anon (anon.delete@this.anon.com) on February 2, 2013 10:21 pm wrote:
> Richard Cownie (tich.delete@this.pobox.com) on February 2, 2013 6:55 pm wrote:
> > Ricardo B (ricardo.b.delete@this.xxxxxx.xx) on February 2, 2013 6:31 pm wrote:
> >
> > > Intel's margins in the x86 business because they manage to deliver a win-win: better products
> > > (performance, power) with lower production costs (area, yield) than the competition.
> >
> > Intel's competition in x86 has been constrained by the patent and licensing
> > issues around the various versions of the x86 ISA.
>
> Yes, although we already demonstrated that this is a red herring for the purpose of discussing
> competitiveness in non-x86 markets, by their way of taking much of the RISC/UNIX market.


What you're missing here is that x86 compatibility was actually a big advantage for Intel competing with RISC/Unix, because of the Windows monopoly and the growing influence of Windows Server on that monopoly (i.e. Exchange, AD, file serving)

x86 being compatible on the desktop helped push out RISC/Unix workstations much more quickly than if there had been Unix software as compatible with Exchange and Office as there is today. If Evolution and LibreOffice existed in their modern form in 1995, things might not have worked out quite as well for Intel, or certainly not quite as quickly. The need for everyone to be on Exchange for meeting scheduling, be able to read/write Office documents, and later to access corporate intranets that only functioned with IE and ActiveX killed the Unix workstation market. It was either move to Windows, or have both a workstation and a PC on your desk. The guys who wanted both often were forced to migrate to Linux (on x86) for cost/support reasons.

x86 being compatible on the server allowed various functions to be brought into AD and avoid the headaches of having many unsynchronized authentication systems throughout the organization. Unix eventually became able to use AD for this, but the years it took for this to be a workable option saw most of the lower end Unix server market disappear. Once NT was also established as the defacto file server, mail server and intranet/application server, the numbers quickly grew much larger than the numbers of Unix servers.

It helped that Windows Server was such a piece of crap that it couldn't be trusted to run at loads anything like what a Unix server could, so you needed more for the same amount of work and couldn't put multiple applications on the same server. This meant there were quickly a LOT of Windows Servers in the typical datacenter, making the decision easier for some companies to "standardize" on Windows and make the acquisition of a Unix server require additional justification. And, of course, x86 servers cost a lot less, and combined with the cost advantage of Linux allowed companies to buy Unix servers for less if they use x86, and RISC/Unix was pushed out of the low end server market.

In the phone market, there is no benefit at all to x86 compatibility, ARM has some advantage as the incumbent, though nothing like the massive advantage conferred on Intel by Microsoft when competing with RISC/Unix. For microservers, despite all the talk of ease of porting, there is definitely some residual x86 benefit since the larger Linux servers will all be x86, so Intel only has to come close to ARM to own this market.