On Thu, May 15, 2014 at 12:34 AM, Joran Dirk Greef <
jo...@ronomon.com> wrote:
> I have heard that TCP checksums have a practical probability of not
> detecting corruption in the data packet.
Whether 1 in 16 million is "practical" is pretty debatable.
http://dl.acm.org/citation.cfm?id=347561&dl=GUIDE&coll=GUIDE
> I am trying to find out more about the checksums that TLS uses? I assume
> these provide significantly more protection than TCP checksums? Is an
> application-level checksum still required for better end-to-end protection
> from corruption?
TLS doesn't use "checksums", it uses cryptographic message digests,
which is what you would use in your app layer protocol/data format if
you wanted equivalent levels of protection.
Sam