[dhcwg] co-existence of temp and normal addresses

[Vijay Bhaskar A K <[e-mail]>]

Consider the following scenario.

A client  has 3 normal  addresses  and 2  temporary  addresses  in an IA
identified  by IAID 36.  After some time, the client  releases all the 3
normal  addresses.  Now, it needs  another 1 temporary  address.  So, it
will send a Request,  with IA option having IAID 36,  encapsulating  RTA
option having value 3, indicating it needs an another temporary address.

Now, what server assumes is....

- the client is requesting to assign a temporary address.
- the client is requesting to asssign  normal  addresses  also.  This is
because,  whenever  a  Request  is sent for IA, the  server  is going to
assign new set of normal  addresses to the IA  according  to its policy,
unless or  otherwise,  the server's  database  says that the IA has been
already  populated  with  addresses.  In the later case it assumes to be
the  retransmission  and sends back whatever the normal addresses it has
already allocated.

Thus, server  allocates  normal  addresses also along with the requested
temp  address.  But,  the  client  didn't  have   requested  for  normal
addresses.

*** Thus the mixture of normal  addresses  and temp  addresses  in an IA
leads to mislead in the server's side.

*** The temporay addresses should not be renewed.  In the renew message,
the IA should not carry temporary  address.  This means,  eventhough the
temporary  addresses  lies in the same IA along  with the  other  normal
addresses,  they are treated as seperate set of addresses in the case of
renewal.

Considering the above two points, i have few suggestions.

- Why can't we have separate IAs for normal and temporary addresses?
- The IAIDs for temporary addresses range from 0-1023 and that of normal
addresses range from 1024 onwards.  This is to avoid the  overlapping of
IAID space.

- Since, T1 and T2 values for the IA containing only temporary addresses
make no sense, i feel  that we can have  seperate  TMP_IA  option  which
doesn't have T1 and T2 fields, as shown below.


      0                   1                   2                   3
      0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
     +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
     |       OPTION_TMP_IA           |          option-len           |
     +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
     |                        IAID (4 octets)                        |
     +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
     |  IA  Status   |                                               |
     +-+-+-+-+-+-+-+-+                                               |
     .                            Options                            .
     .                                                               .
     +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+


      option-code          OPTION_TMP_IA (TBD)

      option-len           See section 23.

      IAID                 The unique identifier for this TMP_IA. 
                           It ranges from 0-1023
      
      IA status            Status of the IA in this option.

      options              Options associated with this IA.


- T bit in the address status of IA address option is also not necessary
since we are having seperate IAs for temp addresses.

Please let me know your comments.

~Vijay


_______________________________________________
dhcwg mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/dhcwg