backtop

Print E-mail del.icio.us 73 comment(s) - last by carage.. on Mar 9 at 12:53 AM
Court documents reveal Intel data slated for retention is permanently lost

In an unpublished statement to the U.S. District Court of Delaware, AMD alleges Intel allowed the destruction of evidence in pending antitrust litigation.

According to the opening letter of the AMD statement, "Through what appears to be a combination of gross communication failures, an ill-conceived plan of document retention and lackluster oversight by outside counsel, Intel has apparently allowed evidence to be destroyed."

Intel's current email system automatically purges emails sent or received by its employees every 35 days.  Senior executive data is purged every 45 to 60 days.  Additionally, Intel's backup system recycles every other cycle -- immediately overwriting any backup data during tape rotation.

AMD alleges that more than a third of 1,027 case-specific Intel employees did not receive instructions to retain their data after the 2005 case initiation.  Of the individuals who retained data, AMD alleges the majority did not retain "sent" emails.  These employees, dubbed "custodians," are persons of interest in the legal proceedings. 

According to Intel, 217 of these 1,027 custodians have been "identified," and must retain all data as per instruction of the court.  AMD has the right to identify another 254 employees for court scrutiny of data -- to date AMD has already identified 74 of those 254.

Intel admits the data lapse, claiming "Intel does not have weekly back-up tapes for every custodian on the custodian list. Some were inadvertently not migrated to the server in 2005, and some, who were later identified, were not migrated on such identification.  In addition, some weekly back-up tapes appear to have been recycled."

Intel's letter to Judge Joseph Farnan also explains that layoffs and corporate restructuring in 2006 caused some oversight in the data retention.

"In the course of routine work on the case we learned that a small percentage of post-filing e-mail was not being retained in the way we believed it was.  This led to a broader study and the implementation of new procedures," Intel spokesman Chuck Malloy said to DailyTech.  "We are still actively checking the availability of back-up tapes and secondary sources to find every bit of e-mail in question.  We haven't yet finished that effort."

The AMD report counters, "AMD always believed -- and for good reason, still believes -- that the most probative evidence of Intel exclusion would reside in the electronic files and documents Intel created after the lawsuit started, evidence that Intel would be obligated to preserve."

AMD has asked the court to have Intel supply a list documenting a custodian-by-custodian tally of the retention inventory and any salvageable backup tapes.  

Expect the official trial for the proceedings to start in 2009.  Statements made by AMD and Intel in court may be published later this week.


Comments     Threshold


This article is over a month old, voting and posting comments is disabled
Right Mind
By aurareturn on 3/5/2007 11:12:13 PM , Rating: 2
Any consumer in the right mind would want AMD to win. You can't have competition if your competition is a monopoly.




RE: Right Mind
By deeznuts on 3/5/07, Rating: -1
RE: Right Mind
By Dactyl on 3/6/2007 12:11:37 AM , Rating: 1
Monopolistic behavior is possible even with some competition. Intel may have violated rules against monopoly-type behavior. That's what this case is all about.

How does Intel's behavior harm the consumer? If AMD had been making more money during 2000-2004, they would have been able to spend more on R&D, for instance to bring Barcelona out faster.

And it's the threat of CPUs like Barcelona that forces Intel to make such great CPUs like Kentsfield and Tigerton. Wanna bet Intel is adding more floating point power to Tigerton in order to counter Barcelona? Those CPUs are going to make Alan Wake very fun.


RE: Right Mind
By Viditor on 3/6/2007 2:10:28 AM , Rating: 2
Well, more to the point:

1. A monopoly isn't a behavior...I think you mean anti-competitive behavior from a monopoly.

2. It harms the consumer in the following way.
a) Intel sets individually targeted sales rebates for each store or OEM, saying that if 90% of your sales are Intel in the previous month, we will give you $x.xx as a reward.
b) If you've studied retail marketing, you'll know that this rebate always goes to the OEM's profit line and is never passed on to the consumer (how could it be since the amount isn't known until the month is past?).
c) By assuring themselves that OEMs are far less likely to use AMD chips (that rebate was a HUGE part of their profit margin as Dell's little fiasco can attest to), Intel is free to price their chips at a higher amount because competition (free market) is no longer setting the price.


RE: Right Mind
By Lakku on 3/6/2007 4:03:11 AM , Rating: 2
Alan Wake will run just fine on a Kentsfield right now, so yes, I would assume future CPUs would be fairly good at the task as well. Kentsfield appears to be their dev CPU and is the target, and it has been what Alan Wake has been shown on thus far. Valve is also using Kentsfields to develop future Source games and engine refinements. You can get a good future CPU right now, but waiting isn't a bad thing either. At any rate, Alan Wake, per the devs, runs best on quad core. They said the physics thread alone takes up 80% of a core, so a dual core would have less to handle everything else and a single core would be almost useless.


RE: Right Mind
By deeznuts on 3/6/2007 6:05:46 AM , Rating: 3
I know what the case is about. But monopolies themselves are not illegal. Trying to make your company a monopoly is not necessarily illegal (it's competition). It's when you engage in anti-competitive or unreasonable practices that is illegal. Regardless if you are trying to become a monopoly or just to compete, it's the practices themselves that concern the government.

If everybody just stopped buying AMD processors tomorrow and everyone started buying Intel, and Intel did nothing wrong but compete, no foul.

The term monopoly is a bit of a red herring. Whether a company is trying to become a monopoly or just wants to compete unfairly, it's the practices, not the intention, that is the heart of any antitrust case.


