Re: Definitions of "Console" and "CLI" expanded

[George Jones <gmj@pobox.com>]

On Wed, 26 Nov 2003, Mike O'Connor wrote:

> :Well, sure, but HTTP is an application protocol. Ethernet is a kind of
> :communications hardware. Saying "Ethernet or USB console" is one
> :thing. Saying "HTTP instead of SSH CLI" is another. Mixing apples and
> :oranges here is bad.
>
> Agreed.  I actually refer to the loose wording of "IP stack" in some
> comments I sent to opsec-request that George hasn't responded to/fed
> back to the list.

Apologies.  Holidays.  Work.  Had to respond to review solicitied
by Area Directors and get -03 out.   I read commands and tried
to incorporate.

On the CLI/RS232 vs. Ethernet/HTTP question, it seemed like
a) RS232 is here now (and has been for 30+ years) b) CLI is
here now and does the job c) reality is there is movement
away from RS232 and is likely to be more, no matter what
a new BCP says.

What I tried to do was:

  - Document the benefits of the RS232/CLI approach
  - Strongly recommend RS232/CLI
  - Allow for other types of connections (USB, Ethernet)
  - Per sentemnt on this list, NOT imply that HTTP/HTML
    was viewed suitable, complete replacement for CLI
    (else xmlconf would probably not be a working group)

On the services question, I added a definition up front:

03>   Service.
03>
03>      A number of requirements refer to "services". For the purposes
03>      of this document a "services" is defined as "any process or
03>      protocol running in the control or management planes to which
03>      non-transit packets may be delivered".  Examples might include
03>      an SSH server, a BGP process or an NTP server.  It would also
03>      include the transport, network and link layer protocols since,
03>      for example, a TCP packet addressed to a port on which no
03>      service is listening will be "delivered" to the IP stack, and
03>      possibly result in an ICMP message being sent back.

Hope that was close.

Thanks.

---George Jones