought of a small leak of info
If I allow different length messages, then an attacker could deduce the type of message you are sending by tracking the size of the packets. Not sure how big of a problem this is. The problem with fixed sized messages is that it will waste a fair amount of bandwidth.
Ah, maybe this is one of the differences in privacy levels!
Yes, I like that. At lower levels of privacy, such small info leakage is accepted,but at higher levels of privacy, all messages will be of fixed size
James
You could always add random length padding to your messages. However, this may be more trouble than it's worth for very small differences in size over fixed length.
good idea!
this could help, but I cant easily calculate the leakage as it becomes quite complex
Do you have documentation somewhere on the types of messages and their format? I'd like to take a look.
Most all messages are like this:
[{"requestType":"processutx","NXT":"8989816935121514892","utx":"0201b93a3d01d00225c5fed2690701cf06f267e7c227b1a3c0dfa9c6fc3cdb593b3af6f16d65302
f8c71b555df3ec27c000000000000000000e1f50500000000000000000000000000000000000000
0000000000000000000000000000000000000000000000000000000000000000000000000000000
0000000000000000000000000000000000000000000000000000000000000000000000000000000
0000444a53bf9828e869e80300000000000031007b22617373657442223a2231363339323939383
439333238343339353338222c2271747942223a2231303030303030227d","sig":"2dec3f78c73a4181e8ac4417eff4036176313935f79667ea8c6bc88a4bfd438f","full":"c04e8b6d34244aa5aef98f095197e9b8d1c8a8a3c3d4e6774c3fc955fcaac6e9","time":1406084344},{"token":"d79fth95u83cu087m4js5pr7oqktvg537dctmf7sclmv3thq7ktbibpgp0f3ecg1uk4fnk98cml5uvf
k1pqqe0em15c699vt1rs4admc3tugtv1jj59btt4rcj0e9pipkhqrs7m61mehd2krk9i819nvjlmhe0
tv"}]]
I keep adding new API calls daily. the first part of JSON is the actual API call, the second is a token that uses the first JSON object as the input. The processutx is one of the longer requests. There are currently over a dozen, probably another dozen more will be added this month. I just write them as I need them since they are easy enough to add.
James
P.S. I know, I should compress all the zeroes and I plan to use my jsoncodec once the API calls stabilize, just need a words list and should get nice 2 to 3x compression ratio