prifi

prifi copied to clipboard

For various performance settings

lbarman

I'm guessing since the PriFi-lib messages are marshallable anyway, we could have a message ``` type PriFi_Message struct { *sda.TreeNode interface{} } ``` that could hold, code and decode any...

lbarman

RelayState contains way too much information; we need to split between: 1) stuff that does not change at runtime 2) stuff regarding the communication state etc etc.. We need to...

lbarman

Trustees only reduce the number of keys. If clients check that the relay is behaving correctly, there is no added security. We could have a significant performance gain without the...

lbarman

lbarman

Rename with `ingress/egress`

lbarman

Not loosing data when resync'ing

1

To be done after issue https://github.com/lbarman/prifi_dev/issues/14. When we make sure that the PriFi protocols eventually reboots and is able to communicate no matter what, then we should make sure that...

lbarman

In the paper, A `slot` is a upstream cell, for one client A `round` contains N `slot`, one for each client An `epoch` is a succession of rounds between two...

lbarman