Jun 25th, 2012
Talk around Microsoft’s new upcoming desktop operating, Windows 8, centers almost completely on its dramatically departing Metro User Interface, borrowed from the Windows Phone which, in turn, borrowed it from the ill-fated Microsoft Zune. Apparently Microsoft believes that the third time is the charm when it comes to Metro.
To me the compelling story of Windows 8 comes not in the fit and finish but the under the hood rewiring that hint at a promising new future for the platform. In the past Microsoft has attempted shipping Windows Server OS on some alternative architectures including, for those who remember, the Digital Alpha processor and more recently the Intel Itanium. In these previous cases, the focus was on the highest end Microsoft platforms being run on hardware above and beyond what the Windows world normally sees.
Windows 8 promises to tackle the world of multiple architectures in a completely different way – starting with the lowest end operating system and focusing on a platform that is lighter and less powerful than the typical Intel or AMD offering, the low power ARM RISC architecture with the newly named Windows RT (previously WoA, Windows on ARM.)
The ARM architecture is making its headlines as Microsoft attempts to drive deep into handheld and low power devices. Windows RT could signal a unification between the Windows desktop codebase and the mobile smartphone codebase down the road. Windows RT could mean strong competition from Microsoft in the handheld tablet market where the iPad dominates so completely today. Windows RT could be a real competitor to the Android platforms.
Certainly, as it stands today, Windows RT has a lot of potential to be really interesting, if not quite disruptive, with where it will stand upon release. But I think that the interesting story lies beneath the surface in what Windows RT can potentially mean for the datacenter. What might Microsoft have in store for us in the future?
The datacenter today is moving in many directions. Virtualization is one driving factor as are low power server options such as Hewlett-Packard’s Project Moonshot which is designed to bring ARM-based, low power consumption servers into high end, horizontally scaling datacenter applications.
Currently, today, the number of server operating systems available to run on ARM servers, like those coming soon from HP, are few and far between and are mostly only available from the BSD family of operating systems. The Linux community, for example, is scrambled to assemble even a single, enterprise-supported ARM-based distribution and it appears that Ubuntu will be the first out of the gate there. But this paucity of server operating systems on ARM leaves an obvious market gap and one that Microsoft may be well thinking of filling.
Windows Server on ARM could be a big win for Microsoft in the datacenter. A lower cost offering broadening their platform portfolio without the need for heavy kernel reworking since they are already providing this effort for the kernel on their handheld devices. This could be a significant push for Windows into the growingly popular green datacenter arena where ARM processors are expected to play a central role.
Microsoft has long fought to gain a foothold in the datacenter and today is as comfortable there as anyone but Windows Servers continue to play in a segregated world where email, authentication and some internal applications are housed on Windows platforms but the majority of heavy processing, web hosting, storage and other roles are almost universally given to UNIX family members. Windows’ availability on the ARM platform could push it to the forefront of options for horizontally scaling server forms like web servers, application servers and other tasks which will rise to the top of the ARM computing pool – possibly even green high performance compute grids.
ARM might mean exciting things for the future of the Windows Server platform, probably at least one, if not two releases out. And, likewise, Windows might mean something exciting for ARM.