This document is a more restrictive version summary of [RFC 7844],
where the keywords (key words
[RFC 2119]) commented in
RFC7844 comments
are actually replaced. Use diff
to see specific differences between these
two documents.
See :ref:`questions` for a summary of the questions stated here.
Note
- Extracts from the [RFC 7844] marked as literal blocks.
- Replacements are marked as parsed literal with the keyword replaced in bold
Note
See :ref:`implementation` for a summary of the messages implementation
[RFC 7844]:
SHOULD randomize the ordering of options
If this can not be implemented MUST order the options by option code number (lowest to highest).
[RFC 7844]:
MUST contain the Message Type option,
MUST NOT contain the Client Identifier option, MUST NOT contain the Parameter Request List option. MUST NOT contain any other option.
[RFC 7844]:
MUST contain the Message Type option,
MUST NOT contain the Client Identifier option, MUST NOT contain the Parameter Request List option. MUST NOT contain any other option.
If in response to a DHCPOFFER, MUST contain the corresponding Server Identifier option MUST contain the Requested IP address option. If the message is not in response to a DHCPOFFER (BOUND, RENEW),:
MUST NOT contain a Requested IP address option
[RFC 7844]:
MUST contain the Message Type option, MUST contain the Server Identifier option, MUST contain the Requested IP address option;
MUST NOT contain the Client Identifier option.
- is it always broadcast?
[RFC 7844]
To do not leak when the client leaves the network, this message type MUST NOT be implemented.
In this case, servers might run out of leases, but that is something that servers should fix decreasing the lease time.
[RFC 7844]:
MUST contain the Message Type option,
MUST NOT contain the Client Identifier option, MUST NOT contain the Parameter Request List option. It MUST NOT contain any other option.
[RFC 7844]:
MUST NOT include in the message a Client IP address that has been obtained with a different link-layer address.
[RFC 7844]
MUST NOT use the Requested IP address option in DHCPDISCOVER messages.
MUST use the option when mandated (DHCPREQUEST) If in INIT-REBOOT:
MUST perform a complete four-way handshake, starting with a DHCPDISCOVER
This is like not having INIT-REBOOT state?:
If the client can ascertain that this is exactly the same network to which it was previously connected, and if the link-layer address did not change, MAY issue a DHCPREQUEST to try to reclaim the current address.
This is like INIT-REBOOT state?
Is there a way to know
if
the link-layer address changed without leaking the link-layer?
[RFC 7844]:
If the hardware address is reset to a new randomized value,
the DHCP client MUST use the new randomized value in the DHCP messages
The client should be restarted when the hardware address changes and use the current address instead of the permanent one.
[RFC 7844]
MUST NOT have this option
In the case that it would have this option because otherwise the server does not answer to the requests,:
DHCP clients MUST use client identifiers based solely on the link-layer address that will be used in the underlying connection.
[RFC 7844]
MUST NOT have this option
[RFC 7844]
MUST NOT send the Host Name option.
[RFC 7844]
MUST NOT include the Client FQDN option
[RFC 7844]:
Nodes visiting untrusted networks MUST NOT send or use the PXE options.
- And in the hypotetical case that nodes are visiting a "trusted" network, must this option be included for the PXE to work properly?
[RFC 7844]
MUST NOT use the
Vendor-Specific Information option (code 43), the Vendor Class Identifier option (code 60), the V-I Vendor Class option (code 124), or the V-I Vendor-Specific Information option (code 125),
[RFC 7844]
Implementers SHOULD provide a way for clients to control when the anonymity profiles are used and when standard behavior is preferred.
dhcpcanon
does not currently implement the standard behavior described in
[RFC 2131] in order to keep the implementation simple and
because all existing implementations already implement it
[RFC 2131]:
the allocating server SHOULD probe the reused address before allocating the address, e.g., with an ICMP echo request, and the client SHOULD probe the newly received address, e.g., with ARP. The client SHOULD perform a check on the suggested address to ensure that the address is not already in use. For example, if the client is on a network that supports ARP, the client may issue an ARP request for the suggested request. When broadcasting an ARP request for the suggested address, the client must fill in its own hardware address as the sender's hardware address, and 0 as the sender's IP address, to avoid confusing ARP caches in other hosts on the same subnet.>> The client SHOULD broadcast an ARP reply to announce the client's new IP address and clear any outdated ARP cache entries in hosts on the client's subnet.
- does any implementation issue an ARP request to probe the offered address?
