Re: Last Call: <draft-farrell-perpass-attack-02.txt> (Pervasive Monitori

Folks,

Rather than discussing specific details of the current doc, I though it 
might
be helpful to offer some text. The text below is my take as what I think we
should be saying.

Steve
-------

Recent revelations [cites] about widespread monitoring of Internet 
communications by a number of national intelligence agencies has caused 
great concern in the IETF community.
The IETF security community has always considered passive wiretapping be 
a feasible attack against the confidentiality of Internet traffic while 
in transit (or while in storage en route to an end user). We also 
presume that active attacks are feasible, especially off-path active 
attacks, e.g., DNS cache poisoning [cite]. Even on-path, 
man-in-the-middle (MITM) attacks are viewed as possible, by some classes 
of adversaries.
The IETF has developed a set of security protocols to counter passive 
and active wiretapping attacks, e.g., IPsec, TLS, SSH, S/MIME, SRTP. All 
of these protocols offer confidentiality, via encryption, as a way to 
defend against passive wiretapping. They also employ cryptographic 
message authentication codes (MACs) or digital signatures to provide 
data integrity and authentication.
The cryptographic mechanisms employed by these protocols are effective 
against a very broad range of attacks and adversaries.
The IETF has not ignored the potential for passive and active wiretapping.

In some respects, widespread monitoring of Internet communication by 
nation states ought not be a surprise. They qualify as threats against 
Internet communication, i.e., motivated, capable adversaries. These 
adversaries have extensive technical capabilities and are motivated to 
collect information in support of (their) national security missions. 
Nonetheless, the scale of the monitoring that has been disclosed is 
considerably greater than what most IETF community members anticipated.
The IETF community perceives a qualitative difference between targeting 
and monitoring individual Internet communications vs. very widespread 
(pervasive) monitoring of such communication. As a result, the IETF 
community has reached consensus [cite] to address the pervasive 
monitoring of Internet communications by revisiting our security 
protocol standards with a goal of making it easier for users and service 
providers to offer better protection for Internet communications.
IETF security standards provide a number of mechanisms to protect the 
application layer content of Internet communications. However, most of 
these standards do not emphasize confidentiality for protocol metadata, 
other than what accrues from encapsulating such data at higher layers. 
Based on the disclosures cited above, the IETF community also has 
decided that more emphasis should be placed on protecting such metadata.
The IETF community recognizes that not all aspects of Internet 
communicationsecurity are within the purview of the IETF. Communication 
security may be offered at multiple layers within the protocol 
hierarchy; the IETF does not create standards for all of these layers. 
For example, the IEEE has created encryption standards for layer 2 (LANs 
and MANs). Security for Internet communications also is affected by 
design and implementation choices within hosts and servers, and most of 
these are outside of the scope of IETF standards.
The IETF community recognizes that not all forms of monitoring of 
Internet traffic constitute the sort of attack that have motivated this 
declaration. Many Internet service providers monitor traffic on their 
nets in support of traffic engineering and network management. Such 
monitoring does not require access to application layer content, but it 
does require access to some types of metadata. Many application service 
providers access application layer content as part of providing their 
services. This access is outside of the scope of this effort. Within an 
enterprise network, monitoring of traffic often is performed in support 
of security, e.g., intrusion detection systems, anti-virus and anti-spam 
filtering. This type of monitoring has a legitimate purpose and is 
within the purview of an enterprise, so it too is not considered an 
attack to be addressed by this IETF effort.