Thursday, July 15, 2010

Of Hardware and OSs

Currently, Linux systems take the very high end machines (any machine more powerful than a fully tricked out MacPro {read supercomputers and mainframes}) and the very low end machines (phones, routers, palm-tops, PVRs). In both cases, someone is missing. Microsoft. I do not hate Microsoft. Many people would think that I did (being a Linux fan, and to a lesser extent an Apple fan), but I really don't. I just think that they have failed to adapt to a changing market place, and I feel that they are stifling the progress of IT. Apple, on the other hand, has been aiding. Apple didn't reinvent the phone. Apple showed that a Blackberry/Palm styled phone could be made for the average person. They made it, and they sold it rather well. The iPad has shown that tablets actually are in demand, and for once someone has proven that Windows is not a requirement for consumer computers. The reality is that most people think of a computer operating system as inseparable from the computer itself. So, where are we headed?

Many people are talking about the death of the desktop. This will never be the case. For the vast majority of the population, the iPad is more than enough. When all you do is some web based stuff, some word processing, and some light gaming everything can exist on a server somewhere. Of course, some files would be local (music, movies, photos, maybe some documents) but your applications would be server side, and your OS serves only as a web browser/media player and connectivity device. In this realm, we can look at Android, Chrome, webOS, and iOS and know what is coming (and apparently we look at Kin and Windows Phone 7, and we know what is not coming). Hardware wise, I feel that ARM will be with us for quite some time. Most home users will likely have ARM powered devices, and the form factors we are starting to see emerge now will stick around (laptop, netbook/tablet, phone). This is a little ironic because only a few years ago RISC processors were for workstations only. So, ARM or Atom/Neo based notebooks and tablets? We will leave room for an as-of-yet-unknown small form factor to emerge, but this is basically the home user.

For these servers with increasing loads (and likewise increasing power), we will see Linux remain strong. Apple and Microsoft both have some issues scaling at a good pace, and as we have seen both have compatibility issues with different versions of their own operating systems. This leads me to believe that Linux will gain as OSs become less and less important in the consumer market. One OS can run on your company's phones, tablets, servers, bar code scanners, fork lifts, routers, and workstations? Cool. Count me in. Another benefit that Linux offers is that any company can use it without any legal worry, pricing worry, and so on. Linux based operating systems can also be ported to any future device your company issues, and this resonates with consumers. If they already know how to use it, they are more likely to buy it.

So, where does the desktop go? Anyone doing heavy graphics or photo manipulation, anyone doing heavy audio production, anyone doing CAD work, anyone doing anything that is incredibly resource intensive will need the power that modern desktops can provide. Until such machines can be made to run at considerably lower temperatures, consume considerably less electricity, and obviously be made in smaller form factors they will never be used as tablets, and we will never get away from the existence of desktop computers.

I feel that our OS contenders of the next decade will likely be:
Linux (consumer, server, supercomputer, mainframe, embedded)
Apple iOS (consumer, embedded)
Apple OSX (workstation)
Windows (workstation, server?)
BSD/Solaris (server, mainframe)

ARM (consumer, embedded)
x86 (consumer, embedded)
AMD64 (workstation, server, supercomputer)
RISC [Power/SPARC] (workstation, server, supercomputer, gaming)
z/Arch. (mainframe)

It is possible that another OS could emerge. If Haiku is ever ready, it could easily have success in the rebuilt computer market. With a slight change to its interface, it could easily be ported to tablets as well. BSD and Solaris (if Oracle doesn't slaughter it) both have chances in this future. The server market (if taken back by Linux systems) could easily be handed over to such powerful, resource light, stable operating systems. The compatibility is there (porting among Unices is just too easy). Also, is there anything that NetBSD can't run on?


Grant Johnson said...

I disagree with one bit. I think that x86 will move out of the embedded space, replaced by ARM. Atom will make it hold on for a little while, but as Windows dies in the embedded space, ARM's power consumption will make it a clear win as soon as the OS will not prevent it from being used.

ChromeMagnon said...

You are leaving out one category which will be dominated by Google Chrome OS in future - the embedded desktop or nettop. I can see this being built into every high end TV, settop box and DVR, and I can see this being used as the standard desktop in libraries, schools, homes, hospitals etc. It is simply robust secure, requires no configuration and is stateless - this is a massive improvement for low maintenance and low administration and support cost. Libraries, schools, homes, and hospitals would be stupid not to use it instead of the current fat desktop.

Nicholas said...

I think that Microsoft will stick around in the server dept until there is a good Exchange and Active Directory replacement. So far I haven't seen one.

Post a Comment