[Bloat] ONTs and ITU - T G.988

[Marco Belmonte]

Dave,

Can you describe to me what under buffered would look like? Let's say 
I'm playing Rocket League on a Frontier FIOS connection... When a 
situation arises where the connection is under buffered you might see 
__<Insert DT's Big Brain words here>___.

A lot of interrupts being generated? connections timing out? lag?

I'm guessing that many of the same symptoms that accompanying buffer 
bloat also manifest themselves with buffer bleed?

Marco


On 1/12/2022 3:42 PM, Dave Taht wrote:
> Up until this point I haven't cared all that much about fiber, as most
> of my observations were that it had "reasonable" buffering and all I
> could hold in my head was the 802.11 standards. Since cablemodems are
> well on their way to being fixed, wifi is looking good, and I recently
> came across an ONT that was actually underbuffered, my holiday reading
> consisted of pounding through a ton of fiber specs like this one:
>
> https://www.itu.int/rec/dologin_pub.asp?lang=e&id=T-REC-G.988-201711-I!!PDF-E&type=items
>
> There's a lot of promising info in this spec starting from page 104.
>
> My 1st question is - are there any ONTs that actually do do pause
> frames? Or providers that configure for them?
>
> My second is - what is a cheap way to setup a lab to emulate a good,
> common, version of gpon e2e?
>
> Third - what OS do these things run? Who makes a "good" one?
>