RE: Right Mind
By ElJefe69 on 3/6/07, Rating: -1
RE: Right Mind
By JeffDM on 3/7/2007 10:17:20 PM , Rating: 2
I don't understand why people think that Tigerton is some magical next generation processor, the silicon is exactly the same as Kentsfield except that it will operate in four-socket computers to make a 16 core server. Basically, it's the Xeon MP chip. You won't be playing a video game on a Tigerton-based computer unless you have far more money than sense or you buy one many years from now. It's for servers that cost tens of thousands of dollars.


RE: Right Mind
By Viditor on 3/6/2007 2:13:17 AM , Rating: 2
quote:
How is Intel a Monopoly? AMD exists doesn't it?

This is a common misconception...
A legal monopoly doesn't mean that there is only one company, it means that there is only one dominant company. Usually a company with 50% of the market or greater is considered a monopoly...


RE: Right Mind
By cheetah2k on 3/6/2007 3:44:20 AM , Rating: 2
Of course Intel was a monopoly.... I mean, we all know about the threats to DELL and others that they made if they outsourced tech elswhere (eg. AMD).

And funnily enuff (or coincidence), DELL's bottm line is now suffering after it picked up AMD as an OEM.... Makes you wonder what stunt Intel pulled when this happened.. Intel probably hit them with a 25% cost increase in their bulk buy agreements.


RE: Right Mind
By SacredFist on 3/6/2007 4:11:17 AM , Rating: 2
Isn't Intel like a biopoly?


RE: Right Mind
By RobFDB on 3/6/2007 4:49:28 AM , Rating: 2
The word you're looking for is duopoly, and no, the microprocessor market isn't quite a duopoly as duopoly assumes only 2 producers. A good example would be the american political system, but even that isn't perfect due to the existance of independants.


RE: Right Mind
By jebork on 3/6/2007 10:31:30 AM , Rating: 1
Yes, but you never hear AMD say "Our goal is making a duopoly". Throwing stones at the other guy is soooo much more fun.


RE: Right Mind
By ElJefe69 on 3/6/07, Rating: -1
RE: Right Mind
By deeznuts on 3/6/2007 6:12:43 AM , Rating: 1
I think you need to look up the word "legal monopoly." A legal monopoly is one granted by the government and protected by its laws to be a monopoly.

Your definition of a "legal monopoly" .... 50%? did you read this somewhere or did you hear it secondhand? Because you described dozens of industries dominated by what I call "market leaders" "industry leaders" etc.

A monopoly is one seller, no good substitutes, price maker. It's a simple definition and I don't see anywhere where the law deems "de facto" monopolies under any other circumstances. I can see if you go as far as "practically" one seller such as AMD blowing up and VIA now cometing with Intel. But you're talking 2-4% maybe, max.

So it's not a common misconception, it's the definition of monopoly.


RE: Right Mind
By nah on 3/6/2007 7:08:33 AM , Rating: 3
Monopolies in economic theory exist when one firm produces an industry's entire output---the legal definition is 25% of the total output--there is a measurement called the Herfindahl Index which measures the sum of the squared values of the market shares. In perfect competition, with many firms, this index is 0 ( 0squared + 0 squared +....so) where 0 represents each firms market share. In a perfect monopoly the index is 10,000 ( 100squared)---as there is only one firm. In the CPU market the Herfindahl Index would be 75squared + 25squared = 6250.


RE: Right Mind
By Oregonian2 on 3/6/2007 2:53:52 PM , Rating: 2
You're saying with perfect competition the index is zero, and you say the only way (mathematically) for that to happen if every company has a zero market share (so their squares add up to zero).

So there's a perfect competition ONLY when there are an infinite number of competitors with equal shares (so shares are zero, even though that infinite number of zeros add up to 100) or when the market has zero sales with a finite number of companies that each have zero market share (again so the index can be zero).

Sounds like something a college prof who's never been outside of academia would think up. Herfindahl a college-prof type?

To be fair, the index sounds reasonable so long as one is nowhere
near the low-side of the index scale where the lower boundary
condition is silly in practice (the marketplace for star-trek
enterprise sort of transportation craft is a perfect competition because all companies have a zero marketshare).

There must be an associated chart that gives "grades" to indexes
(even if an academia generated one) and that would be interesting to see in order to evaluate such index numbers. In any case, the CPU market has a lot of other players, so it depends upon how one defines the market as well (the market for "intel inside" computers gives Intel a 100% marketshare). There also are things like the SPARC processors that Sun uses, IBM has a lot of processors that their computers (of all sizes) use that aren't intel or AMD, and the imbedded market probably uses a LOT more non-x86 processors than the x86 market as a whole. Things like the ARM processor line are everywhere, it's just not as visible.


RE: Right Mind
By nah on 3/6/2007 10:17:18 PM , Rating: 2
Yes, Herfindahl is college prof type--and the index is something only a prof would love--but it is used by the Federal Government to determine whether mergers are acceptable or not--

As for CPU markets you are right--in terms of numbers--non x86 CPUs constitute a majority of CPUs sold--but in terms of value,it's a different ballgame--when I meant the CPU market I was referring to the server, desktop, and laptop one.


RE: Right Mind
By Viditor on 3/6/2007 10:22:13 PM , Rating: 2
quote:
There also are things like the SPARC processors that Sun uses, IBM has a lot of processors that their computers (of all sizes) use that aren't intel or AMD, and the imbedded market probably uses a LOT more non-x86 processors than the x86 market as a whole

True...and that is what Intel has been trying to say in their arguments.
However, the Microsoft case set the precedent for x86 being a market unto itself, so I believe they have abondoned this argument.