Hi Linda,
Thanks for your good comments here. As one of the co-authors, I try to give my 
replies as inline, more complements are anticipated from the other co-authors.

From: Linda Dunbar
Sent: Saturday, March 25, 2017 4:46 PM
To: John Strassner; Xialiang (Frank); Diego R. Lopez; [email protected]; Aldo 
Basile
Subject: comments to draft-xiabassnez-i2nsf-capability-01

John, Frank, Diego, and Aldo:

Thank you very much for the revision of the 
draft-xiabassnez-i2nsf-capability-01. I think the draft is structured really 
well, and describes a really good methodology for defining NSF capabilities. 
Very good work!

A few minor comments:
3.4.1 Network Security Capabilities
I think the section is mainly describing the Security Capabilities for traffic 
or flows traversing the network. I would think that "Network Security" is 
broader, which covers how to secure access to network elements as well, or 
encryption of links, management of secure keys, etc.
[Frank]: the line we draw between Network Security Capability and Content 
Security Capability is the network layer: the former is used for layer 1~4, the 
latter is used for layer 5~7, although there are some minor co-existence for 
them in layer 4. But basically, there two types of security capability can be 
divided by this mean clearly. So, my question is if we need a formal definition 
for them in the terminology draft?

3.4.2 Content Security Capabilities
Your draft stated that "Content Security" is at application layer. Can you give 
some examples? Is it about some specified content (such as URL, video, or 
something else) can't be accessed by some users?
How is "content" represented? Is it by an "Address"? Specific URL? Or special 
ID?

Should also reference the Section 4.3 which has more description.
[Frank]: see my above clarification. In layer 4~7, the content is represented 
no more by "Address", but by application layer metadata, such as: URL, file 
name, regular expressions of content, etc.
Figure 3 (Page 19):
are all those types of Rule (AuthenticationECAPolicyRule, 
AccoutingECAPolicyRule, ..) matched with the categories of "capabilities" 
described in Section 3.4?
Or all the "capabilities" listed under Section 3.4 are under the 
"SecurityECAPolicyRule"?

Figure 5 (Page 24):

The "event" in Figure 4 ( Page 22) are further classified as "user security 
event", "device security event", "system security event", and "Time security 
event".
But the "Condition" in Figure 5 are classified differently. What are the 
correlations between them? Is "UserSecurityCondition" mapped to 
"UserSecurityEvent"? how about the rest?
[Frank]: they are different and orthogonal. Event is "significant occurrences 
the NSF is able to react to", which is something happened to trigger the 
security policy's execution. But Condition is used during the process of 
security policy's execution to determine which actions will be applied.

What is the difference between "Packet Security Condition" and "Packet Payload 
Security condition"?
[Frank]: the former is for the packet header matching, the latter is for the 
packet payload matching.
Figure 6 (page 25):
Can "Apply Profile Action" apply to both "Ingress Action" and "Egress Action"?
[Frank]: both for each direction and both directions.

Thank you very much for putting together a good document to describe such 
complex subject very clearly.

Linda

_______________________________________________
I2nsf mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/i2nsf

Reply via email to