Facebook Engineering Tool Mimics Dodgy Network Connectivity

itwbennett writes: Facebook has released an open source application called Augmented Traffic Control that can simulate the connectivity of a cell phone accessing an app over a 2G, Edge, 3G, or LTE network. It can also simulate weak and erratic WiFi connections. The simulations can give engineers an estimate of how long it would take a user to download a file, for instance, given varying network connections. It can help engineers re-create problems that crop up only on very slow networks.

They should make a tool

[easyTree]

Which mimics massive and creepy privacy invasion - for profit.

Re:

[easyTree]

More openness ("Here's more detail on all the immoral stuff we do") ? Nope.

Re: They should make a tool

[thedonger]

But it's not really a choice. Too many people use Facebook that it's such an inconvenience to not use it.

That may, in fact, be the dumbest thing written on the internet. At least so far today.

Re:

[cmseagle]

No, but it's dumb to say that you don't have a choice to do so.

That's a waste of time

[WebCowboy]

Why explain anything? Anyone worth my time has my phone number and/or email. The only response required to queries about me and facebook can be "phone me" or "email me" or "text me".

If anything, using facebook is more trouble than picking up a phone or tapping out an email or sms, and NSA dragnets notwithstanding less intrusive as well. So after signing up years ago and getting poked and having sheep thrown at me for a couple months my account has been virtually dormant since.

Re:

[thedonger]

But at least it's just the introductory sentence for a reasonable point, whereas your smug asshole comment is all we got.

Except that "reasonable point" hinges upon the notion that Facebook isn't a choice.

there have been network impairment tools forever

[swschrad]

at least in Internet years. we don't need one more that lays advertisements and not-friends on top of the test data.

ipfw/dummynet

[cdrudge]

So they reinvented what was already available and much more flexible with ipfw [freebsd.org]/dummynet [freebsd.org]?

Network Link Conditioner for OSX

[JonTurner]

For OSX, Apple provides a superb tool for simulating "bad" networks -- slow, dropped packets, high latency, or all the above -- called Network Link Conditioner
more info at --> http://nshipster.com/network-l... [nshipster.com]

Re:

[gnasher719]

Now if they could simulate the behaviour of a BT Infinity router... which will from time to time disconnect from the internet, and then for any http or https request, you get a "helpful" website that tells you what buttons to press on your router.

It's not a big problem if you are for example expecting an xml or JSON response, because what you get is quite obviously not xml or JSON, but for requests that actually expect normal html, this is really awful. It means that whatever request you send, you must a

Re:

[Willuz]

This was also my first thought. However, a simpler tool is beneficial for phone app developers since many of them don't understand the technical differences between 3g/4g/satellite internet/cable/oversubscribed cable/etc. Each of these situations provides unique combinations of bandwidth, latency, and packet loss that require manual tuning from the command line in dummy net to simulate. The great majority of people, even many developers, just think in terms of "speed" and not the basic networking conditions

Re:ipfw/dummynet

[bill_mcgonigle]

So they reinvented what was already available and much more flexible with ipfw [freebsd.org]/dummynet [freebsd.org]?

So they made a linux version of dummynet and released it as Open Source with deployment and configuration tools?

Assholes.

Re:

[Anonymous Coward]

While not novel, I literally was going to roll my own implementation of this sort of tool for the project I'm on, which doesn't have access to ipfw/dummynet so, yeah, I'd say this was helpful. But hey, pitchforks or whatever.

Re:

[pacman on prozac]

It does yes, there is a live CD running it in knoppix on an ethernet bridge: WANbridge [google.com].

Used it a few times, stick it on a dual-NIC PC in the middle of a link and it's great for simulating WAN links with minimal effort required.

Re:

[swb]

This was my thinking.

It's been ages since I've used ipfw as a WAN simulator, but my memory of it is normally around a fairly static kind of configuration of latencies and bandwidth.

Simulating a cellular link that might hop between LTE and 1x kinds of data might be tough to do without some kind of engine which dynamically reprograms dummynets for vastly different bandwidth/latency scenarios to better simulate a node moving between 1x and LTE speeds. When I built a WAN simulator, I did to actually simulate k

Simulation?

[danbert8]

Just try it out using Comcast/Time Warner/AT&T/Verizon internet access! Pick your provider... They all suck.

Re:

[richy freeway]

You're not missing out on anything.

Re:

[mattventura]

There's even an aptly-named tool [github.com] for it!

Just saying...

[Anonymous Coward]

Where I come from, we don't need to SIMULATE such connections................

Re:

[MachineShedFred]

And that means what, exactly?

Salary has no direct link with competence or usefulness.

Pages that get stuck

[jones_supa]

Web pages have always been a bit unreliable technology. Who doesn't occasionally meet a page that is almost loaded, but hangs there waiting for one element to be downloaded? At I meet a few times a week a page that gets "stuck". Then you refresh the page and it's fine. Why does this problem still exist? Can't the browser at least quickly try reloading that element?

Imagine if desktop GUI apps were like that. That some GUI element would just randomly not show up. That would be unacceptable.

Facebook should have Tor built-in for China users

[lazyBob]

Big profits for circumventing shitty Great Firewall.

They are using ...

[CaptainDork]

... Comcast as their beta carrier.

Wait, there's already a tool for this

[Anonymous Coward]

There's already a tool for this, with a funny name: https://github.com/tylertreat/... [github.com]

the tool has multiple bufferbloat-related issues

[mtaht]

I just gave up and filed a bug: https://github.com/facebook/au... [github.com]

So like netem?

[drinkypoo]

netem [linuxfoundation.org] provides Network Emulation functionality for testing protocols by emulating the properties of wide area networks. The current version emulates variable delay, loss, duplication and re-ordering.

Nice work, Facebook. NIH assholes.

Re:

[drinkypoo]

Way to show your ignorance, fuckwad.

OK, so I'm ignorant, and the summary is shit. And don't ask me to RTFA, because if it's about Facebook, I'm not going to.

GenyMotion, Routers

[jtara]

GenyMotion (Android simulator) has this built-in, as does the iOS simulator. This could be done on an open-source router, as well, instead of using a computer behind the router. And, ironically - doesn't Facebook have some powerful Cisco routers that probably have this sort of thing built-in (or optional?)

Already Done

[Cola Junkee]

http://wanem.sourceforge.net/ [sourceforge.net]

I hope they give this to google...

[nomel]

The play store is unusable with an erratic connection. Goes to blank screens if you're trying to open a link to an app from an external source, jumps back to unrelated screens if a problem happens when you're trying to access an apps page, has non-fucking-modal popups that you have to access in your notifications before you can continue using the app if there's a problem during download...ffs.

The *only* app that i've seen handle crappy connections is iheart radio. It'll actually sit there and retry as the b

Re:

[nomel]

Oh, and google maps *is the best* with crappy internet. It'll just silently continue on, using some inertia based guidance that you can see plowing through stoplights, with no indication whatsoever, including it *not giving directions* or extremely delayed directions. No warning or tone or something, just silence. Maybe they fixed this in the last 6 months or so...I was tired of u-turns so switched to Waze.