[Ns-developers] Design Review of ns-3-ipv6-1st (ns-3.3 new feature)

[Sébastien Vincent]

Mathieu Lacage a écrit :
> hi craig, all,
>
> At this point, my main concern is that I would like to know what the
> plan is to deal with the ipv4 API rework under way in tom's tree. i.e.,
> if we merge this ipv6 API which is rather closely structured on the
> current ipv4 API, and, if we change the ipv4 API just after merging this
> ipv6 API, we are going to have two different ipv/ipv6 APIs for next
> release. I have to confess that I would find this suboptimal so, I
> wonder if we could get some sort of commitment from the relevant
> maintainers to work together to ensure that the ipv6 and ipv4 APIs are
> aligned before the next release. I could go as far as suggesting that we
> make this a blocker for next release.
>
>   

Just one thing, the first chunk of IPv6 (ns-3-ipv6-1st) does *NOT* 
contain any of layer 3 (Ipv6L3Protocol) and transport protocol 
(UDP/TCP/ICMPv6).

I asked before if it will be good to have things added to this first 
chunk (Ipv6L3Protocol, ICMPv6, Neighbor Discovery, ...) so that users 
can make scenarios with a Ping6 application. But as Mathieu said it will 
be better to wait about IPv4 API refactoring. What do you think ?

--
Sebastien

> regards,
> Mathieu
>
> On Thu, 2008-10-09 at 11:06 -0700, craigdo at ee.washington.edu wrote:
>   
>> Hi all,
>>
>> We have received a request to merge a new feature into ns-3.3 (the upcoming
>> stable release of ns-3 scheduled for release in December).  Based on our
>> recently updated release process, we have to do a design review of the
>> feature addition, and the current maintainers of the ns-3 codebase have to
>> sign off on it, before we merge it into ns-3-dev and then into ns-3.3.
>>
>> I sent the below email to the maintainers yesterday requesting an "official"
>> design review for Sebastien Vincent's ns-3-ipv6-1st feature.  The
>> maintainers are in general agreement that they want to do the review in a
>> public setting and also solicit input from people on the ns-developers
>> mailing list.  This means some technical emails will probably start flying
>> around regarding possibly abstract questions of IPv6 support in ns-3.  I
>> wanted to provide some context before that happens.
>>
>> So, this email is an introduction to the design review thread for the first
>> piece of the IPv6 support in ns-3.  As far as timing goes, this all needs to
>> be wrapped up by about the end of the month.  Anyone reading should feel
>> free to take a look at the submitted code and voice opinions.
>>
>> Regards,
>>
>> -- Craig
>>
>>     
>>> You are receiving this email because you are a maintainer of 
>>> ns-3 and a new
>>> feature addition is being considered for ns-3.3.
>>>
>>> We have a submission of "ns-3-ipv6-1st" to be considered as a 
>>> new feature
>>> addition for ns-3.3.  As indicated in our release process, 
>>> this submission
>>> crosses maintainer responsibilities and so a design review 
>>> should be done.
>>> Also this could be a fairly significant change in terms of 
>>> usability and API
>>> that could be very visible in the future.
>>>
>>> Before proceeding I'll need some kind of positive 
>>> acknowledgement from each
>>> of you that you don't have any issues and that we should go 
>>> ahead and merge
>>> this patch.
>>>
>>> If there are any discussions about the change, it would 
>>> probably be good to
>>> copy everyone.
>>>
>>> The following is the email from Sebastien 
>>>
>>>       
>>>> -----Original Message-----
>>>> From: Sébastien Vincent [mailto:vincent at clarinet.u-strasbg.fr] 
>>>> Sent: Wednesday, October 08, 2008 2:30 AM
>>>> To: craigdo at u.washington.edu
>>>> Cc: 'Tom Henderson'; 'Mathieu Lacage'; 'Craig Dowell'
>>>> Subject: ns-3-ipv6-1st 
>>>>
>>>> Hi Craig,
>>>>
>>>> Code :
>>>> ====
>>>>
>>>> The repository that contains the first chunk of IPv6 is located at 
>>>> http://code.nsnam.org/vincent/ns-3-ipv6-1st/ and it is 
>>>> synchronized with 
>>>> ns-3-dev. The corresponding patch file is attached.
>>>>
>>>> Summary of changes :
>>>> ==============
>>>>
>>>> The features of this first chunk of IPv6 provide the addresses, the 
>>>> routes, the IPv6 header and some changes in *NetDevice to 
>>>> allow specific 
>>>> ethernet multicast addresses to be forward up to layer 3.
>>>>
>>>> Files added : 
>>>> - src/node/ipv6-route.cc/h;
>>>> - src/node/ipv6-address.cc/h;
>>>> - src/node/inet6-socket-address.cc,h;
>>>> - src/node/ipv6-header.cc,h;
>>>> - src/node/icmp-socket.cc,h;
>>>> - example/test-ipv6.cc.
>>>>
>>>> Changes on existing code :
>>>> - Add virtual method GetMulticast6 and MakeMulticast6Address  in 
>>>> NetDevice and all of its subclasses (CsmaNetDevice, ...);
>>>> - In CsmaNetDevice, we declare some ethernet representation of 
>>>> well-known IPv6 multicast address, and forward up all packet 
>>>> (that match 
>>>> these addresses) to layer 3.
>>>>
>>>> This first chunk is not sufficient to do IPv6 simulations, 
>>>>         
>>> it is just 
>>>       
>>>> the addresses and the changes in NetDevice. In very old 
>>>> previous mail, 
>>>> we planned to include the layer 3 (Ipv6L3Protocol, ...) in 
>>>> the second chunk.
>>>>
>>>> Maybe it will be good to have IPv6 layer 3 protocol, ICMPv6 / 
>>>> Neighbor 
>>>> Discovery protocol and a Ping6 application & example. What do 
>>>> you think ?
>>>>
>>>> Regards,
>>>> --
>>>> Sebastien Vincent
>>>> Network and Research team -  University of Strasbourg
>>>>
>>>>
>>>>         
>>
>>     
>
>
>   


-- 
Sébastien Vincent
Network Engineer at University Louis  Pasteur, Strasbourg
Phone : +33 390244586
Address : LSIIT - Boulevard Sébastien Brant BP 10413
67412 Illkirch CEDEX France