[Ecrit] Comments on: draft-ietf-ecrit-phonebcp-01

[Matt Lepinski <mlepinski at bbn.com>]

Brian and James:

First, two very high level questions.

A) [Is this document supposed to be SIP-Centric?] It is unclear to me 
whether the document is intended as a BCP for SIP User Agents and SIP 
proxies, or whether it is more generally a BCP for end points and 
signaling-path devices which may or may not use SIP. For instance, 
Section 6.4 is written as though SIP signaling were a special case, and 
yet paragraph four of Section 5 indicates that all emergency calls on 
the wire should contain a Route header (which is a SIP-specific 
signaling component). Personally, I don't care whether or not the 
document is SIP centric, or whether SIP is a special case, as long as 
the document is consistent.

B) [Is there consensus on how to mark emergency calls?] The current 
version of phonebcp indicates that (in the normal case when end-points 
perform the LoST mapping) emergency calls are marked by putting the URN 
in a Route header (with a "loose" parameter that I'm not familiar with 
... can someone provide me with a reference) and that the URN is placed 
in the Request-URI when the end-point is unable to perform the LoST 
mapping). Personally, I don't understand the current mechanism because 
I'm not sure what an "ordinary" (RFC3261-compliant) proxy does when it 
sees a service URN in the top Route header. However, given the long 
discussion that occured on call marking in January (See: 
http://www1.ietf.org/mail-archive/web/ecrit/current/msg02963.html), 
perhaps the more important question is "Has consensus been reached that 
(something like) the current mechanism is an appropriate way to mark 
messages?"

---------------------------------
More detailed comments:

Section 2: (paragraph 3 - number 2)
Is it clear what the "visited location's emergency number" means in this 
context?
Consider changing to "the local emergency number for its current 
location" or providing a forward reference to the discussion of 
dial-strings in Section 5.

Section 2: (paragraph 4)
I agree with Barbara that if this is intended to an overview of call 
setup for the special case of an Ethernet connected phone then the 
bullets should match up more clearly with the numbered items in the 
previous paragraph.

Section 2: (last paragraph)
It is unclear whether the antecedent of "it" in the last paragraph is 
[RFC4103] or [RFC4504].

Section 3:
Consider changing "should support" to "SHOULD support"

Section 3: (paragraph 2)
It seems that the point of this paragraph is to say that a certain class 
of devices that communicates over IP networks should support emergency 
calls. However, I find the phrase "using current (evolving) standards" 
to be unclear and in particular I'm not sure what the phrase modifies. 
(E.g. Do you mean "Devices that create media sessions using current 
(evolving) standards and exchange audio ...")

Section 4: (first paragraph)
Consider replacing "the norm" with "required" (or a similar word). I 
think the point is not that automatic location is common/normal but that 
it is necessary since most users are unable to provide accurate location.

Section 4.1: (first paragraph)
I think "cellular" (or "traditional wireless")  is more precise than 
"mobile".

Section 4.2: (first paragraph ... also, the last paragraph)
Consider changing the phrase "the desired result" to something like "an 
equivalent result", or perhaps something even more explicit.

Section 4.2: (first paragraph)
Consider addition a phrase to the last sentence indicating why it is 
recommended that the network support a standardized LCP. (This may not 
be obvious to the reader). Also, consider changing "recommended" to 
"RECOMMENDED".

Section 4.2: (paragraph 2)
Given that the paragraph covers what both devices and access networks 
MUST support, consider changing "For all other devices" to "In all other 
scenarios".

Section 4.3: (paragraph 3)
The term "geo-location" is not used in RFC 3825 to describe a 
lat/lon/alt - style location. (Instead it uses terms like a 
"coordinate-based geolocation"). Also, given that the "geolocation" 
header allows for civic locations, I think using geo-location here is 
potentially confusing.
Note: This comment also applies to the use of "geo reported" in the 
third paragraph of Section 6.3

Section 4.4:  (last paragraph)
Consider re-writing the second-to-the-last sentence as: "Certain 
commonly-used techniques for measuring location create a conflict 
between the time it takes to generate a precise location and the desire 
to route the call quickly." (Also, in the following sentence I think 
"precise" is a better word than "accurate")

Section 5: (paragraphs 5 and 6)
Paragraph 5 says that devices MUST mark calls using a service:sos URN. 
However, paragraph 6 says that mapping from dialstring to URN SHOULD be 
done by the endpoint. Either both paragraphs should use "MUST" or else 
they should both use "SHOULD".

Section 5: (paragraph 8)
I don't understand what it means to be "roaming" or "nomadic" in a 
system where there is no "visited network"

Section 5: (last paragraph)
The last sentence of this paragraph seems to be redundent. Consider 
deleting it.

Section 6.1:
I agree with Barbara that the use of IPSec should not be prohibited by 
this BCP.

Section 6.1: (number 6)
These instructions are for the User Agent,  P-Asserted-Identity headers 
and Identity headers should not be inserted by the UA

Section 6.1: (number 10)
This item is unclear to me. Is the author's intention that this item 
handles the case where a UA does not know its own location? If so, I 
guess that updating this item should be deferred until consensus is 
reached on location-hiding.

Section 6.1:
Consider Re-ordering the items as:
11, 10, 12, 14, 15, 13
Also, consider deleting 12, I believe it is redundent given 9, 10 and 11.

Section 6.2: (number 1)
Consider adding an example after "... URN appropriate for the emergency 
dialstring." That is, consider adding (e.g., ...)

Section 6.2: (number 3)
I'm afraid that I don't understand the meaning of this sentence.

