PlayStation 4 (codename Orbis) technical hardware investigation (news and rumours)

[novcze]

The Fujitsu bridge chip is marketed for its "high-speed encryption hardware support", which is a likely reason for Sony using it.

I'm looking at hi-res photo of PS4 motherboard from ifixit.com teardown and datasheet of usb/sata bridge, and

- IC used in PS4 is MB86C311B which does not have an encryption function (AES)
- surprisingly enough USB 2.0 line is also connected to sub-system chip (not only USB 3.0 one)

 

[DrJay24]

I was watching the archive of Giant Bomb's PS4 stream and it seemed that the UI would get really laggy when you were trying to install, play and download stuff all at once. It would be interesting to see if an SSD would improve the experience in those situations where the contention for the HDD is getting bad. That's harder to test, of course.

I've done plenty of that, has no effect on the UI. I've had my queue DLing five thing, game in the background, chatting and UI is perfect.

 

[pch]

SAA-001(PS4 main board):

Bluetooth/Wifi controller:
Marvell Wireless Avastar 88W8797
http://www.marvell.com/wireless/avastar/88W8797/

Gigabit Ethernet Controller:
Marvell 88EC060-NN82

HDMI transmitter:
Panasonic MN86471A(HDCP enforced)
http://www.semicon.panasonic.co.jp/en/products/logic/hdmi/#general

USB 3.0 MTT Hub Controller
Genesys Logic GL3520
http://www.arrivalelectronics.co.uk/uploads/GL3520-21+Datasheet_101.pdf

Network Co-Processor:
SCEI CXD90025G (only thing i could find)
http://www.chipworks.com/en/technical-competitive-analysis/resources/blog/inside-the-sony-ps4/

USB 3.0 / SATA bridge LSI(have Cortex M3 MCU-attached to an SPI Flash chip)
MB86C311B
http://pdf.dzsc.com/99999/20101029175627721.pdf

Flash:
Macronix MX25L25635FMI-10G(CMOS MXSMIO-SERIAL-256M-BIT)
http://datasheet.octopart.com/MX25L25635FMI-10G-Macronix-datasheet-17291205.pdf

Macronix MX25L1006E/25X10CLYA1 Winbond (2 on a board-CMOS SERIAL FLASH 1M-BIT)

http://postar.sino-star.com/public/uploads/20120412174638_676.pdf

or

http://www.winbond.com/NR/rdonlyres/96FEFC44-0598-4395-AA0C-B1CA855121BB/0/W25X10CL.pdf

RAM:
Samsung K4G41325FC-HC03 (GDDR5 RAM)
http://www.samsung.com/global/busin...t/graphic-dram/detail?productId=7824&iaId=759

Samsung K4B2G1646E-BCK0 (2Gb SDRAM DDR3-1600- RAM for SCEI CXD90025G network co-processor)
http://www.samsung.com/global/business/semiconductor/file/product/ds_k4b2g1646e_rev121-0.pdf

APU:
SCEI CXD90026G
http://www.chipworks.com/en/technic...a-look-at-sonys-playstation-4-core-processor/
a fan on it is
Delta KSB0912HE-CK2M (12V/1.40A)

Regulators:
International Rectifier 35858 N326P IC2X


Thermistors and fuses:
Anything with TH or F (meh!)

Clock synthesizer:
IDT 6V41265NLG
https://www.idt.com/document/ovr/timing-solutions-overview

No idea what this is:
SCEI 1327KM44S
39A207 1328 E1 3FU

Bluray board:
BDP-010 chips on are:
Renesas SCEI R8J32841FP1
BD7763EFV 325 T62
STM8ED 9H A07 VG MYS 331Z
Microchip Technology 312 3536A

PSU:
ADP-240AR(100-240V~ 2.5A 50/60Hz to 12V 17A/4.7V 18.6A)

Hard-drive:
Hitachi Travelstar Z5K500
http://www.hgst.com/hard-drives/mobile-drives/7mm-thin-and-light-drives/travelstar-z5k500

 

[SlimJim]

Ah, that would make sense. If MS do that I expect they'll do it the same way they do it with hardware storage encryption on PC, which doesn't require a bridge chip.

Regards,
SB

Given their recent console efforts: they would use the USB bridge as well, but simply rebrand it: probably go for CryptVault. (see: SHAPE, DataMove Engine, etcetera).

@phc: wow, that is really impressive. If PS4 was not coming out in my region next week, I would probably buy all these components and try to built a PS4 myself

 

[liolio]

I don't think that article from CHipworks was posted, it is a more detailed take on their early finds:
http://www.chipworks.com/en/technic...a-look-at-sonys-playstation-4-core-processor/

THe SoC is indeed pretty big at 348mm^2

EDIt, whereas I did not see the article here, Alstrong has been both fast and stealthy as he did update the "consoles die sizes" thread.

 

[Silent_Buddha]

Given their recent console efforts: they would use the USB bridge as well, but simply rebrand it: probably go for CryptVault. (see: SHAPE, DataMove Engine, etcetera).

@phc: wow, that is really impressive. If PS4 was not coming out in my region next week, I would probably buy all these components and try to built a PS4 myself

I really don't see why they would need to. Hardware encryption on Windows machines doesn't require any bridge chips. Support is baked into the chipset/APU and HDD and comes at pretty much no speed penalty when done in that way. See reviews for the Crucial M500, for example.

The only problem then is that user replaceable HDDs would become problematic, but since you won't officially be able to change the internal HDD, that isn't a problem for the machine.

Regards,
SB

 

[novcze]

Hardware encryption on Windows machines doesn't require any bridge chips.

Part used in PS4 is is simple USB/SATA bridge and have nothing to do with encryption.

