Re: Back to Merkle Hash Trees...

> 
> 
> >From: Olaf van der Spek <[email protected]>
> > Elliott Mitchell wrote:
> > > aids cryptanalysis. As lower levels (TCP) will take care of errors, why
> > > add redundancy at the application layer?
> > 
> > Maybe because the TCP checksum isn't (very) strong.
> 
> It is more than strong enough to fulfill its intended purpose. Mainly
> guarding against errors due to line noise and other accidental errors
> during packet handling.
> 
> The TCP checksum is sufficiently strong that if you get a packet where
> the checksum verifies and yet has data different in any way from what the
> other end sent, you can say beyond a reasonable doubt that it was caused
> by deliberate action.
> 
> Note protocols like FTP which send large binary files across the Internet
> without any additional checksums at all. Yet FTP works with very high
> reliability. Sites publish MD5 checksums in an attempt to thwart
> modification by attackers, not because there is even the slightest
> concern of FTP/TCP failing.

This is not universally agreed on. See for example the adding of checksums
to iSCSI accompanied by arguments for the error rate of the TCP checksum for
large transfers (rfc3385).

Justin


 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/BitTorrent/

<*> To unsubscribe from this group, send an email to:
    [email protected]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/