Dell skunkworks brews ARM server future
Cortex A9. LAMP stack. Eye of newt
The ARM RISC processor owns the smartphone and handheld market that is starting to rival the traditional desktop and laptop PC in terms of functionality, and there is a very good chance that the chip will soon start making its way into the server racket.
Not from the bottom up, like the x86/x64 chips did, but from an oblique angle with hyperscale Web companies looking to cram more computing into less space while consuming less power.
At least, that is the thinking of Paul Prince, chief technology officer of Dell's Enterprise Products Group, which designs its general-purpose PowerEdge and semi-custom PowerEdge-C cloud boxes. Prince runs a team of several hundred engineers who in addition to creating those generic x64 servers also monkey around with different technologies.
Prince and his team have watched the ARM chip as it moves from handhelds into netbooks and other devices like routers and switches. The next logical place for ARM chips, once they get some support for virtualization and larger main memories, is in servers.
A lot depends on how the ecosystem develops around the impending "Osprey" Cortex-A9 MPCore, which scales to four cores, and the future "Eagle" Cortex-A15 multicore ARM processor, which will scales to a theoretical sixteen cores.
ARM Holdings doesn't make chips, but rather licenses designs, and last month the company unveiled the Eagle design and has the entire IT industry spoiling for a chip-slap. (If AMD had any sense, it would license the ARM design and scare the hell out of Intel.)
A year and a half ago, as it turns out, Dell's skunkworks in the Enterprise Products Group put together some baby servers using the Cortex-A8 processors (variants of which are in Apple iPhones and iPads as well as in Motorola Droids and many other devices) and using a 3-inch by 3-inch motherboard form factor.
"There are a lot of people dabbling with these ARM chips"
Dell got its BeagleBoard mobo, which is created by Circuitco Electronics and distributed by Digi-Key, slapped on a Linux-Apache-MySQL-PHP software stack, and then put them through the server paces.
The BeagleBoards support Windows CE, Linux, or Symbian operating systems, and Circuitco Electronics sells two single-board computers: one based on the Texas Instruments OMAP35X and the other based on an ARM Cortex A8.
These are very small and very inexpensive machines ($149 for the former and $179 for the latter.) The top-end board has a chip that runs at 1 GHz and that has 512 MB of main memory in addition to 256 MB of flash embedded on the board.
"There are a lot of people dabbling with these ARM chips," says Prince. Dell, in fact, is waiting to get its hands on some of the multicore Cortex-A9 MPCore chips, which will scale up to four cores on a single chip. "For a technology perspective, I am pretty bullish that there is an opportunity for a lot of use cases in servers starting with the A9 chips."
By the way, the ARM Cortex-A8 chips, which range in speed from 600 MHz to 1 GHz or higher, don't have error correction on their memory controllers and are therefore not perfectly suited to enterprise jobs; but if the platform is a cluster that has data protection and failover built into a cluster, then this doesn’t matter quite so much. (If they get a memory error, you reboot them. No big D so long as you have load balancers dishing out work to your cluster nodes.) The Cortex-A9 and Cortex-A15 chips do have ECC scrubbing on their memories, which makes them more suitable for servers.
Chip maker Marvell, which bought Intel's XScale ARM RISC business five years ago, is hoping to get its dual-core variant of the Cortex-A9 chip into the field by the end of the year. As El Reg has previously reported, a startup called Smooth-Stone has raised $48m in funding to design ARM-based servers.
In the summer, there were rumors Facebook was putting ARM-based servers into its data center in Oregon, and of course it would be interesting to see Apple deploy ARM-based servers in its massive North Carolina data center.
Prince and his colleagues at Dell are going to be putting Cortex-A9 skunkworks servers together as soon as chips and motherboards are ready, and will load them up with a LAMP stack to get the feeds and speeds of what these systems can do.
So when the next hyperscale data center looking for bespoke servers to radically cut back on power consumption comes knocking on Dell's door, the company will have another option beside Intel Atom and low-voltage Xeon, Opteron and low-voltage Athlon, and VIA Technologies Nano chips.
Dell is not expecting a lot of business for such machines, but it has to be ready to capture whatever there will be. That is the secret of the success of the Data Center Solutions unit of Dell, which has quietly become a dominant force in Dell's overall server shipments.
"ARM is moving the core forward with the Cortex-A15, and the ecosystem is, too," says Prince, adding that it the ecosystem is what will add the memory capacity, I/O features, and other functions to an ARM chip design that make it suitable for a server. "This feels like it is moving at the right pace."
Virtualization of memory
Some would argue that the ARM chips are not evolving fast enough to take on Intel and AMD in the server racket. As El Reg explained in August, the Cortex-A15 design includes a feature called large physical address extension, which will translate 32-bit virtual memory addresses to 40-bit or larger physical addresses.
This gives hypervisors and operating systems running on the Eagle machines some sleight of hand to address more physical memory than the 32-bit limit allows (at 4 GB). (Similar virtualization of memory was used in 16-bit and then 32-bit PC server chips as Intel moved from the desktop to the servers, so there is some precedence for this.)
ARM's own specs for the Cortex-A15 calls for "home and Web 2.0 servers" in quad-core configurations with the Eagle chips running at between 1.5 GHz and 2.5 GHz and with virtual machines atop the cores and support for more than 4 GB of main memory in the box.
Don't get the wrong idea. We are not going to see Dell launch a PowerEdge-A series lineup until ARM chips go mainstream in servers.
"I don't see ARM taking over the enterprise server market," says Prince. "But ARM could be a very important niche."
Ubuntu Linux and Android Linux already run on ARM chips, and so does Microsoft's Windows Embedded CE. There is no reason why the real server versions of any of these platforms - and others - can't be tweaked to run on ARM-based servers.
Lots of hyperscale customers control their own code-bases, and porting to a new architecture is no big deal for their techies. If the cost per unit of performance per watt numbers work out, it is cheaper to pay a few programmers to retune the apps for a new architecture than it is to build a new data center or three.
Dell did not provide access to its benchmark results on the LAMP stack running on the skunkworks Cortex-A8 machines, but you can see a performance test done by ARM Holdings that pits a netbook using a 1.6 GHz Atom processor against a prototype machine using a dual-core Cortex-A9 processors here.
See if you can tell if there is much of a difference between the two. ®