Using GPS timing to detect the bufferbloat problem

[justjed]

Considering the level of expertise here in hardware and small-quantity PCB fab, combined with enthusiasm for open-source, I thought perhaps some folks would be interested in a proposed project to attempt to measure network latency caused by IP bufferbloat, using GPS timing.

FYI: The bufferbloat projects, see also coverage on /.

[cyberteque]

LOL

A while ago this "expert in everything" I know was wondering if we could detect gravity waves by using ping times between servers, over the internet!

This sounds similar!

Unless you live somewhere like, Naru, geographical location has very little to do with your internet data path.

[justjed]

This is something that's more tangible than gravity waves, I think. And ESR isn't yet another wannabe engineer who just heard about Tesla waves. The bufferbloat problem is man-made thing, existing in a well-defined domain, and therefore it seems reasonable to me that it should be measurable. Whether that's possible in the way he's looking at it, I don't know. But at least initially, it sounded like something interesting.

[Udo Klein]

Why should we care and why should this be tackled with GPS? I do not get both. Latency issues can be measured without GPS. Also I do not see why latency is such a big deal anyway. If I need low latency I keep my data local.

[cyberteque]

wth do tesla waves have to do with gravity waves?

my point was -

the whole idea behind the "internet", as she was invented by ARPA, was to make a redundant, multi path network.
That way if the ruskies nuked one routing point or server, it would "automagically" use another router node.

when you send a packet of data from your modem, there is no guarantee it will go through the same data path.

another classic quote from yet another "expert on everything",
I was talking about getting slow ping times, he pipes up with,

"It's not so much ping times as a problem of latency"

[justjed]

wth do tesla waves have to do with gravity waves?

Not a thing, it was just an example for sake of contrast. Though I wouldn't be surprised if there are people out there who think they're connected.

my point was -
[snip]
[/quote]

Yeah, I know.

I wasn't trying to stir up the pot over here. Just thought that the hardware aspects of it might be something someone might be interested in. There's plenty of debate on that blog post about the merits of the approach. And I, not being a networking guru, find it somewhat interesting, but I can't judge whether it's out in left field.