On Sun, Jun 23, 2013 at 12:47 PM, Joel M. Halpern <[email protected]>wrote:

> I think that the particular example you propose will onot help.
> Some tunneling protocols only requrie action at one end to initiate a
> bi-directional tunnel.  I that case, presumably, the I2RS client would only
> need to interact with that one end.  Other protocols require action /
> configuration at both ends to establish a bi-directional tunnel.  In that
> case, the i2rs client would need to interact with both ends.  (And other
> protocols are even stranger.)
>
> If your question is whether i2rs includes a protocol between i2rs agents
> to coordinate action across boxes, separate from existing protocol
> mechanisms, then I believe that the answer is a clear "no".  There is is
> agent-agent protocol in scope for the i2rs work.
>
>
No, my question is how many I2Rs agents does a client need to
contact to perform a task that requires changes on multiple routers?



> Yours,
> Joel
>

Andy


>
> On 6/23/2013 11:59 AM, Andy Bierman wrote:
>
>> Hi,
>>
>> If an I2RS client wants to set up a tunnel between 2 routers
>> (as a lame example), does it send 1 request to 1 I2RS agent
>> or 2 requests, 1 to each I2RS agent/router?  If 1 request, is the
>> southbound
>> protocol between the I2RS agent and at least 1 router proprietary
>> or part of the standard?
>>
>> I think this draft should be clear about what is in scope.
>> It seems to say that the southbound protocol is out of scope.
>>
>>
>> Andy
>>
>>
>> On Sun, Jun 23, 2013 at 2:16 AM, Abdussalam Baryun
>> <[email protected] 
>> <mailto:abdussalambaryun@**gmail.com<[email protected]>>>
>> wrote:
>>
>>     I beleive if any thing not shown then they are not allowed by I2RS, if
>>     it is shown then it is the way the I2RS will work to solve the
>>     problem. So I will understand that I2RS client does not talk to NE but
>>     only to the agent, that is why I suggest that we don't show any sign
>>     that the client can talk to any other as long it is out of scope.
>>     Therefore the drawing methodology (figure 1) is : (if out of scope of
>>     the protocol then should not be shown, and if future work of protocol
>>     it may be shown).
>>
>>     AB
>>
>>     On 6/22/13, Andy Bierman <[email protected]
>>     <mailto:[email protected]>> wrote:
>>      > Hi,
>>      >
>>      > I just re-read the framework and problem statement drafts.
>>      > Only 1 minor issue in the problem statement draft:
>>      >
>>      > The 'I2RS Agent' is shown as a single box in figure 1.
>>      >
>>      >  1) Does this mean the protocol between the "broker" and the
>>      >      NEs is proprietary, or just not shown?
>>      >
>>      >  2) Does this mean that an I2RS Client never talks directly to an
>> NE
>>      >      or does it mean all I2RS Agent functionality is available on
>>     all NEs?
>>      >
>>      > Nit, sec. 6, para 4:
>>      >    - the lack of standard data models is the problem of the
>>     NETMOD WG,
>>      >      not really the NETCONF protocol.
>>      >    - s/may help define needed/may require help defining needed/
>>      >
>>      >
>>      > Andy
>>      >
>>
>>
>>
>>
>> ______________________________**_________________
>> i2rs mailing list
>> [email protected]
>> https://www.ietf.org/mailman/**listinfo/i2rs<https://www.ietf.org/mailman/listinfo/i2rs>
>>
>>
_______________________________________________
i2rs mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/i2rs

Reply via email to