On Friday 17 December 2010 21.59, Kong Posh Bhat wrote:

Greetings,

We are trying to develop an architecture that is based around the premise that there is a special purpose local HTTPS server with a well known address. We intend to reserve the port number with IANA and define a well known local address in each of the following address spaces: 10.0.0.0, 172.16.0.0, 192.168.0.0, as well as in some private IPv6 address ranges.

When a device comes up, as part of its startup logic it tries all these well known addresses one by one, until there is a hit. If there is no hit, the process terminates. However, if there is a hit, the device will make a special HTTPS request to this server, which in turn will deliver some management bootstrapping information to the device.

One of my distinguished colleagues thinks that this constitutes an IP address abuse. Is that so? I do not seem to find any reference to this on the IANA Abuse FAQ site (http://www.iana.org/abuse/faq.html).

Thanks in anticipation. I really appreciate your patience with me on this issue.

I'm not a ripe representative. I do however have a suggestion : why don't you send a multicast packet to a known multicast address ? That way a client with any address ( not only 1918 ones) be functionally equivalent The range would be limited by the networks multicast config, but you could always reach a server on your local wire.

Regards,

Kong Posh Bhat

Standards Research Lab

Samsung Telecommunications America

Ph: 972-761-7450 (Desk); 214-766-1743 (Mobile)

Fax: 972-761-7631

-- Peter Håkanson There's never money to do it right, but always money to do it again ... and again ... and again ... and again. ( Det är billigare att göra rätt. Det är dyrt att laga fel. )

On Fri, Dec 17, 2010 at 02:59:29PM -0600, Kong Posh Bhat wrote:

We are trying to develop an architecture that is based around the premise that there is a special purpose local HTTPS server with a well known address. We intend to reserve the port number with IANA and define a well known local address in each of the following address spaces: 10.0.0.0, 172.16.0.0, 192.168.0.0, as well as in some private IPv6 address ranges.

the topics of service discovery, service location, zero configuration networking and the like has been addressed in multiple ways, some of which have been standardized within the IETF. There might be a better starting point for evaluating existing solution frameworks against your requirements than this list.

One of my distinguished colleagues thinks that this constitutes an IP address abuse. Is that so? I do not seem to find any reference to this on the IANA Abuse FAQ site (http://www.iana.org/abuse/faq.html).

Some people would probably agree with your distinguished colleague, but there is a subtle difference between the "abuse" you quoted him mention and the abuse this working group is addressing. On the former - there is a reason some of the schemas mentioned above have been developed. For instance, you cannot be sure that the addresses you are trying to probe are within the same administrative domain/realm and there is no method to limit the scope of these packets. Even though RFC 4085 "Embedding Globally-Routable Internet Addresses Considered Harmful" aims at globally routable addresses, similar logic is likely to apply. This connects to the latter, network abuse as an operational phenomenon - see the charter at <http://www.ripe.net/ripe/wg/anti-abuse/>. -Peter