Evaluate Weigh the pros and cons of technologies, products and projects you are considering.

The ARM-based server vies for first in your data center

Are the alluring functions of ARM servers enough to cause a reshaping of data center operations? Or will the x86 server continue its domination?

ARM-based servers are light on power usage and heavy on processing power. Will those features be enough to loosen...

the stranglehold that x86 systems have on the general-purpose server marketplace?

The x86 chip is a go-to for low-cost, high-performance servers. These systems have gradually replaced higher-end Unix and proprietary microprocessor systems in enterprise data centers. Intel is the main player in the commodity server microprocessor space, appearing in nearly 95% of x86 servers, while Advanced Micro Devices Inc. (AMD) has most of the rest.

ARM-based servers are making their way into the data center. Rather than fighting a single ARM competitor, Intel faces an ad hoc consortium. ARM chips gained traction in the mobile market and ARM-based designs are now slowly moving into the enterprise server space. Chipmakers like AppliedMicro and AMD, and hardware suppliers like HP, develop ARM-based servers.

The first wave of ARM-based devices started when HP partnered with AppliedMicro on its ProLiant m400 server cartridge for the Moonshot 1500 microserver chassis in 2014. The enterprise-class 64-bit ARM v8 server is positioned as a Web tier/caching box for service providers, commercial Internet providers and the ARM server/mobile software development community.

ARM vs x86 systems

The x86 servers use complex instruction set computing (CISC) that operates with direct memory and system registers. The instructions supported by x86 enable the processors to run large, complex workloads.

ARM is a reduced instruction set computing (RISC) architecture system. It only operates on registers with a few instructions for loading and saving data to and from memory. One benefit of the RISC design is power efficiency. An ARM-based processor with a clock speed of 1 GHz fits into a 2-Watt thermal design power (TDP), while Intel's Atom processors used in microservers struggle to dip below a 5-Watt TDP, according to market research firm Moor Insights & Strategy in Austin, Texas. Lower TDP numbers generally represent lower overall power draw. Consequently, replacing standard x86 1U servers with ARM systems could result in power savings of as much as 35%.

Each architecture runs different operating systems. ARM-based servers are the domain of mobile systems, such as Android and iOS. The x86 processor is associated with Windows and Linux. While ARM-based systems have plenty of end-user software, little of it exists at the server level. These new servers depend on vendors and enterprises that modify software stacks to be compatible with the ARM architecture. The prevalence of Windows and Linux servers gives the Intel server ecosystem an advantage.

Intel is incredibly protective of its license on the x86 instruction set architecture. Only AMD and VIA Technologies have the ability to develop and build new x86 processors.

ARM Holdings is essentially an intellectual property licensing company. The RISC architecture can be licensed by anyone who wants to build a processor. This is deigned to lower the overall cost -- it is much less expensive for vendors to license an instruction set architecture than to develop their own. The goal is to make designing an ARM device practical for a group of vendors rather than a single supplier.

ARM development

Vendors have been sending mixed messages on whether the new ecosystem for ARM server chips will take shape.

AppliedMicro and HP are still on the ARM server bandwagon, but early supporters, such as Nvidia and Samsung AG, have moved away from the idea.

Red Hat and SUSE have ported their Linux operating systems to work on ARM processors. Oracle has developed versions of Java for the chips, and Hadoop runs on ARM architectures as well. In terms of applications, users can write their own software. Large technology firms such as Google and Facebook have such capabilities, but it's too resource-intensive a task for most enterprises.

About the author:
Paul Korzeniowski is a freelance writer who specializes in data center issues. He has been writing about technology for two decades, is based in Sudbury, MA and can be reached at paulkorzen@aol.com.

Next Steps

Can ARMs servers compete with traditional servers?

Answers to common questions about 64-bit ARM processors

What ARM means for tablets and desktops

This was last published in April 2015

Dig Deeper on Server hardware strategy

PRO+

Content

Find more PRO+ content and other member only offers, here.

Start the conversation

Send me notifications when other members comment.

By submitting you agree to receive email from TechTarget and its partners. If you reside outside of the United States, you consent to having your personal data transferred to and processed in the United States. Privacy

Please create a username to comment.

-ADS BY GOOGLE

SearchWindowsServer

SearchEnterpriseLinux

SearchServerVirtualization

SearchCloudComputing

Close