Hi,

Is there anyway to secure a Xilinx NGC file from being reverse engineered ?
Xilinx has a ngc2edif  utility to convert the binary ngc file into a 
readable netlist.
Still work of course but makes it a lot easier to copy a design.

So anyway to secure a NGC file ?

Thanks.

Jim 

Jim,

Copying the bitstream is trivial, so why bother with the NGC?

What is the vulnerability you are analyzing?

Who is your threat? (a major government, or an individual hacker...)

Not knowing what you are trying to protect (and why), we can't provide
you with an answer.

Austin

http://syndicated.synplicity.com/Q306/aldec.html

ngc2edif will not extract encrypted cores.

Is it perfectly secure?  I'll leave that answer as an exercise for the
reader....

'Greg

On May 9, 12:26 am, "fpganut" <r...@myhouse.com> wrote:
> Hi,
>
> Is there anyway to secure a Xilinx NGC file from being reverse engineered ?
> Xilinx has a ngc2edif  utility to convert the binary ngc file into a
> readable netlist.
> Still work of course but makes it a lot easier to copy a design.
>
> So anyway to secure a NGC file ?
>
> Thanks.
>
> Jim

"austin" <austin@xilinx.com> wrote in message 
news:g022gp$p032@cnn.xsj.xilinx.com...
> Jim,
>
> Copying the bitstream is trivial, so why bother with the NGC?

  Copying is fine. I don't care about copies.

>
> What is the vulnerability you are analyzing?

   I have some circuits implementing some unique algorithms. I  don't want 
someone to figure out
  the algorithm by extracting the circuit and figuring out how it operates.

>
> Who is your threat? (a major government, or an individual hacker...)

   A motivated competitor.

   Thanks.

 Jim

>
> Not knowing what you are trying to protect (and why), we can't provide
> you with an answer.
>
> Austin 

On Fri, 9 May 2008 13:09:40 -0700 (PDT), SoyAnarchisto
<greg.daughtry@gmail.com> wrote:

>http://syndicated.synplicity.com/Q306/aldec.html
>
>ngc2edif will not extract encrypted cores.
>
>Is it perfectly secure?  I'll leave that answer as an exercise for the
>reader....

Is there an encyrption method XST understands?

On May 10, 10:19 am, Muzaffer Kal <k...@dspia.com> wrote:
> On Fri, 9 May 2008 13:09:40 -0700 (PDT), SoyAnarchisto
>
> <greg.daugh...@gmail.com> wrote:
> >http://syndicated.synplicity.com/Q306/aldec.html
>
> >ngc2edif will not extract encrypted cores.
>
> >Is it perfectly secure?  I'll leave that answer as an exercise for the
> >reader....
>
> Is there an encyrption method XST understands?


XST has encryption built in.  To get access to the encryption tools,
you need to talk to your FAE about becoming part of the Xilinx
Alliance Program.

http://www.xilinx.com/products/alliance/overview.htm

Regards,

John McCaskill
www.FasterTechnology.com

Jim,

A motivated competitor has the legal right to reverse engineer your
product -- you do need to protect yourself.

A patent would be nice, but often a patent is too expensive, and that
may not deter a competitor.

If using Virtex II, IIP, 4, or 5, I suggest the lithium coin cell
battery from Ray-O-Vac, and the use of encryption.  That is a minium of
15 years of battery life (to 2.0 volts), and since the key is maintained
down to less than 1 volt, Ray-O-Vac is unable to extrapolate the
lifetime beyond their 2 volt end of life (they said "as close to forever
as you can get").  As long as their is reactive lithium in the cell, it
is able to push out electrons.

Set top boxes use the lithium battery solution for example (they know
how to protect themselves).

If using Spartan, there is no comparable encryption solution.

Some customers use the 3AN (with internal prom) as they feel that has
sufficient barriers to reverse engineer (no signal of the configuration
appears externally).

Identifying a small part of the design, and reverse engineering just
that part is tough, but not unheard of.

The other choice is to hide, or obscure the design as best you can:  I
know that the crypto community chants "there is no security in
obscurity" but obscurity is used by most anyways (causes the thief to
look for something easier).

Obscurity is the solution favored by natural evolution (so it does work
well enough ...).

Austin