ietf
[Top] [All Lists]

Re: [paws] Last Call: <draft-ietf-paws-problem-stmt-usecases-rqmts-12.txt> (Protocol to Access White Space (PAWS) Database: Use Cases and Requirements) to Informational RFC

2013-02-13 12:15:02
Hi SM,

I responded inline to your latest comments, below, where I took additional
action.

Thanks,

Tony


On Wed, Feb 13, 2013 at 4:33 AM, SM <sm(_at_)resistor(_dot_)net> wrote:

Hi Anthony,

At 17:00 12-02-2013, Anthony Mancuso wrote:

In Section 1.1:

  "Academia and Industry have studied multiple cognitive radio [2]
   mechanisms for use in such a scenario."

The reference seemed odd.  It took me some time to understand that it was
put in to address a comment.  However, the first (external) reference that
defines that is a 404.

TODO(amancuso): fix link


Ok.


 I think the MUST below is one of the two boilerplate words you had a
comment on. Is there another?


See the section after the Abstract and Section 2.1.


    TM: Now i see the two. I removed the second (redundant) RFC 2119
wording and reference.



 In response to earlier comments, this section was reworded and "MUST"
changed to "must". I assume this is what you were questioning.


Ok.


 ----

In Section 4.2:

  "A simplified operation scenario of offloading content, such as video
   stream, from the a metered Internet connection to the a WS connection
   consists of the following steps:"

What is a metered Internet connection?
----
Usage is monitored (metered) and paid for. I rephrased to clarify: "more
congested or costly Internet connection, such as a metered
(fee-based-on-usage) wire, wireless, or satellite service."


In my opinion you don't really have to get into all that.  I'll defer to
the author.


     TM: Agreed. Simply referred to the "metered" Internet service as a
paid service.



 cleaned up protocol and operational requirements in response to earlier
comments. This is a good example of a use case, but didn't fit comfortably
as a P or O requirement, and was deleted from requirements sections.


Ok.


 White Space Database (used to be defined in terminology). Changed it
simply to "database".


Ok.


 This is a rather large container of worms. We kicked it around for about
a half-hour and concluded that since this doc is for the Internet
Engineering Task Force, most readers will understand (and forgive).
Nonetheless, this is a good point, so I generalized it as follows (while
retaining the commonly used (and loosely-understood) term: "The database
and the master device MUST be connected (e.g., through the Internet)."


I agree about that container. :-)  The draft is trying to convince me that
white space is better for the Internet connection (see my previous comment
about metered).  The requirements is about a protocol to Access Spectrum
Database.  IP connectivity is a lower-layer issue.  I am ok with the above
text.




 This section only intends to state the general requirement. It is my
understanding that the current draft of the protocol specification
specifies JSON encoding of location parameters that is compatible with
GEOPRIV.


Ok.

Regards,
-sm