New ARM-based Moonshot servers from HP exemplify workload-specific computing

One of the developing trends in computing, relevant to both enterprise and service providers alike, is the notion of workload-specific or application-centric computing architectures. These architectures, optimised for specific workloads, promise improved efficiencies for running their targeted workloads, and by extension the services that they support.

Share

One of the developing trends in computing, relevant to both enterprise and service providers alike, is the notion of workload-specific or application-centric computing architectures. These architectures, optimised for specific workloads, promise improved efficiencies for running their targeted workloads, and by extension the services that they support.

Earlier this year we covered the basics of this concept in “Optimise Scalable Workload-Specific Infrastructure for Customer Experiences”, and this week HP has announced a pair of server cartridges for their Moonshot system that exemplify this concept, as well as being representative of the next wave of ARM products that will emerge during the remainder of 2014 and into 2015 to tilt once more at the x86 windmill that currently dominates the computing landscape.

Specifically, HP has announced the ProLiant m400 Server Cartridge (m400) and the ProLiant m800 Server Cartridge (m800), both ARM-based servers packaged as cartridges for the HP Moonshot system, which can hold up to 45 of these cartridges in its approximately 4U enclosure.

These servers are interesting from two perspectives – that they are both ARM-based products, one being the first tier-1 vendor offering of a 64-bit ARM CPU and that they are both being introduced with a specific workload target in mind for which they have been specifically optimised.

The m400 is based on Applied Micro Products STORM 64-bit ARM SOC, with 8 64-bit X-Gene I cores and up to 64 GB of memory and local SSD on each cartridge. The server is targeted at web caching  workloads.

Preloaded with a new Ubuntu Linux release, HP claims that this server will yield a 35% total cost advantage over competing server offerings for its targeted workload. We are unable to verify this specific claim, but look forward to more details from partners as their initial references sites at Sandia National Labs and University of Utah along with partners Cannonical and Nginx release more detailed performance data.

The m800 is an even more extreme embodiment of workload-specific computing. Each m800 cartridge contains 4 Texas Instruments KeyStone II SOCs, each consisting of 4 32-bit ARM cores and 8 DSPs. HP has specifically targeted this unit at real-time signal processing and what they describe as real-time data processing, epitomised by their marquee reference customer PayPal who described the use of the m800 at the most recent ISSC conference for real-time event processing of complex events in their online payments systems (see the PayPal video here).

TI itself lists an even more encompassing set of applications for the KeyStone series, including high performance computing, media processing, video conferencing, off-line image processing and analytics, gaming, security digital video recorders (DVR/NVR), virtual desktop infrastructure, medical imaging. HP offers these cartridges with Canonical Linux as well as software from Infochips and ENEA for real-time Telco development.

Taken as a whole, this announcement sends two signals about HP’s approach to the systems market – that ARM is still a potentially viable platform, and that workload-specific systems have an increasing place in an ecosystem that grows ever more complex as processing demands escalate. Forrester expects further announcements from the ARM ecosystem as well as more system vendors fielding workload-optimised devices over the next 12 months, powered by both ARM and x86 CPUs, with the possibility of activity surfacing from the OpenPower consortium as well.

For service architects both in enterprise settings as well as CSP environments, these are exciting developments, expanding the number of options for squeezing the largest number of transactions out of a service for the smallest number of dollars, albeit at the expense of having to evaluate these additional offerings against incumbent platforms.

Posted by Richard Fichera

"Recommended For You"

Intel stresses in-memory computing with 15-core server chip First ARM benchmarks - The race is on