Hello Ankur,
Sorry for the late reply. Let me take it from a bigger perspective.
RFC 3046 defines the Option-82 as a container that can hold a set of suboptions. The same RFC establishes two initial suboptions:
- Type 1: Circuit ID
- Type 2: Remote ID
Suboption numbers are currently numbered by IANA and this is the current type allocation list:
http://www.iana.org/assignments/bootp-dhcp-parameters/bootp-dhcp-parameters.xml#relay-agent-sub-options
The server behavior over diverse Option-82 suboptions is defined with respect to individual suboptions, i.e. the processing of different suboptions by a DHCP server may differ. Now, both Circuit ID and Remote ID suboptions are defined as opaque values that should not be internally interpreted by the server:
Section 3.1, Agent Circuit ID Sub-option
Servers MAY use the Circuit ID for IP and other parameter assignment
policies. The Circuit ID SHOULD be considered an opaque value, with
policies based on exact string match only; that is, the Circuit ID
SHOULD NOT be internally parsed by the server.
Section 3.2, Agent Remote ID Sub-option
DHCP servers MAY use this option to select parameters specific to
particular users, hosts, or subscriber modems. The option SHOULD be
considered an opaque value, with policies based on exact string match
only; that is, the option SHOULD NOT be internally parsed by the
server.
Furthermore, the entire Option-82 is intended to be processed by the very device that originated it. Other devices are not supposed to act upon a Option-82 value they have not originated themselves (although there is no 100% guarantee provisioned to make sure that a device won't confuse a different Option-82 for its own).
From all of this it follows that both Circuit ID and Remote ID may contain arbitrary values and they are totally vendor-dependent. So, yes, the format of Circuit ID and Remote ID as used by Cisco for DHCP Snooping purposes is proprietary - it uses the correct suboption header of the suboption type and length and follows with proprietary data - but that is expectable and actually meant to be so. If there is any more general meaning to the Circuit ID type of 0 and Remote ID type of 0 as used currently by Cisco DHCP Snooping I would like to know if myself (including the other possibilities) but otherwise, all these values are purely at the vendor's discretion because no other device is going to parse them.
Does this help a bit? Please feel welcome to discuss further!
Best regards,
Peter