- is it issued after DHCPOFFER and before DHCPREQUEST, or after DHCPACK and before passing to BOUND state?
Currently, there is not any probe
Sending DHCPDISCOVER [RFC 2131]:
The client SHOULD wait a random time between one and ten seconds to desynchronize the use of DHCP at startup.
- is the DISCOVER retranmitted in the same way as the REQUEST?
[RFC 2131]:
a client retransmitting as described in section 4.1 might retransmit the DHCPREQUEST message four times, for a total delay of 60 seconds
[RFC 2131]:
In both RENEWING and REBINDING states, if the client receives no response to its DHCPREQUEST message, the client SHOULD wait one-half of the remaining time until T2 (in RENEWING state) and one-half of the remaining lease time (in REBINDING state), down to a minimum of 60 seconds, before retransmitting the DHCPREQUEST message.
[RFC 2131]:
For example, in a 10Mb/sec Ethernet internetwork, the delay before the first retransmission SHOULD be 4 seconds randomized by the value of a uniform random number chosen from the range -1 to +1 Clients with clocks that provide resolution granularity of less than one second may choose a non-integer randomization value. The delay before the next retransmission SHOULD be 8 seconds randomized by the value of a uniform number chosen from the range -1 to +1. The retransmission delay SHOULD be doubled with subsequent retransmissions up to a maximum of 64 seconds.
- the delay for the next retransmission is calculated with respect to the type of DHCP message or for the total of DHCP messages sent indendent of the type?
- without this algorithm being mandatory, it'd be possible to fingerprint the the implementation depending on the delay of the retransmission
- how does other implementations do?
[RFC 2131]:
DHCP clients are free to use any strategy in selecting a DHCP server among those from which the client receives a DHCPOFFER message. client may choose to collect several DHCPOFFER messages and select the "best" offer. If the client receives no acceptable offers, the client may choose to try another DHCPDISCOVER message.
- what is a "no acceptable offer"?
- which are the "strategies" to select OFFER implemented?
- different algorithms to select an OFFER could fingerprint the implementation
[RFC 2131]:
The client collects DHCPOFFER messages over a period of time, selects one DHCPOFFER message from the (possibly many) incoming DHCPOFFER messages The time over which the client collects messages and the mechanism used to select one DHCPOFFER are implementation dependent.
- Is it different the retransmission delays waiting for offer or ack/nak?, in all states?
Currently, the first OFFER is chosen
[RFC 2131]:
Times T1 and T2 are configurable by the server through options. T1 defaults to (0.5 * duration_of_lease). T2 defaults to (0.875 * duration_of_lease). Times T1 and T2 SHOULD be chosen with some random "fuzz" around a fixed value, to avoid synchronization of client reacquisition.
T1 is then calculated as:
renewing_time = lease_time * 0.5 - time_elapsed_after_request range_fuzz = lease_time * 0.875 - renewing_time renewing_time += random.uniform(-(range_fuzz), +(range_fuzz))
And T2:
rebinding_time = lease_time * 0.875 - time_elapsed_after_request range_fuzz = lease_time - rebinding_time rebinding_time += random.uniform(-(range_fuzz), +(range_fuzz))
The range_fuzz is calculated in the same way that systemd
implementation
does
- what's the fixed value for the fuzz and how is it calculated?
- The "fuzz" range is not specified, the fuzz chosen could fingerprint the implementation.
[RFC 7844]:
There are scenarios in which a client connecting to a network remembers a previously allocated address, i.e., when it is in the INIT-REBOOT state. In that state, any client that is concerned with privacy SHOULD perform a complete four-way handshake, starting with a DHCPDISCOVER, to obtain a new address lease. If the client can ascertain that this is exactly the same network to which it was previously connected, and if the link-layer address did not change, the client MAY issue a DHCPREQUEST to try to reclaim the current address.
- is there a way to know if the network the client is connected to is the same to which it was connected previously?
For the sake of simplicity and privacy dhcpcanon
does not currently
implement the INIT-REBOOT state nor reuse previously allocated addresses.
In future stages of dhcpcanon
would be possible to reuse a previously
allocated address.
In order to do not leak identifying information when doing so,
it would be needed:
- to keep a database with previously allocated addresses associated to:
- the link network where the address was obtained (without revealing the MAC being used).
- the MAC address that was used in that network
It is possible also that dhcpcanon
will include a MAC randomization module
in the same distribution package or would require it in order to start.