Section 6.3: (paragraph 3)
The sentence that begins "This can be an enclosing ..." is unclear to 
me. Are you suggesting that when a PSAP coverage region is complex, that 
a LoST server SHOULD return a simple polygon that (a) contains the 
location of the device; and (b) is entirely contained within the PSAP 
coverage region?

Section 6.5: (number 3)
This seems to imply that proxies should expect to SUBSCRIBE to Presence. 
Do you mean that proxies should expect the PSAP to SUBSCRIBE to Presence?

Section 6.5: (number 4)
I'm not very familiar with Session Timers, can you provide a reference.

Section 6.6:
I'm afraid I don't understand the parenthetical remark after the "Call 
Forward" bullet. (Consider removing the remark or re-wording it if it is 
important).

Section 7:
Given the text in Section 4.4 it seems to me that the first paragraph of 
Section 7 is redundent and should be deleted.

Section 10.2:
Given the difficulty of implementing location signing in a useful 
manner, I think that that either paragraph 2 should either be removed or 
else it should reference some other document that explains location 
signing in more detail. (That is, one paragraph does not do location 
signing justice and it seems irresponsible to strongly recommend 
location signing without providing additionaly guidence to the implementor).

-------------------------------------------------------------------
Minor Nits: (That don't change the meaning of the text)

Section 2: (last paragraph)
Put a space between [RFC4103] and "media"

Section 3: (paragraph 1)
Add commas to second sentence "Future PSAPs will, however, support ..." 
and remove the extra period at the end of the sentence

Section 4.1: (paragraph 2)
Change "... where it is the access network that knows the location ..." 
to "... when only the access network knows the location ..."

Section 4.4: (paragraph 1)
Change "... process engaged from establishing a VPN ... " to "... 
process engaged by establishing a VPN ..."

Section 4.4: (paragraph 2)
Change "... related to the mobility of the device and ..." to "... 
related to the degree of device mobility and ..."

Section 4.4: (paragraph 2)
Combine the final 2 sentances as follows:
"When a device is aware that it has moved, for instance when it changes 
access points, the device SHOULD refresh its location."

Section 4.4: (last paragraph)
Change "... getting more recent location ..." to "... getting updated 
location ..." or "... obtaining a fresh location"

Section 5: (first paragraph)
Consider re-writing as:
"A device (or a downstream signaling element) identifies an emergency 
call by an "address", which in most cases is a dialstring, although 
other user interfaces may be used."

Section 5: (paragraph 2)
First sentence has too many words modifying the word "element". Consider:
"Note: It is undesirable for a user-interface to enable a user to place 
an emergency call by pressing a single button."

Section 5: (paragraph 3)
Consider changing the first sentence:
"... in other countries there are several 3 digit numbers used for 
different types of emergency calls."

Section 5: (paragraph 6)
Change "... some entity needs to ..." to "... some entity on the 
signaling path must ..."

Section 5: (paragraph 9)
Change "... from North America, the home ..." to "... from North 
America, then while in North America the home ..."

Section 5: (last paragraph)
Add a close parenthesis ")" after "dialstrings."

Section 6:
Change "... is expected be supported ..." to "... is expected to be 
supported ..."

Section 6.1:
Change "signaling Method" to "signaling method" (lower case).

Section 6.1: (number 1)
Change "To: SHOULD" to "Request-URI: SHOULD"

Section 6.1: (number 2)
Add a space between [I-D.rosen-iptel-dialstring] and "with"
Also change "sips MUST be ..." to "a sips URI MUST be ..."

Section 6.2: (number 1)
Change "If it finds it it MUST:" to "If it finds the dialstring it MUST:"
Also, change "... for the endpoint" to "... of the end device." (note 
that period was missing)

Section 6.3: (paragraph 3)
Non-parallel sentence structure. Consider re-writing the last sentence as:
"In the case of civic location, the LoST server SHOULD report that the 
same mapping is good within a community name or even a street, as this 
is helpful for WiFi connected devices that roam and obtain civic 
location from the AP to which they connect."
(Or perhaps "Despite the fact that civic location is uncommon for mobile 
devices, the LoST server SHOULD ...")

Section 6.3: (paragraph 4)
Change "... URI of the service URN ..." to "... URI of a service URN ..."

Section 6.3: (last paragraph)
Re-write the last sentence as: "The proxy then replaces the Request-URI 
with the resulting PSAP URI."
Or perhaps "The resulting PSAP URI then replaces the Request-URI"

Section 6.4: (first paragraph)
Consider adding the phrase "Once the mapping to a PSAP URI has been 
performed," to the begining of the paragraph (to improve the flow of the 
document).

Section 6.6:
Change "The emergency dialstrings ..." to "Emergency dialstrings ..."

Section 7: (paragraph 3)
Change "For calls send with ..." to "For calls sent with ..."

Section 8: (paragraph 1)
Change "... media streams on RTP ... " to "... media stream using RTP ..."
or perhaps "... media streams via RTP ..."
Also, consider re-writing the 4th sentence as:
"Future IP-enabled PSAPs should accept a wider array of potential media 
types."

Section 8: (paragraph 2)
Add a period between "the offer" and "Silence suppression".

Section 10:
Change "... it specifies use of several ..." to "... it specifies the 
use of several ..."

Section 10.1: (paragraph 2)
Change "... DHCP is the LCP [RFC3118] ..." to "... DHCP is the LCP, 
[RFC3118] ..." (add a comma)
Also, change "... spoofing would be ..." to "... spoofing is ..."

Section 10.1: (last paragraph)
Add an 's' to change "Client SHOULD" to "Clients SHOULD"

Section 10.2: (last paragraph)
Change "... signaling would help significantly." to "... signaling helps 
significantly."



_______________________________________________
Ecrit mailing list
Ecrit at ietf.org
https://www1.ietf.org/mailman/listinfo/ecrit