Re: [lmap] One example of a "passive measurement peer"

[Nalini Elkins <[e-mail]>]

Guys,

If I look in http://tools.ietf.org/html/draft-eardley-lmap-terminology-02, I see the definition of measurement peer as:

Measurement Peer: The function that receives control messages and test packets from a Measurement Agent and may reply to the
Measurement Agent as defined by the Measurement Method. 

Is this correct?

Now, imagine the following scenario in a hybrid measurement technique:

1.  In the MA, the hybrid technique is to have the MA to append x bytes of diagnostic data to each passive packet.  For example, via a command to the operating system.

2.  The MA resides in the client

3. It would be ideal to have the partner that the client is having the session with (let's call it "server") also participate in this appending of x bytes of diagnostic data from his side.  (Leaving aside for the moment the potential for this to be a DoS vector.  That is a separate discussion). 

4.  In this case, could the MA not ask the "other end" to participate in extra data gathering?  This would be a control message from an MA as defined above.

Would this be an example of a Measurement Peer in a hybrid scenario?

I ask because this is what we are thinking we may want to do.

What am I missing?  Is my understanding of this OK?  Please advise.

Thanks,

Nalini Elkins
Inside Products, Inc.
[phone]
www.insidethestack.com



________________________________
From: Aamer Akhter (aakhter) <[e-mail]>
To: Juergen Schoenwaelder <[e-mail]>; "Bugenhagen, Michael K" <[e-mail]> 
Cc: 'Brian Trammell' <[e-mail]>; 'Dan Romascanu' <[e-mail]>; 'Matt Mathis' <[e-mail]>; "'[e-mail]'" <[e-mail]> 
Sent: Tuesday, March 11, 2014 8:03 AM
Subject: Re: [lmap] One example of a "passive measurement peer"


I agree on the removal of 'passive' in 'passive measurement peer' but if I remember correctly where was some discontent with just that as the solution. 

Do we have any clarity on where the discontent was coming from 

-----Original Message-----
From: lmap [mailto:[e-mail]] On Behalf Of Juergen Schoenwaelder
Sent: Tuesday, March 11, 2014 10:59 AM
To: Bugenhagen, Michael K
Cc: 'Brian Trammell'; '[e-mail]'; 'Matt Mathis'; 'Dan Romascanu'
Subject: Re: [lmap] One example of a "passive measurement peer"

Hi,

I do not think we need the term 'passive measurement peer' in the framework so perhaps the best thing is to not spent cycles trying to define it.

/js

On Tue, Mar 11, 2014 at 01:36:23PM +0000, Bugenhagen, Michael K wrote:
> Just my 2 cents - but a word on future proofing... 
> 
> The word passive is contextual...
> 
> Which directly infers that we should "contextualize" it buy saying "x" passive.
> As the technology disturbers add new dimensions that also contain passive/active stuff this is going to get more confused....
> 
> We're seriously having a problem with this at NFV / Cloud technology SDO's where we've virtualized NIC's (Vnic's) and Vswitches both of which can be suspended (stopped)..
> 
> It may prove difficult to do passive measurements on a virtual x,y,z component.
> Given Broadband Modems have this virtualized stuff, we should consider future proofing our terms.
> 
> Summary - contextualizing passive may be a great thing to future proof... for virtualization, those proof of concepts are in the works now.
> 
> Cheers,
> Mike
> 
> 
> 
> -----Original Message-----
> From: lmap [mailto:[e-mail]] On Behalf Of Brian Trammell
> Sent: Tuesday, March 11, 2014 8:30 AM
> To: Dan Romascanu
> Cc: Matt Mathis; [e-mail]
> Subject: Re: [lmap] One example of a "passive measurement peer"
> 
> hi Dan,
> 
> Hm. I don't really consider this passive measurement -- we discussed this a bit in the ippm registry design team, but I'm pretty sure there's a difference between (1) metrics derived from the operation of a protocol at one of the endpoints participating in it and (2) metrics derived from the passive observation of packets emitted by those protocols. I've always thought of passive measurement as the latter, and (1) as something like "endpoint measurement", since metrics in (1) can also be derived from internal state at each endpoint not synchronized over the protocol or otherwise hard to derive. 
> 
> One could say that a midpath observation point (i.e. any observation point other than an endpoint) has as many "passive peers" as there are observable senders and recipients, while an "endpoint measurement agent" would have a single "endpoint peer".
> 
> (This becomes a much more interesting distinction as soon as you 
> consider encrypting absolutely everything you can, of course. At that 
> point everything you're reduced to endpoint measurement for everything 
> other than the "envelope" and/or things you derive through behavioral 
> traffic analysis. But that's a separate discussion, I think.)
> 
> In any case, I stand by my statement that I'd like to see peers defined _only_ as much as is absolutely necessary to make sense of the resulting control and reporting protocols.
> 
> Cheers,
> 
> Brian
> 
> 
> On 10 Mar 2014, at 14:31, Romascanu, Dan (Dan) <[e-mail]> wrote:
> 
> > My other example of a 'passive management peer' is an RTP peer which receives RTCP messages about the state of the other side of the connection and the parameters of the received traffic at the other end. All these are part of RTP/RTCP, there is no interaction between the controller and the MP, so I believe it can be considered passive in LMAP terms.
> >  
> > Regards,
> >  
> > Dan
> >  
> >  
> > From: lmap [mailto:[e-mail]] On Behalf Of Matt Mathis
> > Sent: Friday, March 07, 2014 12:11 PM
> > To: [e-mail]
> > Subject: [lmap] One example of a "passive measurement peer"
> >  
> > From the Internet Archive:
> >  
> > http://web.archive.org/web/20071005130953/http://www-didc.lbl.gov/SC
> > NM/
> > 
> > Thanks,
> > --MM--
> > The best way to predict the future is to create it.  - Alan Kay
> > 
> > Privacy matters!  We know from recent events that people are using our services to speak in defiance of unjust governments.   We treat privacy and security as matters of life and death, because for some users, they are.
> > _______________________________________________
> > lmap mailing list
> > [e-mail]
> > https://www.ietf.org/mailman/listinfo/lmap
> 
> _______________________________________________
> lmap mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/lmap
> 
> _______________________________________________
> lmap mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/lmap

-- 
Juergen Schoenwaelder           Jacobs University Bremen gGmbH
Phone: [phone]         Campus Ring 1, 28759 Bremen, Germany
Fax:   [phone]         <http://www.jacobs-university.de/>

_______________________________________________
lmap mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/lmap

_______________________________________________
lmap mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/lmap