[Monotone-devel] Re: partial pull #3 - calling conventions

[Lapo Luchini]

Christian Ohler wrote:
> Ok.  Still, verifying signatures on 10MB worth of data is very likely
> faster than verifying them on 71MB worth of data.

Not really: that very 10MB contains all the data that is RSA-signed...
the files themselves are "only" hashed in order to get a smaller value
to be actually signed.

SHA-1 hashing 70MB worth of data is a 4 seconds work on my PC.
OK, it would take a bit more because those 70MB worth of data is not a
single file but many smaller files (the is a bit of overhead to "close"
the hash).

The problem is that the other 10MB are actually used to recreate
revision, that is, gigabytes of text data to be hashed and verified
against signatures. The "simple" hash of the other 70MB is a little thing...

(I'm working on assumptions and incomplete knowledge of the internals,
please someone correct me if I'm wrong)

> Or is all of the cryptographic verification that monotone does during
> pull really needed at that time due to security considerations (e.g. to
> ensure that the server doesn't perform a DoS attack on the client by
> sending an infinite chain of junk revisions)?

Is having a 1 hour pull and "fast" checkout worse than having a "fast"
pull and a 1 hour wait on the first checkout?
Sooner or later, the check is to be done, if we want to keep the current
level of "correctness assurance" of the data...

    Lapo