Hi Faraz,
I agree that this is a viable use case for OSPFv3. However, I don't see it as 
one that can be handled with autoconfiguration. I will try and clarify this in 
example in the next revision. 
Thanks,
Acee 

On Sep 27, 2012, at 10:04 AM, Faraz Shamim (sshamim) wrote:

> Acee,
> 
> I am not sure why you think that the same instance can not be used for the
> ISP link. The fact that you will enable OSPFv3 in a passive mode, instance
> ID really won't matter here.
> 
> Faraz
> 
> On 9/24/12 11:42 AM, "Acee Lindem" <[email protected]> wrote:
> 
>> Hi Faraz,
>> 
>> On Sep 24, 2012, at 12:11 PM, Faraz Shamim (sshamim) wrote:
>> 
>>> Hi Acee,
>>> 
>>> In my point 2) below, I am not saying to enable OSPFv3 towards ISP for
>>> the
>>> purpose of forming OSPFv3 neighbor relationship with ISP. I am saying
>>> that
>>> in eBGP case, you do need to enable OSPFv3 on the link towards the
>>> service
>>> provider(in passive mode) because this ISP link is needed to be known as
>>> an IGP because BGP will cary this next-hop "as is" in its iBGP updates
>>> unless we do next-hop-self. This is one aspect of interaction between
>>> BGP
>>> and IGP where you need to carry the eBGP next-hop in your IGP.
>> 
>> I understood. What I meant was that this would not be an autoconfigured
>> instance of OSPFv3 and it would most likely not be the same instance as
>> is running on non-ISP facing interfaces.
>> 
>> Thanks,
>> Acee 
>> 
>> 
>>> 
>>> Thanks,
>>> 
>>> Faraz
>>> 
>>> On 9/24/12 10:47 AM, "Acee Lindem" <[email protected]> wrote:
>>> 
>>>> Hi Faraz, 
>>>> Thanks for review.
>>>> 
>>>> On Sep 21, 2012, at 1:43 PM, Faraz Shamim (sshamim) wrote:
>>>> 
>>>>> Hi Acee,
>>>>> 
>>>>> Not opposing this idea but have 4 comments here:
>>>>> 
>>>>> 1)
>>>>> 
>>>>> In section 5.2.1:
>>>>> 
>>>>> The OSPFv3 Auto-Configuration (AC) LSA has a function code of TBD and
>>>>> the S1/S2 bits set to B'01' indicating Area Flooding Scope.
>>>>> 
>>>>> I think it should say S2/S1 as mentioned in RFC 5340. Since S2 comes
>>>>> before S1 in packet format so it make sense to use the same sequence
>>>>> in
>>>>> your text.
>>>> 
>>>> Yes - will update.
>>>> 
>>>> 
>>>>> 
>>>>> Also, not sure what you meant by letter 'B' next to '01'.
>>>> 
>>>> I mean these are bits. I will see whether or not this is used
>>>> elsewhere.
>>>> 
>>>>> 
>>>>> 2)
>>>>> 
>>>>> In section 2.2:
>>>>> 
>>>>> 
>>>>> For example, if manual
>>>>>     configuration or an other condition indicates that an interface
>>>>>     is connected to an Internet Service Provider (ISP), there is
>>>>>     typically no need to employ OSPFv3.  However, note that in many
>>>>>     environments it can be useful to test whether an OSPFv3 adjacency
>>>>>     can be established.  In home networking environments, an
>>>>>     interface where no OSPFv3 neighbors are found but a DHCP prefix
>>>>>     can be acquired may be considered as an ISP interface.
>>>>> 
>>>>> 
>>>>> 
>>>>> In a situation when you are running eBGP with an ISP and iBGP
>>>>> internally,
>>>>> it is required to enable OSPFv3 on the ISP link unless you are using
>>>>> next-hop-self in BGP. So in that case excluding ISP link is a bad
>>>>> idea.
>>>>> Also in this case enabling OSPFv3 with passive-interface would make
>>>>> sense.
>>>> 
>>>> You certainly wouldn't run OSPFv3 in the same routing domain with the
>>>> ISP
>>>> box. 
>>>> This is really only an example but I could qualify OSPFv3 as
>>>> "autoconfigured OSPFv3".
>>>> 
>>>> 
>>>> 
>>>>> 
>>>>> 3)
>>>>> 
>>>>> In section 2.4:
>>>>> 
>>>>> OSPFv3 interfaces MUST auto-configure the default HelloInterval
>>>>>     and RouterDeadInterval as specified in [OSPFV3].
>>>>> 
>>>>> I think the "auto-configure" part is confusing here. We do not
>>>>> "configure"
>>>>> the default values when we enable OSPFv3 on an interface. A better
>>>>> sentence in my opinion would be:
>>>>> 
>>>>> OSPFv3 interfaces MUST use the default HelloInterval
>>>>>     and RouterDeadInterval as specified in [OSPFV3].
>>>> 
>>>> Ok. 
>>>> 
>>>>> 
>>>>> 
>>>>> 4) In section 5.2.1, wouldn't it be better to use the field
>>>>> name/letter
>>>>> instead of the actual value?
>>>> 
>>>> These fields are referred to by the bit names in RFC 5340.
>>>> 
>>>> 
>>>> 
>>>>> For example instead of mentioning 1 0 1 it
>>>>> should say U S2 S1 and in the explanation section it should mention
>>>>> the
>>>>> default value for AC LSA. Same logic goes for using TBD instead of
>>>>> "LSA
>>>>> Function Code"
>>>> 
>>>> This is pretty much standard text for adding an OSPFv3 LSA. The text
>>>> excerpted below
>>>> includes the name of the bits, value, and the designation.
>>>> 
>>>> The OSPFv3 Auto-Configuration (AC) LSA has a function code of TBD and
>>>> the S1/S2 bits set to B'01' indicating Area Flooding Scope.  The U
>>>> bit will be set indicating that the OSPFv3 AC LSA should be flooded
>>>> even if it is not understood.
>>>> 
>>>> Note that we want to specify these bits precisely rather than the
>>>> defaults. 
>>>> 
>>>> Thanks,
>>>> Acee 
>>>> 
>>>> 
>>>>> 
>>>>> 
>>>>> 
>>>>> Thanks,
>>>>> 
>>>>> Faraz
>>>>> 
>>>>> On 9/17/12 1:58 PM, "Acee Lindem" <[email protected]> wrote:
>>>>> 
>>>>>> We had some support of making
>>>>>> http://www.ietf.org/id/draft-acee-ospf-ospfv3-autoconfig-03.txt an
>>>>>> OSPFv3
>>>>>> WG document at IETF 84. The impetus is:
>>>>>> 
>>>>>> 1. OSPFv3 being considered as homenet routing protocol and
>>>>>> autoconfiguration is a requirement.
>>>>>> 2. Work on OSPFv3 configuration has started in the past in other WGs
>>>>>> but not finished.
>>>>>> 3. Current proposal has gone through a couple revision cycles with
>>>>>> incremental comments addressed.
>>>>>> 
>>>>>> Is anyone opposed?
>>>>>> 
>>>>>> Thanks,
>>>>>> Acee _______________________________________________
>>>>>> OSPF mailing list
>>>>>> [email protected]
>>>>>> https://www.ietf.org/mailman/listinfo/ospf
>>>>> 
>>>> 
>>> 
>> 
> 

Attachment: smime.p7s
Description: S/MIME cryptographic signature

_______________________________________________
OSPF mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/ospf

Reply via email to