Dell's Scrappy 'Copper' ARM-Based Server Initiatives
Jun 5, 2012 5:00 AM PT
Dell's recent launch of its "Copper" initiatives and ecosystem, which aim to develop commercial ARM-based servers, proves one point: IT industry watchers love a fight, especially when it pits a youthful, upstart technology against an established heavyweight vendor.
There are numerous others to choose from, but the Microsoft vs. Linux slugfest of the late 1990s is a case in point. Plus, that match-up is an interesting example of how often initial expectations of such events are essentially incorrect and change substantially over time.
When Linux first arose, open source aficionados suggested, often in vehement terms, that the hard-working OS with the cute little penguin mascot would eventually grow mighty enough to gobble up Windows' PC market position like so many kippered herrings. Despite their grandiose strategizing, that never happened and most likely never will.
However, with the help of some very interested big buddies, including Dell, IBM and HP, Linux prospered in the data center, providing businesses viable, affordable alternatives to -- and bulwarks against -- Microsoft's Windows Server. In addition, the open source development model sparked new models of collaboration -- some of which inspired what we now call "social networking." Plus, today's fully mature Linux ecosystem is driving numerous new innovations, including cloud computing and Big Data analysis.
What does any of this have to do with Dell's "Copper"? Just as many once said Linux could defeat Microsoft, some today are suggesting that ARM is capable of playing David to the Goliath of x86-based systems. How realistic is this? About as likely as a penguin waltzing into Redmond and saying the town ain't big enough for it and Steve Ballmer.
I'm not knocking ARM. It's a great technology with scads of market wins. In fact, ARM today is more mature and successful than Linux was at the time loyalists were prepping it for its prime-time prize fight.
While ARM-based servers are certainly intriguing for certain kinds of workloads and IT infrastructures, though, ranking them out of their class or pitting them directly against Microsoft is likely to result in ARM being twisted right out of its socket.
Which is why I'm heartened to see how Dell is approaching this issue: leveraging ARM's proven power efficiency and system footprint features to position it as a platform for Web front-end applications (LAMP stack -- Linux, Apache, MySQL and PHP), lightweight OSes and distributed applications.
In fact, ARM support is available today or coming from LAMP, Ubuntu, Fedora, OpenStack, Oracle Java, KVM and Hadoop. Hadoop was highlighted in Dell's Copper announcement for a very good reason: Because of the likely interest in ARM/Hadoop solutions among the company's myriad Data Center Systems (DCS or hyperscale) customers -- including four of the five largest search engines, four of the five largest computing clouds, three of the four top social media sites, and 13 of the world's 15 most-visited websites. That not only suggests a potentially lucrative commercial market, but also highlights the degree to which Dell's Copper efforts are being driven by customers.
That's a particularly interesting point. Dell has seldom been the first mover in emerging markets, preferring to bide its time until there are financial incentives to pursue. Yet the success of Dell's DCS solutions -- IDC has placed the company No. 1 in this market during 9 of the past ten quarters -- suggests its moves here are worth watching.
In fact, cultivating relationships with clients and partners by providing Copper seed servers and investing in additional engineering and IP related to ARM-based systems should help Dell broaden its offerings and achieve its goal of being an end-to-end IT solutions provider.
In addition, despite the relatively low profile today of microservers -- the class of systems in which ARM CPUs are generally intended to reside -- many forecasts suggest they are in for robust near-term growth. Gartner says the segment will account for 15 percent (with 82 percent CAGR) of the overall server market by 2016, while Oppenheimer has aggressively targeted a 21 percent market share (with 95 percent CAGR) in the same time frame. In other words, the work Dell is doing today with DCS customers and ARM ecosystem partners, including the Texas Advanced Computing Center (TACC) and OS/application developers, could well lead to significant future opportunities.
Pennies From Heaven
In times of drought, it isn't unusual for thirsty communities to try cloud seeding -- dispersing copper iodide into clouds, which offers a sphere for moisture to condense around and often triggers precipitation. While today's data centers aren't exactly running dry, you could say that the initiatives related to Dell's Copper seed servers aim to condense customer interest around ARM-based servers in a similar way.
If those efforts succeed, Dell's DCS customers, including those focused on cloud computing, could open up their IT infrastructures and pocketbooks even more to the company.
Initially, Copper servers aren't likely to drive enough revenues to match the winnings from a championship prize fight. Over time, though, ARM-based systems and services should provide valuable experience and freshening showers of future profits for Dell.