Re: [Detnet] DP solution (Flow identification)

Stewart,

Thanks, I'm aware of that trend, the only question is how much of "the
old" situation actually remains in the deployed networks?

/Loa

On 2018-01-23 19:26, Stewart Bryant wrote:
> Loa,
> 
> That used to be the case, but I am not so sure that it is still the 
> case, mainly due to the cost of the linecards.
> 
> If a node does have mixed functions egress is normally on a specialist 
> line card, and so long at the duplicate deletion happens on the egress 
> of that line card we have a single point of perfection where can 
> de-duplicate.
> 
> - Stewart
> 
> 
> On 23/01/2018 10:57, Loa Andersson wrote:
>> Balazs,
>>
>> Yes I understand your point, what I tried to sat (a bit glibly) is 
>> that find true P-nodes is not that easy. Most nodes have some edge
>> functionality anyway.
>>
>> /Loa
>>
>> On 2018-01-22 22:25, Balázs Varga A wrote:
>>> Hi Loa,
>>> I think this case is rather different. The "Packet Ordering" function
>>> is rather specific and You need the right order only at the egress.
>>> Cheers
>>> Bala'zs
>>>
>>> -----Original Message-----
>>> From: Loa Andersson [mailto:loa@pi.nu]
>>> Sent: 2018. január 22. 14:12
>>> To: Stewart Bryant <stewart.bryant@gmail.com>; Balázs Varga A 
>>> <balazs.a.varga@ericsson.com>; Andrew G. Malis <agmalis@gmail.com>
>>> Cc: DetNet WG <detnet@ietf.org>
>>> Subject: Re: [Detnet] DP solution (Flow identification)
>>>
>>> Stewart,
>>>
>>> On 2018-01-22 18:42, Stewart Bryant wrote:
>>>>
>>>>
>>>> On 19/01/2018 09:09, Balázs Varga A wrote:
>>>>>
>>>>> Also during the Elimination
>>>>>
>>>>> we can simple concentrate on duplicate detection and out-source
>>>>> in-order delivery to DN4.
>>>>>
>>>>
>>>> In the middle of the network it may be sufficient to do approximate
>>>> duplication detection. That is to say delete a packet if it has the
>>>> same s/n as its predecessor (may be with a very tiny window) and let
>>>> the egress do anything more complex.
>>>
>>> This has been tried before, but if I remember correctly. there turned 
>>> out to be very few true P nodes.
>>>
>>> /Loa
>>>>
>>>> - Stewart
>>>>
>>>>
>>>> _______________________________________________
>>>> detnet mailing list
>>>> detnet@ietf.org
>>>> https://www.ietf.org/mailman/listinfo/detnet
>>>>
>>>
>>
> 
> _______________________________________________
> detnet mailing list
> detnet@ietf.org
> https://www.ietf.org/mailman/listinfo/detnet

-- 

Loa Andersson                        email: loa@pi.nu
Senior MPLS Expert
Huawei Technologies (consultant)     phone: +46 739 81 21 64