[j-nsp] Juniper route age reset behavior

[Niall Donaghy]

Hi Jeff,

 

Niall> I suggest the default should be on-the-wire as this incurs the minimal CPU hit, as I understand it.

Jeff> Please consider the extra CPU hit to be marginal.  It's the memory hit that's the main concern for having more than one timer.

 

I should also add that as I previously stated, I find the ‘on-the-wire’ age to be the most intuitive one / ‘rule of least surprise’, etc.

I read the age as ‘age of route to destination {via this BGP speaker}’, as I presume Magnus (OP) also intuits.

 

HTH, and again, thanks for your diligence in soliciting user input on this one. :)

 

Br,

Niall

 

From: Jeff Haas [mailto:jhaas at juniper.net] 
Sent: 19 April 2018 21:00
To: Niall Donaghy <niall.donaghy at geant.org>
Cc: Magnus Bergroth <bergroth at nordu.net>; juniper-nsp <juniper-nsp at puck.nether.net>
Subject: Re: [j-nsp] Juniper route age reset behavior

 

Niall, 

 

I'll answer clarifying questions, but hope to remain mostly silent while people offer their opinions.

 

On Apr 19, 2018, at 3:46 PM, Niall Donaghy <niall.donaghy at geant.org <mailto:niall.donaghy at geant.org> > wrote:

Jeff> Thus, a knob is being considered to have both the "only on-the-wire" impact or the "whenever the route's properties change" impact.

 

Can I suggest a knob with three options – the two you mention, plus an option to maintain /both/ timers? If that’s not feasible (for reasons you already eluded), so be it.

My preference is indeed for ‘best of both’.

 

I don't believe we'll be adding a second timer at this time.

 

We have proposals in mind that may permit variable sizing of some of our core data structures.  Once we have such a thing, we may consider such optional memory-hungry features.  I know I want to do several things with them. :-)

 

 

Jeff> The question being: What's the default?

 

As you indicated, quite a few folk on older hardware will be sensitive to resource impacts.

I suggest the default should be on-the-wire as this incurs the minimal CPU hit, as I understand it.

 

Please consider the extra CPU hit to be marginal.  It's the memory hit that's the main concern for having more than one timer.

 

-- Jeff