Dhruv
Please see inline
On 08/16/2011 07:17 AM, Dhruv Dhody wrote:
Dear Ramon and PCE-WG,
(1) Encoding of Domain Seq in IRO.
I like your suggestion to have new IRO object (Object class 10, object
type 2).
Imho it is a good trade-off. Using type 1 would not be backwards compatible
//
(2) Area Encoding
I have kept the subobject similar to IPv4/IPv6 subobjects. But I have
no objection to your proposed encoding either.
Indeed, that's what I thought. My proposal was motivated by: a) ISIS
area restriction and b) alignment issues, e.g. in 32 bit machines
Consider this
#include <iostream>
#include <cstdint>
struct A {
uint8_t type;
uint8_t len;
uint32_t addr;
uint16_t resv;
};
struct B {
uint8_t type;
uint8_t len;
uint16_t resv;
uint32_t addr;
};
int main () {
std::cout << sizeof(A) << std::endl;
std::cout << sizeof(B) << std::endl;
return 0;
}
The output is 12 and 8. The compiler adds padding to align the address.
Wrt to AS Number subobject (type=32) as defined in RFC 3209 allow only
2 octet AS Number.
Indeed, nothing changes w.r.t to the already existing AS number
sub-object, but I was proposing _another_ new sub-object to also take
into account 4-byte AS. See http://tools.ietf.org/html/rfc4893 Since you
are already proposing Area sub-objects, you may as well include 4-byte AS
WG Chair/Area director: How to handle this --
(a)Update to RSVP RFC 3209?
(b)Errata to RSVP RFC 3209?
(c)New subobject to be defined in this draft to handle 4-byte AS Number?
//
If you add Area sub-objects, you may as well add 4-byte AS. We can later
on address this appropriately since as you mention it is generic and may
need to be addressed by ccamp?
If adding RBNF gives more clarity to draft, am all for it!
The RBNF may be overkill and/or may be simplified :) but the important
point is that a detailed, generic guideline could be added and the RBNF
tries to provide some formal information structure
Ramon: I wanted to thank you for such detail review, as you have
suggested detailed text to the draft; we can co-author the draft so
that all your suggestions/texts is correctly incorporated.
Thanks
R.
_______________________________________________
Pce mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/pce
