MS Patent for DRM OS

[Ken Brown]

Could be vulnerable to hot-wire conditions - floods of interrupts & so
on. I don't think M$ would want to be selling an OS that took a security
outage every time the disk got too busy.

Anthony Naggs wrote:
> 
> In message <20011213145708.E51502@colon.colondot.net>, Matthew Byng-
> Maddick <ukcrypto@lists.colondot.net> wrote
> >
> >I also note the idea that it gets the time off a "trusted" time server,
> >what will the "trust" be based on, and how will it cope if not networked.
> 
> Not a "time server", a clock.
> 
> >Surely this doesn't imply that I have to be on a live, internet-connected
> >network, to run software?
> 
> The trusted clock would tick at a fast rate, but probably slower than
> the processor clock.  The purpose would be to detect the operation of
> debugging tools in the vein of Numega's SoftIce.  If a debugger was
> active then the code would take longer to execute.
> 
> The trusted clock would have to be part of a hardened core to the host
> platform.  Perhaps incorporated in the microprocessor.  The timers in
> Pentium class CPUs may be sufficient for this purpose.
> 
> Applying this rapidly becomes much harder in an environment such as a
> PC, where the processor is frequently interrupted by hardware requiring
> supervision.  Allowing for some time to be used by genuine interrupts
> loosens the time constraint that can be applied, possibly allowing
> debugging tools to probe the DRM code.
> 
> Regards,
>        Tony