> Or, more accurately, Linux stopped Unix in its tracks
Thats hardly true. When Linux was ramping up steam the Big Iron companies were the only real option for stability and performance. Also for years the largest sites on the internet ran FreeBSD, not Linux.
What killed Big Iron was the cost, pure and simple; its much easier to scale horizontally with lots of smaller machines than piles of giant machines. We can thank open source software and operating systems for that.
To some extent, Microsoft also helped kill Big Iron as well. The advent of IIS and SQLServer made it possible to do real work on commodity hardware. PC Graphics advances killed the high end Unix Graphics workstation business too.
well, let's see. in '99 I could buy a $50K SGI with IRIX or a $2K linux machine. With the exception of graphics and heavy floating point, the linux machine did everything the irix machine did. So I bought 6... and saved myself $25K.
Judging purely from my own experience, I would say yes. I was a Solaris user from about 1995 until about 2000.
Around about 2000-2001, I bought a new computer with a large (for that time) hard drive that was (IIRC) about 30 gigs. I found that my Solaris x86 was not able to access the whole of that disk due to disk capacity limits in its configuration.
On the off chance, I also tried installing Linux. It had no problems at all in working with that large(!) disk.
(Note: I am trying to remember back roughly 20 years, I may have some numbers wrong.)
That was when I put my Solaris away, ending 10 years of using several commercial unices, and moving across to Linux.
I would say Intel and AMD killed alternate CPU architectures. Linux was the knife, but the arm (sorry) wielding the knife was the instruction set.
ARM destroyed the others in the mobile platform. If linux hadn't existed to become Android, I am sure some other general purpose OS would have filled the void.
VX works was not capable of doing this. Nor was OS/2. So in some senses, yes, Linux was necessary. Why SPARC or MIPS didn't persist as a major component of worldwide platform lies in questions around how they were targetting size, cache, speed, die yield. Perhaps if Sparc had been spun off as an independent foundry it could have succeeded but I suspect Solaris wasn't heading to the pricepoint where it worked for a killer-platform.
So maybe the meme is "price killed commercial unix"
8 comments
[ 2.7 ms ] story [ 30.7 ms ] threadThats hardly true. When Linux was ramping up steam the Big Iron companies were the only real option for stability and performance. Also for years the largest sites on the internet ran FreeBSD, not Linux.
What killed Big Iron was the cost, pure and simple; its much easier to scale horizontally with lots of smaller machines than piles of giant machines. We can thank open source software and operating systems for that.
To some extent, Microsoft also helped kill Big Iron as well. The advent of IIS and SQLServer made it possible to do real work on commodity hardware. PC Graphics advances killed the high end Unix Graphics workstation business too.
TLDR; Linux only contributed, its not the cause.
Around about 2000-2001, I bought a new computer with a large (for that time) hard drive that was (IIRC) about 30 gigs. I found that my Solaris x86 was not able to access the whole of that disk due to disk capacity limits in its configuration.
On the off chance, I also tried installing Linux. It had no problems at all in working with that large(!) disk.
(Note: I am trying to remember back roughly 20 years, I may have some numbers wrong.)
That was when I put my Solaris away, ending 10 years of using several commercial unices, and moving across to Linux.
ARM destroyed the others in the mobile platform. If linux hadn't existed to become Android, I am sure some other general purpose OS would have filled the void.
VX works was not capable of doing this. Nor was OS/2. So in some senses, yes, Linux was necessary. Why SPARC or MIPS didn't persist as a major component of worldwide platform lies in questions around how they were targetting size, cache, speed, die yield. Perhaps if Sparc had been spun off as an independent foundry it could have succeeded but I suspect Solaris wasn't heading to the pricepoint where it worked for a killer-platform.
So maybe the meme is "price killed commercial unix"
NetBSD for Sparc has existed since 1993 or so.
OpenSPARC has existed since 2005 and multiple operating systems have run on Sparc since the 90s.
Linux gained adoption the same way OSX did.