Rumors are circulating that CPU designer ARM will soon announce a 64-bit chip. This would remove the last theoretical blockage that prevents ARM chips being taken seriously in the enterprise space, although it’s not yet clear if ARM intends to invade such markets or is merely offering greater flexibility for manufacturers who have given ARM chips a home in the majority of cellphones, handhelds and portable devices produced recently.
Whatever the case, because ARM licenses its technology to third-party manufacturers, it’s unlikely we’ll see any actual 64-bit ARM implementations for at least a year or two.
ARM is very late to this market: AMD and Intel have both sold 64-bit chips since the beginning of this century, and now it’s nearly impossible to buy a computer that doesn’t contain a 64-bit processor.
ARM is an entirely different CPU architecture compared to that offered by Intel or AMD, and this presents some potential limitations when it comes to enterprise use: Microsoft Windows won’t work on ARM chips (although there have been rumors this could change), although Linux is at home on an ARM chip as on x86, and already dominates in the enterprise space. But this means Windows virtualization is virtually impossible too.
However, it’s easy to imagine a service such as Amazon’s EC2 offering virtualized Linux instances without the user being aware that it’s an ARM setup, and these could be cheaper than equivalent x86 instances (perhaps even making for a “budget EC2” service). Storage cloud services such as the architecture-independent Amazon S3 could easily switch to ARM.
In the world of enterprise computing the limitations of 4GB of memory are easily reached, and this is especially true in the world of virtualization and cloud computing. ARM’s most recent chip design, the Cortex-A15, attempts to bypass this limitation with its Large Physical Address Extensions (LPAE), which allows access to up to 1TB of memory. However, the chief benefit of a move to pure 64-bit architecture is chiefly psychological; 32-bit simply isn’t taken seriously in the corporate space, and would make a very hard sell for the likes of pointy-haired bosses.
It’s not yet clear if ARM’s new architecture will overcome the traditional criticism that ARM chips are slower than equivalent Intel and AMD chips. However, as I discussed recently, a move to ARM architecture in the data center could provide other compelling benefits, including significant heat reduction and therefore a reduction in the cost of cooling. Quite how significant a reduction is yet to be seen, although there have been rumors that Facebook might soon make the jump.
It’s also possible that ARM’s new architecture will bring with it various CPU architecture enhancements that boost performance. However, bearing in mind that ARM chips see most sales right now in the mobile phone arena, it’s likely that multimedia data manipulation will be a strong focus, rather than anything that enterprise users are likely to find useful.
That said, the Cortex-A15 includes hardware virtualization technology but again it’s believed this is aimed at mobile users, with the goal of running two operating systems side-by-side–one for personal usage, and one for business.
Keir Thomas has been writing about computing since the last century, and more recently has written several best-selling books. You can learn more about him at http://keirthomas.com.