Reverse Engineering the Nike+ FuelBand's Communications Protocol

An anonymous reader writes: Security researcher Simone Margaritelli has reverse engineered the Bluetooth low-energy communications protocol for his Nike+ FuelBand SE, a wrist-worn activity tracker. He learned some disturbing facts: "The authentication system is vulnerable, anyone could connect to your device. The protocol supports direct reading and writing of the device memory, up to 65K of contents. The protocol supports commands that are not supposed to be implemented in a production release (bootloader mode, device self test, etc)." His post explains in detail how he managed this, and how Nike put effort into creating an authentication system, but then completely undermined it by using a hard-coded token. Margaritelli even provides a command list for the device, which can do things like grab an event log, upload a bitmap for the screen, and even reset it.

Comments

OMG the Horror!

[Anonymous Coward]

Now we know how many Calories he burned. we are doomed...

Re:OMG the Horror!

[AchilleTalon]

Better, you can let him think he hasn't burn enough calories and make him running forever.

Re:OMG the Horror!

[macs4all]

Now we know how many Calories he burned. we are doomed...

Exactly!

Does EVERYTHING have to have the utmost in security?

What I didn't see was any mention of "personal information" or "remote code injection/execution" that MIGHT call for strong encryption.

So the question for Nike becomes "How much will having a bigger, faster microcontroller to encode/decode secure communications cost us (and therefore the consumer)?"

And is it necessary?

Undermined security w/ hardcoded token?

[Anonymous Coward]

Developer: That's insecure.
Phil Knight: Just do it.

So, what's the practical concern of this?

[TWX]

Are people going to reverse-engineer these things so that when they're worn into secure facilities, they inject-attack systems in the secure facility?

Are they going to act as a vector to attack other Bluetooth Low Energy capable security systems?

I simply want to know what kind of maliciousness can be achieved through exploiting bugs in a very, very special-purpose device.

Re:So, what's the practical concern of this?

[CastrTroy]

It's interesting that you bring that up. Many secure facilities won't allow people to bring in cell phones or other devices. But it's actually quite hard to distinguish some regular wrist watches from one with cameras or communications devices in them. I think if you really want to have a "secure" facility, then you pretty much have to limit people to bringing in no electronic devices whatsoever.

Re:So, what's the practical concern of this?

[oodaloop]

I work in a secure facility, and activity tracking wristbands (among many other things) are forbidden.

Re:So, what's the practical concern of this?

[DarkOx]

That was my first thought too. There is an obvious privacy, implication. Maybe in some really contorted situation you could induce someone to do something dangerous like convince a diabetic they have done a whole ton of walking this morning and therefore should eat more sugar than normal and similar attacks.

I think the big issue is the potential to use this as a vector to introduce malware to the phone or PC the owner interfaces the device with. Not sure how practical that is.

Re:So, what's the practical concern of this?

[gmack]

I think the big issue is the potential to use this as a vector to introduce malware to the phone or PC the owner interfaces the device with. Not sure how practical that is.

The big issue will be that people will use this to display rude things on random people's armbands.

Re:So, what's the practical concern of this?

[Anonymous Coward]

I work in a secure facility, and activity tracking wristbands (among many other things) are forbidden.

What about posting to Slashdot on the clock?

Re:So, what's the practical concern of this?

[Frosty Piss]

The big issue will be that people will use this to display rude things on random people's armbands.

So you are saying the Anonymous script kiddies will use this to advance their devilish plan to "out" pedos and bring down ISIS?

Re:So, what's the practical concern of this?

[CaptainDork]

The real issue is the inherent philosophy that the Internet of Things is such a trivial group of devices that security is not important.

What's the threshold, and who establishes that?

It's obviously an important discussion, given the attention it has generated.

screw fitness bands.

[nimbius]

from a social standpoint these devices are near and helpful. From a FOSS standpoint these devices are intrusive and treat their users like cattle. Check out Fitbit for example, the largest provider of digital harvesting/tracking hardware. the privacy policy insists they sell de-identified data (because metadata is a dirty word these days) to third parties. So if you're wondering why health insurance companies are pushing biometric competitions at the workplace using subsidized devices its because your health is not their primary concern. Determining an accurate insurance rate for a component of workers that are at heightened risk for diabetes, heart attack, and alzheimers is what they care about. Your corporation in turn cares about your health, and might reward you with water bottles or gift cards to sporting goods stores that, in turn, might turn into a newer fitbit/fuel.

the protocol used to affect data and function of the device is trivial, Galileo and libfitbit hacked this a few years ago. The real problem is your biometric data which is transferred across the device in an AES/md5 header encrypted blob. This violates countless freedoms of the application, starting with 0. The key to decrypt this data doesnt exist for you, and hence you're tethered to a website and a product that if it were ever usurped by say, fitness applications on your phone, would go bankrupt, shutter its doors, and leave you with a nice chunk of plastic that showed numbers and belched motivational platitudes. the real work in these devices should be decrypting the collected data without the use of the companies respective servers and web resources.

