By 1997, Citrix was reaping in a great deal of success based on their WinFrame 1.x product line. This software was based on Windows NT 3.51 and instead of just patching a 3.51 installation, Citrix actually shipped a full modified version of NT 3.51. This work had been done by a small core of developers between 1993 and 1995. Work in 1996 focused on supporting the upcoming NT 4.0 release from Microsoft. This version was unofficially called WinFrame 2.0. Citrix had worked hard to support the latest changes and had already merged and tested the code base by early 1997.
Early in 1997, Microsoft declared that it would no longer license the operating system code to Citrix for NT 4.0. In fact, it became clear that Microsoft wanted such work to stay in house. Microsoft had some ventures (investments) in an alternative solution for making NT multiuser. I don’t remember the name of this group but I’m sure someone out there will. The message to Citrix was to either sell the multiuser component or Citrix wouldn’t have this part of the market anymore. Tough negotiations proceeded. The stock price sank. Speculation was rife. Many Citrites of the time saw this as a threat to the survival of the company.
Later that year Citrix and Microsoft struck a deal to sell the multiuser code to Microsoft for something like 185 million dollars over five years. Microsoft would not buy ICA but would instead develop RDP. RDP was originated from another software acquisition. Citrix would no longer be allowed to ship its own versions of operating systems and would be only given a restricted access to the source with no rights to build derivative works from such source.
The company, obviously, survived this crisis. I think it was mostly due to the strength of the work to do multiuser and I would give strong credit to John Richardson, Brad Pedersen, and Mike Discavage. Brad is still with the company and is the Chief Architect and the sole Citrix Fellow. He started with the company in 1989 and obviously is one of the originating developers. His strong Unix background made all the difference during the early years of the company.
1997 was the year I left Citrix to move to Australia. I was offered a Systems Engineer job for Citrix but decided I didn’t want to travel to all the English speaking countries in the southern hemisphere as part of my job. I started with a local reseller in Brisbane instead.
Live near Brisbane, Australia. Software developer currently focused on iOS and Android. Avid Google Local Guide
Red Circle Blog 
The licensing problem I heard was actually a fault of someone inside Citrix. They had not negociated the reseller license and by the time they actually went to, Microsoft had changed thier mind due to a certain fear of some network devices that may give competition.
The initial WinFrame 1.x sales were 100% profit for Citrix however Microsoft now wanted Citrix to pay something like 1,000 per sale. Citrix wouldn’t agree so like you said Microsoft attempted to provide this solution in house. This was done in two teams, one was internal and another was through an acquision of something named “prologue” or something.
The problem was that they could only scale those solutions to 5 users. Citrix was showing users being scaled to 180. The secret was “Session Space” then patented by John Richardson (Now, incidently of Microsoft). John had stolen memory from the file system cache manager to create session space to get around the issue of the window manager (win32k), display drivers and printers eating up Virtual Address space, something that was not a problem in NT 3.51 since this was all in user mode.
So instead of attempting to completely rearchitect those components to be “session aware” such as CDM, they recieved their own address space. This then worked for sessions in the kernel as user space does for processes and thus allowing code to be shared, global data to be seperate and not inflat virtual address space of the kernel.
John Richardson now works for Microsoft.
Very interesting feedback.
It matches fairly well with what I heard happened. I’ve never heard the figures for what Microsoft wanted. From what I heard at the time, it was not negotiable to get the source license for modification for NT 4.0.
Years ago it was considered a secret that Microsoft was an investor in Citrix. I think it would go back to 1989.
1997 really marked the beginning of big changes in the relationship.
Citrix seems to be losing it’s most valuable engineers, those who know and understand Terminal Services including the inventor of Seamless over the last year. There have been several other key players leaving, including an influencial leader in your office.
There seems to be a mass exodous from the company recently; how has this damaged the moral and how can you keep replace years of knowledge and experience with new hires. Any thoughts?
With such a large company, it is likely that people will eventually leave. Very few of the original Citrix people are still around. I’m sure the same would be true at places like Microsoft.
I don’t think the moral has been deeply affected. There are still plenty of talented people left. In some cases, it has become easier for people to rise because the top layers have decided to move on.
Some key people in the Citrix Redmond branch have left and I think this in part what you are referring to. Two of the people went straight to Microsoft. I’m sure they would have seen a much bigger opportunity awaiting them on the other side of the fence.
I’m sure things could be better in any organization and I’ll also sure that management has picked up on the need to retain people. It is very hard to hire for some of the jobs that engineers do. This is especially true in cases where the code is poorly documented and the programmer was very reticent about sharing information about what he or she has done.
I’m not as worried about core knowledge since I personally know people that have either acquired core knowledge or could learn it fairly quick.
So, in summary, I don’t think it is as bad as it seems. I also think that with all the acquistions of the last few years tends to make up for any losses since essentially Citrix is buying talent from other companies in new fields. There are plenty of smart people in the world.
Please be assured that the “team” that worked on Seamless (myself included – I modestly add :)) and made it what it is today still very much works at Citrix.
With regards to our core technologies including TS, we have several engineers with extremely deep knowledge of these core elements. I happen to manage some of the most knowlegeable engineers in these core elements. It always amuses me when I read in the press about the exodus of our core talent as I know we in the Citrix development engineering team are sucessful because we are a team and thus so much more than just the few engineers (some of whom admittedly were brilliant and one in particular case my technical mentor from my days as a junior engineer) who have moved on.
[…] The included article by Kevin Maney written in 1997 does a great job capturing the elements of what happened during the 1997 Microsoft/Citrix crisis. There is a real sense of urgency for survival. As I have said before, it should not be forgotten. It is dangerous to be idly complacent. […]
[…] time you're no stranger to the Citrix rumor mill. At one time there were stories about Citrix being doomed as a company, acquired by Microsoft, acquired by Oracle and even IBM or Cisco acquiring them. Of course […]