 

[SlimJim]

Part used in PS4 is is simple USB/SATA bridge and have nothing to do with encryption.

Agreed. Sony probably have their own encryption methods anyway.

 

[Deleted member 11852]

Agreed. Sony probably have their own encryption methods anyway.

Doubtful. Security yes, encryption no. Encryption is very hard to get right unless you really know what you're doing and few commercial organisations have this expertise. Almost nobody rolls their own encryption these days and when they do, it's generally bad. That's why we have standards like AES, and before that, DES.

AES is solid as long as key management and access is equally solid.

 

[bkilian]

Agreed. Sony probably have their own encryption methods anyway.

Not sure I'd trust any Sony specific encryption method. I mean they're the reason we had the term "Sony Random" (A random number you pick right at the beginning and then never change) on the teams I worked on.

 

[patsu]

Yes... regarding that. In retrospect, I was wondering if they were visited by NSA folks.

 

[BoardBonobo]

Yes... regarding that. In retrospect, I was wondering if they were visited by NSA folks.

The NSA helped create the standards we are so reliant on anyway, DES, AES, SHA, et al. They, and other security services, have access to everything anyway. No need to encourage anyone.

And as DSoup said, creating an encryption that is mathematically sound is a very hard\complicated process. Increasing the key size in order to exponentially expand the combinations is the easiest and, in many ways, the most practical method to use.

 

[SlimJim]

Not sure I'd trust any Sony specific encryption method. I mean they're the reason we had the term "Sony Random" (A random number you pick right at the beginning and then never change) on the teams I worked on.

"DRM" is probably more in line with how I meant encryption: as a means to secure the platform.

How would your teams call your own drm/encryption methods btw? I think they would quickly change the subject because the 360 didn't even last a month (going by the EU release date). But correct me if I am wrong.

On the subject of 'encryption', I think the PS4 will be the most secure platform ever. Going by the PS3 and what they learned from it: also the Vita showed that no new vectors have been discovered.

A secure platform will create a strong bond with developers. And it's probably essential to the survival of some consoles: Imagine piracy on the rival platform is super easy. Wouldn't that discourage devs from targeting that platform, especially if PS4 is really easy to develop for along with those minimal risks?

 

[Grall]

The NSA helped create the standards we are so reliant on anyway, DES, AES, SHA, et al.

In at least one case, the NSA deliberately encouraged/inserted vulnerabilities into the encryption scheme standards, to make it easier for them to crack encrypted data. Very dangerous and shortsighted behavior when you actually think about it, but alas... [Insert joke about military intelligence or lack thereof here.]

 

[pMax]

the reason we had the term "Sony Random"

Indeed, was very silly.
Yet, one implementation error may happen -and if it were not for that monstrous error, PS3 would still be not bricked, as the chain of trust would be intact.

That was not the same destiny of xbox.

I am curious to what will happen this generation for ps4 and xb1, but the backpedaling on June's DRM left me speechless (I swear I would have loved to be a security guy in MS those days...).

 

[patsu]

It's pretty amazing that they could recovery PS3 after it's been hacked massively. ^_^

 

[bkilian]

"DRM" is probably more in line with how I meant encryption: as a means to secure the platform.

How would your teams call your own drm/encryption methods btw? I think they would quickly change the subject because the 360 didn't even last a month (going by the EU release date). But correct me if I am wrong.

On the subject of 'encryption', I think the PS4 will be the most secure platform ever. Going by the PS3 and what they learned from it: also the Vita showed that no new vectors have been discovered.

A secure platform will create a strong bond with developers. And it's probably essential to the survival of some consoles: Imagine piracy on the rival platform is super easy. Wouldn't that discourage devs from targeting that platform, especially if PS4 is really easy to develop for along with those minimal risks?

The 360 was only ever able to run unsigned code once, when an error crept into the hypervisor in one update, which was very quickly fixed (It was a buffer overrun, if I recall correctly). The main hack on the 360 was MS's misplaced trust in the DVD drive bios, which never resulted in unauthorized game code running. Not being able to trust the hardware is one reason why MS wanted to move to online activation only this generation. It doesn't matter how good your encryption, if someone has access to the hardware, they can find a way around it.

What Sony did was implement an encryption standard, but get it wrong. There is only one rule for encryption: Never, never, never, implement your own. If even one outside encryption specialist had looked at Sony's implementation, the PS3 hack story could have been very different.

 

[TheWretched]

And as you can see with PC, online activation doesn't work either... it's a fools errand, if you ask me. Everything gets cracked. It's just a matter of time... but PSP openness was pretty massive^^

 

[bkilian]

Oh, absolutely. No MS engineer would ever call their system "hack-proof". They just call it "hack-resistant". Given enough time and energy, any system can be cracked. What the console makers want to do is make the time and energy greater the the cost of just buying the damn game.

So the things they want to not have is a hack that can be easily applied in software. The 360 hacks, which required things like resoldering chips on the ODD, basically guaranteed that the barrier to entry was high enough that it didn't become an epidemic. The PS3 hack, which required nothing more than plugging in a USB dongle was a bit more worrying for them. They did a stellar job of mitigation in the end though.

Compared to the previous generation, the 360 and PS3 were amazingly engineered devices, security wise. If they learned as much from this go around as they learned from the previous ones, game pirates will have a really difficult time ahead.

 

[Shifty Geezer]

If not for Sony's stupidity in the implementation, PS3 would have been remarkably robust. That same approach this gen, executed properly and with further refinements, should be enough to stave off mass piracy. The major issue is when your platform is hard to hack, you'll attract champions like Hotz trying to prove their brilliance, in it for the challenge. You need to keep them from success long enough that no piracy industry can evolve, as the pirates themselves have know interest in the hacking and won't invest in it.