Guarding Against Physical Attacks: The Xbox One Story [video]
platformsecuritysummit.com
platformsecuritysummit.com
> In this talk, we will discuss how we achieved this for the Xbox One.
While Xbox' technical security measures are very interesting, I dare to say that most effective measure was allowing to install homebrew apps by end users[0]. The only use for cracking Xbox One would be to enable piracy and exploit creators are almost universally against that.
Also, PS4 has been jailbroken up to firmware 5.07 (patched in March 2018), and it is known that exploits for newer firmwares exist but are not (yet?) publicly available.
[0] https://github.com/wiired24/ProjectPegasus/blob/master/binaries/Installing APPX Packages.pdfThat's not strictly true, cracking could also allow alternative os loading.
Settlement was supposedly $65 but it looks like affected people got a whole $10 for their trouble: https://www.reddit.com/r/PS3/comments/9izcpm/has_anyone_rece...
There were some clusters made out of PS3s, presumably the operators kept these from updating, but they wouldn't be able to replace failed units with new ones anymore: https://arstechnica.com/gaming/2010/05/how-removing-ps3-linu...
So when the DRM servers are shut down, these devices will become just useless paperweights (if nobody manages to crack them).
"User goes to Walmart, buys console and game. Drives to cabin with no internet. Sets up console and plays game".
Clearly the original design goal didn't require an initial phone home. That may have changed.
Some points from the talk:
- custom 28nm AMD CPU
- off-die devices distrusted
- encrypted memory
- physical threat model: 10-game budget
- pre-PSP "Secure Processor"
- custom SP firmware by MS
- minimal hypervisor
- subset of Windows
- secure boot & remote attestation
- Azure Sphere Pluton = Xbox derivativeDecidedly unsettling.
Cheating does happen on PC. I've seen it maybe twice in the last 15 years. Wouldn't call that rampant.
Losing at a video game does not mean your opponent is a cheater. This might sound dismissive but the number of times I was called a cheater when barely managing to win has me convinced that people just don't like losing.
Some games have lots of cheaters. Some don't have any.
Doesn't mean they cheated, new game plus is a thing, and the summoning levels (and now weapon matchmaking) are easily gamed to allow you to invade people less well-equipped than you are.
I've seen videos of actual cheaters, who hit you with a homing soulmass and it breaks all your equipped weapons and armor. Or they never run out of stamina. Or whatever. But I've never seen one in person, and I've wasted far too much of my life on that game.
I don't think most people know how to evaluate who is cheating and who is not, and it's taken as fact that cheaters are rampant so claims of cheating are believed by default.
I'm definitely not saying that cheats don't exist, or that people don't use them. I'm saying the community at large appears to have jumped on the problem as a way to save face. "Oh he beat me, but it's only because he's wallhacking!"
However, I am calling the god-like players who own me hackers, when I watch their kill-cam jump rapidly to each persons head with absolutely precision shot for shot. It gets into an area of response time where the human feedback loop takes longer than the response-and-inputs being delivered.
Getting accused of being a cheater is actually more common. One time some guy spent 10 minutes frantically trying to convince everyone that I was aimbotting...on Pyro. I wasn't even doing that well.
That's exactly what I'm talking about though. The fact that cheats exist is used to save face when people are losing (or even not winning by a large enough margin) in a video game.
It's okay to lose in a video game. I do it all the time. There's no shame, you don't have to make excuses, it's a game.
Valve Anti-Cheat (VAC) is probably one of the best anti-cheat technologies out there given the size of the user-base and the time it has been deployed.
I think I can safely say that every competitive online game with a sizeable player base has a cheating problem. How bad that problem is varies - my impression is that Overwatch is pretty good (I've never noticed cheaters, only heard reports), whereas Rainbow 6 Siege is much worse - I have noticed cheaters several times and have had MMR awarded due to matches lost to people later banned for cheating.
I'm told that Dark Souls is the perfect storm of peer-to-peer and trusting clients, but I've only seen videos of cheaters, never in-person. And I've created many characters purely for invasions and summoning.
Amusingly enough, one of the two times I referenced above was actually in Minecraft of all places. We started noticing one player getting resources far faster than everyone else, faster than was reasonable for even a very good/lucky player. I mean multiple nether stars in the span of half an hour, that kind of thing. I think they were filling up books and that somehow let them duplicate items.
re: the business model - Put a price on flipping that mode switch. That solves the business model problem too.
He even brings this up in the talk, mentioning Dreamcast.
How it died early on, because of piracy - developers didn't want to produce games for dreamcast - so no one bought console because no one made games for it.
Its a vicious cycle that cannot be stopped otherwise.
The Dreamcast was killed because of the PS2, according to Wikipedia (but I'm not knowledgeable about the issue).
https://en.wikipedia.org/wiki/CD_Projekt
https://en.wikipedia.org/wiki/GOG.com
https://www.mcvuk.com/no-evidence-that-piracy-affects-video-...
Sega had a tricky financial situation and consumer trust issues because of their last 1.5 platforms (Saturn and 32X). EA did not support the platform because Sega wouldn't give them an exclusive license for sports games. The Playstation 2 had an unbelievably huge hype train before release. The cord for the controller came out of the wrong side. The economy was getting weird.
If piracy was a big issue, I think Sega would have sold more hardware, and the software sales would have been low, but Wikipedia says the hardware sales were low, and the software sold 8-1 with consoles.
Sega with the previous generation consoles burn consumer and developers alike. People were reluctant to buy the Dreamcast, PS2 hyper machine was also making exaggerated claims about the PS2 when magazines interviewed random people at EB Games and baggage the all said the would wait for the PS2.
Someone did an in depth analysis of the Dreamcast sales as related to piracy. If piracy was such an issue then sales of the console itself should have increased because people were buying the console and not game but this was not the case. Console sales never increased.
People who claim the Dreamcast die due to piracy simply do not know SEGA's history
youtube-dl -f bestvideo[ext=mp4]+bestaudio[ext=m4a] -o "XBox One Story by Tony Chen.mp4" https://www.youtube.com/watch?v=U7VwtOrwceo
start "" "XBox One Story by Tony Chen.mp4"Explanation/example: https://thejh.net/misc/website-terminal-copy-paste
Previous HN discussions: https://news.ycombinator.com/item?id=10554679 and https://news.ycombinator.com/item?id=5508225
No need to tell it best quality; it does that automatically.
What they claim sounds impressive to CEOs and such, but everybody in the industry knows that anybody with straight hands can probe a chip, and somebody with ion beam equipment can do pretty much anything he wants with it.
For example, chips from credit cards, said to be "physically impossible" to tamper with, are known to be copied by somebody in Eastern Europe.
If somebody becomes a victim of credit card cloning, proving the bank that it was not you who withdrew money on other side of the globe becomes nearly impossible without going to a court.
In the video he specifically talks about an economic threshold - it must cost more than 10 games ($600) to mod a console. Above that price it's not worth it, so direct chip beam attacks aren't relevant to their threat model. That is the direct 'resource requirement' as a result of their business goal - prevent piracy.
Attacks and exploits can have scaling properties much like other software.
We can do something expensive (say $200k attack cost) to break the platform and sell 10k $100 mod chips.
The defender's threat model can't discount expensive attacks; the constraint is that an expensive one-off attack must not enable a cheap bypass. This is hard!
Those types would only need to get a single key to decrypt all downloads.