Hi,

I fully support the statements "TLS_AES_128_CCM_8_SHA256 is not suitable for 
general use" and "MUST NOT be used without additional safeguards". CCM_8 has no 
place in general non-constrained DTLS usage.

I do however not understand the logic behind the following classifications that 
have been done in recent TLS WG documents:

CCM_8                      MUST NOT be used without additional safeguards
Group key authentication   NOT RECOMMENDED
psk_ke                     RECOMMENDED

Of these three I would say that CCM_8 is by far the least worrisome security 
problem. Even in systems with 32 bit tags, the tag length is likely very far 
down on the list of most severe practical security problems. Non-PFS key 
exchange and symmetrical group keys on the other hand are very real practical 
security problems that are exploited daily.

Cheers,
John


_______________________________________________
TLS mailing list
TLS@ietf.org
https://www.ietf.org/mailman/listinfo/tls

Reply via email to