"There's a new class of technologies poised to shake up the Server market - the Microserver, sometimes called the ARM Server. But the Microserver will do more than shake up the X86 server market. It is going to shake up the software market as well." That's the message software and technology executive Chris Piedmonte, CEO and founder of Austin, TX-based Suvola Corporation delivered as we discussed the topic over coffee earlier this week. As I wrote in "ARM Server Microservers Seek to Transform Cloud, Big Data", ARM Server represents an entirely new generation of highly parallel server-on-a-chip ("SoC") computing that offers compelling benefits for many enterprise-class applications. And some analysts, including Oppenheimer's Equity Research group predict that these new ARM Servers will take a 20% bite out of that $45 Billion traditional Server market by 2016. That's big news. Chris' 30-year background in enterprise application software and highly scalable parallel systems gives him a unique and valuable perspective on this. He's well- known for being the principal inventor of a revolutionary new mathematical data management technology - and from 2005 to 2012 was the cofounder and CTO of a company focused on employing massively parallel systems to implement this data processing technology. Microserver Predictions Chris' enthusiasm about the technology is evident, "I first became aware of Microservers when researching new technologies for accelerating data management systems." He continued to tell me how he quickly realized that these systems were more than just a densely-packaged parallel processing platform. "The technology included a high bandwidth network fabric, vastly reduced power consumption and integrated management technologies. With the ability to package dozens and dozens of servers into a single chassis, I realized that this technology could be the next step in enterprise server technology," he continued. Ask him about the current state of the Microserver market, and he's quite clear - it's early stage, but is going to ramp at a remarkable rate because ARM Servers are highly suitable for many mainstream enterprise computing needs. "Microservers will begin to take hold in the enterprise starting in 2014. These systems will be initially used as file and media servers, network infrastructure hardware such as web servers, load balancers, SSL encoders/decoders and such. They will then begin to be used as LAMP stack [Author's note: LAMP is a combination of free, open source software. The acronym LAMP refers to the first letters of Linux (operating system), Apache HTTP Server, MySQL (database software), and PHP, Perl or Python (programming languages) - principal components to build a viable general purpose web server] implementations for SaaS offerings and eventually for full enterprise transactional and big data systems. Cloud providers will also embrace the technology as a means of providing Cloud services for less money due to the vastly lower power requirements and dense packaging. If the analysts have it right, 20 percent of the enterprise server technology will be Microserver based in the next several years. Opportunity for Software Vendors "This platform needs enterprise software specifically designed for this platform. The benefits that the server technology will bring to the enterprise are clear, but the current state of the industry doesn't have a large group of companies providing software ... in fact, it's difficult to find any enterprise software for the ARM-based Microserver platforms. A large, untapped market is being created. That's why I founded Suvola - to help speed up the adoption of this next generation technology by providing the enterprise software required to make complete solutions based on this server technology." Mr. Piedmonte has high hopes for the ARM Server market - and some pretty big ambitions for his company - "Suvola intends to move quickly to become the leading enterprise software company delivering products for the ARM-based Microserver market. As the market is expected to grow to over $10B in less than four years, Suvola will be there to help create it. Suvola has been developing relationships with the key IP licensees like Applied Micro and Imagination Technologies, SoC chip providers like Calxeda and ARM, chassis companies like Boston Limited, Penguin Computing, AAEON/Asus, MBX Systems and others. We are working with all these companies to help shape and guide the technology providers to create great platforms for enterprise software." The Verdict Chris tells a compelling story. People far smarter than I am see the promise in this emerging market. One such person is Gartner Group co-founder David Stein, who commented "the advent of Microserver systems incorporating hundreds to thousands of processors well may revolutionize enterprise computing". Various research groups have estimated the ARM Server market to be between 10% (HIS iSuppli Research) and 20% of the Server market by 2016. Already this year, Microservers are on target to hit 290,000 units versus 88,000 last year - a stunning 230% year-on-year growth rate. Given their diminutive size (you can fit dozens quad-core servers into the space used by a single traditional server), their "green computing" appeal (a Calxeda ECX-1000 quad-core Server with build-in Ethernet, SATA Controllers and 80 Gigabit Interconnect Switch uses only about 5 Watts of power at full power), and promises of better reliability at a reduced total cost of ownership - I think Mr. Piedmonte's enthusiasm is warranted. Challenges Facing the ARM Server Market This discussion would not be complete without acknowledging some of the challenges facing the Microserver. When I ask Chris about why Microservers aren't more prevalent, given their compelling advantages, he drives home the point that "impressive hardware isn't enough", and then elaborates on this point - "for this market to take off, we need solutions based on the technology, not just servers with operating systems and Java compatibility." He then continues to identify what he sees as the root cause of this: "the big software vendors aren't paying attention to this space right now - there aren't enough servers being shipped to merit the investment. And very few ARM developers are experienced in enterprise-class computing - they're focused on platforms like tablets and cell phones, and that's an entirely different mindset." As Dave Stein notes with respect to ARM Server software, "It remains to be seen how drastic will be the hardware and software architectural changes needed to accommodate the transition from minimal multiprocessing to massively-parallel multiprocessing, but it's a good bet they will be substantial." It's clear that neither the traditional set of ARM developers nor the traditional X86 software developers have what it takes to exploit the massive parallel capabilities of the ARM Server platform - building and optimizing software for scalability in massively parallel systems requires a mindset and skillset that few traditional software developers or architects in either of those camps possess. The opportunity for software (and hardware) vendors in the Microserver market is tremendous, and the challenges are significant. But if anyone can make this happen, it'll be folks like Chris and his team at Suvola. Source: HOLLIS TIBBETTS, Sys-Con Media, AUGUST 25, 2013, www.sys-con.com/.../2762446
↧