Re:screw fitness bands.

[tlhIngan]

the privacy policy insists they sell de-identified data (because metadata is a dirty word these days) to third parties

Metadata is NOT de-identified data. Metadata is data about data, while de-identified data is anonymized data.

Metadata would be for example how often and when you upload your results to their website, but nothing on what you ran or for how long and all that (that's data). The data itself would be your track, pace, location and all that information, tied to you.

De-identifying the data would mean advertisers get access to your track, pacing and other stuff, but with no name attached, and maybe even missing a few reporting points so your address isn't obvious by looking at the endpoints.

It's not that metadata is a bad term - it's reasonably accurate because it's the difference between say, a pen recorder and a wiretap recorder (ohe records details about the call, the other records the call itself). Or recording IP headers over recording packet contents.

You deal in metadata a lot - a file name is metadata - it's not a part of the file's contents (the data), just like the date and other details. You can get access to file metadata quite easily even if you can't read the file itself (and it's not possible to read the file without being able to access the metadata).

Re:screw fitness bands.

[sjames]

Keep in mind that the better insurance gets at assessing risk, the less value it has.At some point they get the risk pinned down with sufficient accuracy that you come out better by putting the premiums into a savings account until needed.

Of course the root problem is mistaking insurance for a solution to outrageously overpriced healthcare.

Is anybody surprised?

[gstoddart]

In what way should anybody be surprised that a wearable, wireless device has implemented security in a completely incompetent way?

These are products which are intended to be cool, shiny, and pretty ... but secure? Not even a little.

I continue to be unsurprised by this crap, and I continue fairly firm in my indifference to owning any of this stuff ... and the same goes the for "Interweb of Stuff"; I assume that out of the gate it's going to be insecure and stupid.

Unless companies have actual legal liability for shit security, you'll continue to see shit security.

So just don't buy it if you value security or privacy -- because they're all pretty much designed to upload your information to analytics companies anyway.

Data mule-ing and brick-ing?

[xxxJonBoyxxx]

As I understand the analysis, this exploit could be used to turn Fuelbands into data mules. It could also let someone temporarily brick all the Fuelbands within range (could be fun at the start of a marathon or at the gym).

>> Cmd_Bootloader: Set the device to bootloader mode ( basically it locks down the device, the official app won't work either ... only resetting it with the usb cable will unlock it ).
>> Cmd_SampleStore: Use the device memory to store a custom object (!!!)

Cool.

[the_skywise]

I've got one of these and I honestly don't care if the band isn't secure. Sure they can get my motion data and I'm sure some nefarious insurance company could install bluetooth readers worldwide to pilfer my data, confirm that I'm not moving enough and raise my rates but... heck any good scale and heart rate meter would tell you that and probably be a lot more useful.

But now that it's been cracked -
I'd like to see somebody use this to make an android version of the software.

I'd like to see somebody interface with the iPhone better so when I get a call it could display the name of the caller or an appointment reminder on my band (ooh, wotta concept!)

And maybe, just maybe somebody would write a download tool that'd COLLECT THE DATA BETTER THAN NIKE DOES!!!! Because right now I average about 1 bad read a month. (Killer if you're trying to go for the consecutive days award or, y'know, actually, accurately *monitor* your activity levels)

Fuck it, who fucking cares about that token anyway

[burtosis]

Let's just use 0xff 0xff 0xff 0xff 0xff 0xff .... !

I laughed pretty damn hard. Best laugh I've had today.

Didn't they learn the lesson of the PC?

[zoffdino]

This whole IoT concept is treating security as a joke. In the first of wave computing, the mini-computers (particularly Windows) treated security as an after-thought. That created the virus-laden era of the 1990s and early 2000s. The second wave, the "new" smart phone, learned the lessons, and use sandboxes, walled garden, permissions, encryption, tokenization, etc. pervasively. It's not fool-proof but at least the door is locked. Now we are approaching the third wave, the Internet of Things, and manufacturers think these devices are so personal that no security is needed. What do they say about people who don't learn any history?

Re:Didn't they learn the lesson of the PC?

[itzly]

Those who learned from history are doomed to watch others repeat it.

Re:Targeting the DIY community?

[burtosis]

Yes like bricking everyone's band as they run past at a marathon or popular path. Perhaps adding a mandatory PWNT message on the band.

Im thinking this is an unintended concequence of a smug and vindictive type benevolent overseer decision.

There are better ways to get DIY support IMO.