OK, this topic has taken some interesting twists, but piqued my interest to look into it a little further. Ken got back with the most interesting answer here.
VMware co-founders Rosenblum, Bugnion and Devine were all involved in three main Stanford comp science projects named Hive, SimOS and Disco. All of these were written to work on a Stanford hardware project, the Flash multiprocessor, based on a MIPS Risc Processor.
The last Co-Founder, Wang, is recognized as providing "contributions to the development of Berkeley UNIX".
All this information is pretty much right off the VMware web page with some supplements from Stanford's CS Dept documents.
This would lead me to conclude that in order to build ESX, it wouldn't take much for the design concpets of SimOS and Disco to be rewritten for x86 architecture and may have borrowed largely from what was then the only 'relatively free' OS code that ran on x86, Berkely UNIX. The relative commonality between BSD Unix and Linux drivers probably is why ports from Linux drivers are 'relatively simpler' to execute, I don't think that it is 'trivial' however or the HCL wouldn't be quite so limited.
VMware co-founders Rosenblum, Bugnion and Devine were all involved in three main Stanford comp science projects named Hive, SimOS and Disco. All of these were written to work on a Stanford hardware project, the Flash multiprocessor, based on a MIPS Risc Processor.
The last Co-Founder, Wang, is recognized as providing "contributions to the development of Berkeley UNIX".
All this information is pretty much right off the VMware web page with some supplements from Stanford's CS Dept documents.
This would lead me to conclude that in order to build ESX, it wouldn't take much for the design concpets of SimOS and Disco to be rewritten for x86 architecture and may have borrowed largely from what was then the only 'relatively free' OS code that ran on x86, Berkely UNIX. The relative commonality between BSD Unix and Linux drivers probably is why ports from Linux drivers are 'relatively simpler' to execute, I don't think that it is 'trivial' however or the HCL wouldn't be quite so limited.
OK, this topic has taken some interesting twists, but
piqued my interest to look into it a little further.
Ken got back with the most interesting answer here.
VMware co-founders Rosenblum, Bugnion and Devine were
all involved in three main Stanford comp science
projects named Hive, SimOS and Disco. All of these
were written to work on a Stanford hardware project,
the Flash multiprocessor, based on a MIPS Risc
Processor.
The last Co-Founder, Wang, is recognized as providing
"contributions to the development of Berkeley UNIX".
All this information is pretty much right off the
VMware web page with some supplements from Stanford's
CS Dept documents.
This would lead me to conclude that in order to build
ESX, it wouldn't take much for the design concpets of
SimOS and Disco to be rewritten for x86 architecture
and may have borrowed largely from what was then the
only 'relatively free' OS code that ran on x86,
Berkely UNIX. The relative commonality between BSD
Unix and Linux drivers probably is why ports from
Linux drivers are 'relatively simpler' to execute, I
don't think that it is 'trivial' however or the HCL
wouldn't be quite so limited.
piqued my interest to look into it a little further.
Ken got back with the most interesting answer here.
VMware co-founders Rosenblum, Bugnion and Devine were
all involved in three main Stanford comp science
projects named Hive, SimOS and Disco. All of these
were written to work on a Stanford hardware project,
the Flash multiprocessor, based on a MIPS Risc
Processor.
The last Co-Founder, Wang, is recognized as providing
"contributions to the development of Berkeley UNIX".
All this information is pretty much right off the
VMware web page with some supplements from Stanford's
CS Dept documents.
This would lead me to conclude that in order to build
ESX, it wouldn't take much for the design concpets of
SimOS and Disco to be rewritten for x86 architecture
and may have borrowed largely from what was then the
only 'relatively free' OS code that ran on x86,
Berkely UNIX. The relative commonality between BSD
Unix and Linux drivers probably is why ports from
Linux drivers are 'relatively simpler' to execute, I
don't think that it is 'trivial' however or the HCL
wouldn't be quite so limited.
Thanks for this background. Where did you read this background come from on VMware's web-pages - can you give us a link?
Regards
Mike Laverick
RTFM Education - VMware
Thanks for the info, George... probably the first bit of actual information in this thread
