In Part III: How Microsoft Manufactures Statistics with Gartner|IDC

THIS IS the second part of a short series that began here. It reveals the ways in which IDC and Gartner interact with Microsoft [1, 2, 3, 4, 5, 6, 7].

Comes vs. Microsoft Exhibit PX07168 [PDF] from September 2002 shows why a 'study' against GNU/Linux is being manufactured.

Microsoft's Jim Allchin, for example, states in this debrief that "we are not on a path to win against Linux" where he mentions (in point 6) the absolute necessity to "get a study done". To quote in context, "We MUST get a TOC study done. Cost is a first thing on everyone’s mind right now given the economy and pressure on cost reduction. I am not sure what the final decision was on the IDC study. We REALLY need some here. I think billv/bob own this. If the IDC report won’t cut it, then we get another one done."

Among those involved in this correspondence we find Paul Flessner, who wrote about Dell's GNU/Linux dealings: "We should whack them, we should make sure they understand our value." Brian Valentine is there too and so is Bill Veghte of the anti-GNU/Linux initiatives [1, 2, 3].

The subject of this discussion is -- as plainly put in the subject line -- "Linux".

The quote above is very real by the way. Allchin is chanting about "Facts", repeatedly (as in "Get the Facts"). Here is how it starts:

My conclusion: We are not on a path to win against Linux We must change some things and we must do it immediately. The current white papers, etc. are too high level and they are not going to cut it. Here are specific actions that I have concluded that we must take.

1. bill’s team must get a couple more hot resources assigned to it immediately who can do a tear down of Linux.

Jeff Jones (of Microsoft) has just had the IDC/IDG-owned magazines [1, 2] publish his security FUD. That was yesterday (an attack on Firefox) and here in this letter we find Allchin writing about the very same lies-based methodology that's so often criticised (counting and aggregating all packages that are peripheral to the operating system):

3. We need a comparison of the security issues published from some place like CERT for Linux vs Windows 2000 or Windows XP. We need to be sure to count all the component pieces of Linux (e.g., apache, samba, navigator, etc.).

Then come more FUD about having to recompile applications in GNU/Linux:

4. We need the technical resource / strategy resource to look for fundamental issues about Linux that customers might not know. One that I thought of while on the trip that I used dealt with the fact you need to recompile your apps, etc when a new release of Linux comes out. I don’t think anyone wants to recompile their apps when they are running them in production, etc. I am sure if we put serious IQ to the situation we can think of many issues.

Check this patent FUD out:

5. We need someone to tear down the indemnification offered from RedHat and IBM to customers. We need to understand exactly the risk a customer is under if a patent lawsuit happens and Linux is challenged. I’d like Dan to own this. There MUST be risks to customers that are being passed on. I want this understood precisely. We need to get the license from IBM given to customers and investigate.

Here is an important bit:

6. We MUST get a TOC study done. Cost is a first thing on everyone’s mind right now given the economy and pressure on cost reduction. I am not sure what the final decision was on the IDC study. We REALLY need some here. I think billv/bob own this. If the IDC report won’t cut it, then we get another one done. Some customers know that Linux isn't really free, but we need to help the other customers see this.

According to Allchin, "Customers are very disappointed (outright angry!) in how our apps interfere with each other." They need to dig some dirt on GNU/Linux, so someone is assigned for the job.

8. We need a paper explaining how we do scale up. Linux is getting in some places based on the fact that customers are trying to scale up and they can't run applications independently/safely on Windows. I will be writing mail about this separately. The paper would explain how to do it on Windows. Customers are very smart about the problems (eg., registry issues, lack of support from 3rd parties if more than a single app is running on the system (even though it works), etc.). We need a story here. I think this has to be someone as smart as davidds, I am not sure who should own this. It is possible that someone Iike blaing should do this. Maybe someone in Paulfle’s team can do it. We must address MSFT server products in this (e.g, sql, exchange, etc.). Customers are very disappointed (outright angry!) in how our apps interfere with each other.

Microsoft needs to manufacture some "quick facts" because, according to Allchin, "The paper [against GNU/Linux that] we have today as I said was laughed at in one place."

10.. We need to put together a single short paper (just a couple of pages - maybe a single sheet) for a leave behind for customers of the 10 questions that they should consider before adopting Linux This should be based on the learnings we get from some of the items above and well as others that we find. These should be as hard hitting as we can be, BUT they need to be factual based, We should be thinking about howto put things like Red Hat’s server price In it to show where it’s going, etc, The paper we have today as I said was laughed at in one place. (I am not sure what one they had seen,) I think Billy needs to figure out who should own this I think someone like vlcg would be great.

Jim Allchin then admits that he is scared when he says under the subject line "Linux":

I am scared.

The third and last part of this series will give most of the 'beef' of this story. The correspondence, in full, is below. ⬆

---

Appendix: Comes vs. Microsoft - exhibit px07168, as text

---

From: Jim Allchin Sent: Friday, September 27, 2002 10:20 PM To: Brian Valentine, Paul Flessner; Bill Veghte, Bob Kefly; Eric Rudder, Dan Neault Cc: Michele Freed Subject: Linux

I am travelling as you know. I have had many VAP round tables (both partner and not), met with many enterprise customers, had many roundtables with the field - both sales roundtable and technical roundtables. I also met with the US seminar group I have seen a tot of customers this week.

