This is pretty cool. I believe it was Bob Metcalf at PARC who had a device in his office that was hooked up to the 3Mbs Ethernet cable and would essentially do the same thing. He built and installed it to get a feel for how much of an issue "collisions" would really be on the network. At the time, the idea that you could just back off when you detected a collision and re-transmit was considered heresy of the highest order and the "experts" all believed that once you got to a certain small number of clients the wire would be overwhelmed with collisions and the network would be unusable. His gizmo proved that wasn't the case.
There's a certain system I work with where random unauthenticated visitors on the internet end up supplying data like name/phone/email, with no validation... And of course, the business wants to somehow convert that into a list of "real people" and start correlating it to other records.
I've been trying to stop anything too terrible from happening by asking them to clarify their business requirements, e.g. what should happen when there is malicious impersonation, or the expected result should be when inconsistencies and overlaps exist.
It's not like there's no value to the data... but I'm afraid they don't really understand the problem are are hoping the magic computer can somehow *poof* garbage into fine cuisine.
“Enrichment” is what they actually want. People think it’s Google, Amazon, Facebook etc selling their data when in reality they are simply letting people target based on it.
On the other hand there are hundreds of companies nobody has ever heard of that do buy, collate, clean, and sell access to the data that apps, browser extensions, windows apps, loyalty card programs, branded credit cards, retailers, and companies that scrape LinkedIn etc will happily sell.
You provide what you have and for a price these “enrichment” services will provide what is essentially a dossier of everything that can be even remotely inferred from the thousands of datapoints they have based on your email/name/phone.
What most people think big tech is doing is actually being done in ways that are far more unsettling by companies with cutesy names and vague services that major companies sign contracts with to improve their signal to noise ratio.
> In CSS, selectors are patterns used to match, or select, the elements you want to style. Selectors are also used in JavaScript to enable selecting the DOM nodes to return as a NodeList.
From [1]. This nomenclature was also used way before CSS even existed; in SQL.
My guess is it is meant tongue-in-cheek seeing as it is a way to render into audio all network traffic. The idea being that if you are being targeted by some nefarious, infamous 3 letter agency who has gotten something nefarious running on your system, this could, in theory, be used to identify traffic you didn't intend to have happen.
Realistically speaking, it's hogwash because you'd have to profile all "loopback" traffic and whatnot, but in theory, with enough time and effort, that dastardly stream of bits could be heard, I guess. The circuit board itself though has quite a bit of punning in it.
Also... If they really want on your databus, they will get on your databus. After a certain point, cost is not an issue.
> if possible disable encryption, then you can profit from not only timing pattern (of white noise), but also listen in on the plaintext payload. the NSA loves plaintext.
Haha, incredible.
On a more serious note this is a really cool idea. Would be interesting to listen to the same origin traffic in different network conditions to hear things like TCP rate control.
Those who recall cars with fm radios and gsm phones will have heard something similar if their car had a random compartment right below the stereo.
When a call came in, before the phone would ring, the radio would emit this very clear tri di di dit dah dah. But only when the phone was very near the radio.
Yes kinda, I would say network activity rather then traffic. Audio signal is going to be in scale of 48Khz while measuring ethernet signal at scale of 100Mhz. At that rate it wouldn't even get more then 1 sample from a full size packet. So really it's polling 48Khz whether or not there was activity during that period. The gimmick is that it uses some analog components. Fully digital you could craft a meaningful audio signal that represents traffic.
some people learn the hard way that blindly copying and pasting things from the interweb directly to a terminal is not always the best of ideas. some times, they're bloody brilliant
The joke is on him - I used to get this functionality for free way back in the day when (what I presume was) RF noise generated during the processing of ethernet traffic would get picked up by my cheap ISA sound card and sent to the speakers. I never built a sequencer out of it though.
Back in college, I would listen to AM radio while I worked on my computer. The radio would pick up electrical noise from the keyboard and the mechanical mouse. I wonder if this sounds anything like it.
I read that the FCC really started cracking down on EM interference because early microcomputers were serious offenders. I'm guessing that's why there's a humongous RF shield inside my Amiga 1200.
Good for network wiring diagnostic. It would be great if it can pipe the noise to Bluetooth audio. I can pair a headphone to it, plug this into the network in another room and still can hear it while checking the line connections on the switch/router.
That audio port is blasting out a total bandwidth of 100 Mbps (4 bits at 25 MHz) versus 768 Kbps for BT audio, assuming a high quality codec (16 bits at 48 KHz), so not without loss.
It would be nice to compare all packets dumped by system and check if there are any unmatching packets running from your computer that were not sent by OS but a closed source hardware inside like Intel ME
The NSA worked with GHCQ to tap Google's fiber between data centers, which at the time, was not encrypted. You can see several presentations including "SSL added and removed here" (reference to the SSL connection being terminated at the Google front end and then transmitted unencrypted to the backend in another data center), as well as an actual BigTable packet from tcpdump that included a user identifier.
If you read The Idea Factory, it shows that AT&T leadership worked closely with NSA and other governmental agencies (on a "secret schedule" so nobody would know who the execs were meeting with) to help them access US phone data.
I'd love to know the extent of what NSA has done between its founding and today; I'm sure they've pulled off some astounding things, and bolluxed up other stuff badly.
I've worked with quite a few ISPs and exchanges. I haven't set up port mirrors for the NSA but I have setup temporary mirrors for the FBI upon request.
Schoolyard bully justifications do not a great nation make. Especially when that nation is the worlds #1 committer of human rights abuses and supporter of genocide^Wmass murder at heinous scale.
Many times over the decades I have constructed a MIDI->Audio cable so that I can record MIDI signals as audio, which as a technique is very useful for measuring and fixing MIDI->Synthesizer engine latency.
Its nice to know that if I ever get around to shipping a product with RTP-Midi (MIDI over Ethernet), I can just get a ready-made Eurorack module to do that testing! :P
reply