Re: [mif] Route option for DHCPv6 - next steps?

[Margaret Wasserman <[e-mail]>]

Hi Alex,

On Mar 29, 2012, at 2:24 PM, Alexandru Petrescu wrote:

> Le 29/03/2012 14:22, Tomek Mrugalski a écrit :
>> On 12-03-29 13:35, Ted Lemon wrote:
>>> Probably the best thing to do is set the router lifetime to the DHCP
>>> renewal time, and not send a lifetime separate from that.
>> We are talking about two timers here: route lifetime and renewal time
>> (governed by information refresh time, defined in RFC4242). Will add
>> clarification text that route lifetime should be greater than
>> information refresh time. Making them equal is not a such good idea as
>> it could trigger race conditions. Otherwise badness will happen.
> 
> We may be talking three timers: route lifetime, router lifetime (ND) and renewal time.

I am not exactly sure what you mean by a "timer" in this sentence?  In the implementations I have done, there is no "timer" (in the embedded systems sense) associated with each route entry.  There is a timestamp that indicates the current end of the valid lifetime for that route (the expiry time).  Before a route is used, the expiry time is checked to see if it is still valid.  Each routing table entry would have (at most) one expiry time associated with it.  Static routes have no expiry time.

Margaret