My conclusion: We are not on a path to win against Linux We must change some things and we must do it immediately. The current white papers, etc. are too high level and they are not going to cut it. Here are specific actions that I have concluded that we must take.

1. bill’s team must get a couple more hot resources assigned to it immediately who can do a tear down of Linux. This resources must help do the items I outline below together with other teams. I am not sure who might be best to assign here. Michele and I have been brainstorming and we don’t have a long list. One person that I think we could consider recruiting to help would be vicg. Another person might be brian hall. These guys aren’t super technical, but they are both good at distilling the essence and packaging stuff. Vic would be better than brian, but I wanted to throw both names out. I would only pick one of them. Just as importantly we need a kynl type of person - someone very technical. I insist that we find someone within a week and assign them on this for the next 2 months. I am not sure who to pick on the technical side. Once you read below I would like some suggestions. I think can be done in 2 months.

2. We need a paper which outlines technically how our system (kernel, web service, file server) is better. I think we have Mark R signed up to write this (robs was supposed to engage him on this). This paper needs to cover things like the facts that we have a preemptive kernel, asychronous I/O, etc Facts... that go to the core of why windows is different and Linux is old unix. Facts. I would have the technical person help with this. There was a technical wrlteup a few years ago by Mark in Windows NT Magazine. We need more on this.

3. We need a comparison of the security issues published from some place like CERT for Linux vs Windows 2000 or Windows XP. We need to be sure to count all the component pieces of Linux (e.g., apache, samba, navigator, etc.). This needs to be fact based. It should be short and sweet. A table would be great. I think Mikenash owns this.

4. We need the technical resource / strategy resource to look for fundamental issues about Linux that customers might not know. One that I thought of while on the trip that I used dealt with the fact you need to recompile your apps, etc when a new release of Linux comes out. I don’t think anyone wants to recompile their apps when they are running them in production, etc. I am sure if we put serious IQ to the situation we can think of many issues.

5. We need someone to tear down the indemnification offered from RedHat and IBM to customers. We need to understand exactly the risk a customer is under if a patent lawsuit happens and Linux is challenged. I’d like Dan to own this. There MUST be risks to customers that are being passed on. I want this understood precisely. We need to get the license from IBM given to customers and investigate.

6. We MUST get a TOC study done. Cost is a first thing on everyone’s mind right now given the economy and pressure on cost reduction. I am not sure what the final decision was on the IDC study. We REALLY need some here. I think billv/bob own this. If the IDC report won’t cut it, then we get another one done. Some customers know that Linux isn't really free, but we need to help the other customers see this.

7. We need a paper (which we may already have) on the productively gains possible on with .NET development over php and j2ee. This must have examples of how productive development is, performance, and operational capability. Eric owns getting this down if it isn’t already.

MS-CC-RN 000001039342 HIGHLY CONFIDENTIAL

---

8. We need a paper explaining how we do scale up. Linux is getting in some places based on the fact that customers are trying to scale up and they can't run applications independently/safely on Windows. I will be writing mail about this separately. The paper would explain how to do it on Windows. Customers are very smart about the problems (eg., registry issues, lack of support from 3rd parties if more than a single app is running on the system (even though it works), etc.). We need a story here. I think this has to be someone as smart as davidds, I am not sure who should own this. It is possible that someone like blaing should do this. Maybe someone in Paulfle’s team can do it. We must address MSFT server products in this (e.g, sql, exchange, etc.). Customers are very disappointed (outright angry!) in how our apps interfere with each other.

9. I think we need a paper on SFU and interop. Customers believe that unix systems are more compatible with each other and more interoperable, I was stunned at the number of customers who had no idea about what SFU could do, We must promote this much more. Billv owns ensuring this is done.

10.. We need to put together a single short paper (just a couple of pages - maybe a single sheet) for a leave behind for customers of the 10 questions that they should consider before adopting Linux This should be based on the learnings we get from some of the items above and well as others that we find. These should be as hard hitting as we can be, BUT they need to be factual based, We should be thinking about howto put things like Red Hat’s server price In it to show where it’s going, etc, The paper we have today as I said was laughed at in one place. (I am not sure what one they had seen,) I think Billy needs to figure out who should own this I think someone like vlcg would be great.

_________

I know there is a lot in this email, I am sorry. It is serious guys, The field does not feel supported by us. We are not giving them what they need to win

Bill/Paul: I need to ask you to take ownership of driving this ahead What I want to see is a package including ALL of these items that we can provide to the field within 2 months (MAX). I am scared. Again.. I want the two people assigned within a week. I want to know who the people are. Eric please help thinking about who the right people are. Please remember NO marketing. Facts. No anger toward Linux. Just facts. Please understand this isn’t up for discussion. I want some solid people assigned ASAP.

I would like a review in 1 month on the progress on this.

thanks,

jim

5/6/2005

MS-CC-RN 000001039343 HIGHLY CONFIDENTIAL

---