( ESNUG 268 Item 3 ) -------------------------------------------- [10/9/97]
Subject: ( ESNUG 265 #2 266 #4 267 #4) Synthesizable IP Is *Unprotected* IP
> It is my general opinion that there is tremendous pressure to avoid
> royalties & marginally higher production prices by re-spinning the design
> with a low price fab that is not offering any IP. (Been there, done that.)
>
> It's not just design IP that is the issue here either. No matter how
> much is costs to develop, nor the advantages to the customer, the customer
> will always want a device in a 500 high performance BGA package for the
> same price of a 208 QFP.
>
> From the inside, I view the IP issue as a huge problem for the silicon
> industry. I personally don't see how semi houses can continue to offer
> this limitless cornucopia of IP with no ROI. Many of your responses have
> indicated that you have to make the production "attractive" enough to
> keep the business. The only way to do that is to bomb the price.
From: Eric Ryherd <eric@vautomation.com>
Hi John,
I'm replying to "Synthesizable IP Is *Unprotected* IP" while at 30,000 feet
on the redeye after the VSI meeting on IP protection (among other things).
As an IP vendor we ONLY ship the HDL source code. We rely on the
trust and honesty of the organization which is purchasing the core to honor
their contractual obligations. We can trust organizations since they
have quite a bit to lose given the legal system in the states. We can't
always trust individuals but that's just the way the world is...
Trying to encrypt the IP or otherwise "hide" the underlying nature of the IP
will only cause infinite support headaches. We've all had our share of bugs
in ASIC simulation libraries that are quicker to fix ourselves than wait for
the ASIC vendor to correct themselves. But if the library is ecrypted, all
we can do is scream at the vendor! I've been in this position myself
several times and certainly don't want to inflict it on any of my customers.
Two interesting points were made at the VSI meeting. First, there is
new technology which was announced which will encode a "signature" into the
unused states of a state machine. This in effect applies a "watermark"
to the design and would allow an IP provider to identify it's IP and
possibly even the path the IP followed to it's final silicon destination.
Thus, the IP provider could prove that it's code is in a particular design.
This is not unlike certain typos in error messages used to prove a certain
unnamed EDA software copyright lawsuit that has recently been in the press.
The second point was that at some point, no matter how good the encyption
of IP is, unencrypted forms of the IP are available to the world. In the
best protected case, you can always peel the lid off a chip and reverse
engineer the polygons (and there are companies making quite excellent livings
doing just this!). More often, netlists of the IP are released to ASIC
vendors or programmed into FPGAs that can easily be retargetted into other
technologies. So encryption doesn't solve the problem of protecting IP.
To sum up... Everyone has to realize the IP has value. You need to pay for
the IP seperately from the silicon costs. If your ASIC vendor has IP you
want, pay for it up front. If they have the best silicon costs, buy it. But
no one wins by trying to lock a customer to a silicon process with IP.
- Eric Ryherd
VAutomation Inc.
---- ---- ---- ---- ---- ---- ----
From: dchapman@goldmountain.com (Dave Chapman)
Dear John,
The psychological abnormality which causes corporations to have royalty
phobia should be exploited, not struggled against. However, I think that
the best bet is to distribute IP as a form of advertising.
- Dave Chapman
Goldmountain
---- ---- ---- ---- ---- ---- ----
From: Dyson.Wilkes@swi055.ericsson.se (Dyson Wilkes)
John,
All of this discussion so far has been about encryption as a mechanisim
for enforcing licensing (or thats how I read it). I am not a lawer so I
am working from an intuitive standpoint. It seems to me there are two
appraoches to licensing: copy protection and copy detection. Is we look
at other forms of IP both mechanisms are used. Is there scope for some
form of watermarking of RTL code? Idealy the watermark would "show through"
in the physical realisation (the ASIC component). The watermark needs to
be incorporated into the logical structure of the design in such a way that
removing distorys the design. There are comparable patented machanisms
for protecting digital images.
- Dyson Wilkes
Ericsson
|
|
