Re: rfc 1122 and ICMP SOURCE HOST ISOLATED

[[e-mail]]

  *> From [e-mail] Thu Nov 10 20:40:12 1994
  *> Date: Thu, 10 Nov 1994 11:11:32 -0800 (PST)
  *> From: Gabriel Montenegro <[e-mail]>
  *> Reply-To: Gabriel Montenegro <[e-mail]>
  *> Subject: rfc 1122 and ICMP SOURCE HOST ISOLATED
  *> To: [e-mail]
  *> Cc: [e-mail]
  *> Content-Type  *> :   *> text  *> 
  *> X-Sun-Text-Type: ascii
  *> Content-Length: 929
  *> X-Lines: 22
  *> 
  *> I have a question about one of the new ICMP destination unreachable codes
  *> defined in rfc1122. As editor of that document, I was wondering if you could
  *> give me some information, or pointers as to where to get it.
  *> 
  *> My question is:
  *> 
  *> 	Is destination unreachable code 8 (source host isolated) usable as a
  *> 	"hard" error? Any info on what the motivation was to define this code?
  *> 
  *> RFC1122 specifies that HOST/NET unreachable codes are "soft" errors which
  *> must be taken as hints. Codes 6-12 are defined in that document, but nothing
  *> is said about whether they should be considered "hard" or "soft". This omission
  *> is documented in rfc1127. 
  *> 
  *> Thanks in advance for any info you can offer,
  *> 
  *> -gabriel montenegro
  *> 
  *> Gabriel Montenegro (Nomadic Computing Group)
  *> Sun Microsystems Inc.               Email: [e-mail]
  *> 2550 Garcia Ave, MS UMPK14-207      Tel: [phone]
  *> Mountain View, CA 94043-1100        Fax: [phone]
  *> 

Gabriel,

I don't at the moment recall exactly what code 8 was supposed to mean
(Jon Postel, did that come from your little list?), but it seems to be
administrative rather than operational.  That is, it would be returned
as a statement about the topology/configuration, not as a slightly
stronger Unreachable.  I would therefore expect it to be "hard".

I am CCing the former WG mailing list, to solicit other opinions.

Bob Braden