Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Intel Corporation (INTC) -- Ignore unavailable to you. Want to Upgrade?

To: Amy J who wrote (136323)	5/30/2001 12:00:18 PM
From: Paul Engel	Respond to of 186894

Amy - Re: "Intel chips suffer massive counterfeiting" Imitation is the sincerest form of flattery. Paul

To: Amy J who wrote (136323)	6/6/2001 1:56:54 PM
From: tcmay	Read Replies (1) \| Respond to of 186894

Intel chips suffer massive counterfeiting theinquirer.net My (Tim's) comments to follow this excerpt: (I really wish SI would get serious about providing better quoting tools...) --begin excerpt-- By the Newsdesk, 30 May 2001 09.40 BST BROKERS, DISTRIBUTORS AND DEALERS in Europe are reporting a vastly increased incidence of re-marked Intel processors ... ...sealed with a counterfeit special "Intel" sticker still in place. This sticker is used to ensure the integrity of the contents - once a box has been opened, it cannot be sealed with the same tape again. That suggests that the operation is high tech and sophisticated. The CPUs use serigraphy with part numbers, again, a hard item to duplicate. The only way to really check whether a part is remarked or not is to send Intel the serial numbers for checking, but many of the chip firm's channel partners - system integrators and distributors alike - are frightened to do so. --end excerpt-- Re-marking has been a fact of life for decades. Down-binned parts are re-marked as higher-bin parts, witht the price delta being pure profit for the re-marker. (This is, ironically, a facet of the "clock chipping" issue, as a part AMD or Intel tested at "1 GHz" may work for most uses at "1.1 GHz," so the duped customer may not care...hell, he may even then take that "1.1 GHz" part and kick it up another notch to 1.3 GHz...then he wonders why Quake is flaky.) There are some good technological solutions, though. I'll spend a minute describing how the best one works. Imagine a device emitting a characteristic series of patterns. For example, a wake-up pattern for RAM bits, or a series of threshold voltages. (What these might be is an implementation detail. A system once developed by Light Signatures used a pattern of lights and darks, scratches and tool marks, on a machined metal part...the canonical application was a scheme for detecting counterfeited automobile and motorcycle parts.) The pattern is typically analog: a graph or pattern of internal current-voltage characteristics, a pattern on the surface, etc. (Again, an implementation detail. An important detail, to be sure, but not necessary to worry about overmuch right now.) Anyway, once this internal pattern (or even scratches and scuffs on the lid, a la the example above) is emitted, here's the sequence: -- the vendor takes the analog pattern and encrypts it with his PRIVATE key. (This key is stored very securely at the vendor's facilities.) -- he stamps the resulting number on the package, or even stores it in the chip. -- a customer or reseller who wishes to verify the signature takes this number and runs it through the PUBLIC key of the vendor. -- the result is the original analog sequence (of lights and darks, scratches, internal voltage-current characteristics, whatever). -- if this sequence is close enough to the analog pattern the customer sees with his own equipment, he has confidence that the part was indeed "signed" by the vendor. Importantly, having the analog sequence is not enough of course to generate a valid "mark number." Also, copying a mark number from another package is not useful, for self-evident reasons. This process is identical to a digital signature of a document. It is effectlvely a way to "sign a physical object." (For the curious, there are issues of wear and tear. In the counterfeit metal part application, the system had to have some level of tolerance for misalignment of the "light signature track" and for normal wear and tear altering some of the original patterns of scratches and grains and swirls and whatnot.) When does it become economical to do this? Clearly an end-user, home customer is not likely to do this. But a major reseller might invest in the equipment. If counterfeiting or cheating on speed bins ever gets to be a major, serious problem, solutions like this one are waiting in the wings. --Tim May