An open source server motherboard specification designed specifically for financial services organizations has been released by the Open Compute Project and AMD at the Open Compute Summit today.

The specification, formerly known as Roadrunner, was started on the back of a napkin at a café in October 2011 at the New York Open Compute event, according to Bob Ogrey, engineering fellow and cloud technical evangelist at AMD. Today, the motherboard specification is known as AMD Open 3.0. It is designed to provide gains in computing flexibility, efficiency and operating cost by simplifying motherboard design with a single base product to address multiple enterprise workloads, including high-performance computing, cloud infrastructure and storage. Fidelity and Goldman Sachs are currently evaluating AMD Open 3.0. Don Duet, managing director and global co-chief operating officer of the technology division at Goldman Sachs, is on the board of directors for the Open Compute Project.

“Along with Fidelity and Goldman Sachs, we have had 12 other companies review the specification,” Ogrey said during an exclusive interview with Wall Street & Technology. “The specs are available to anybody and anyone can use it. This specification is a huge win for lowering technology costs, which is a huge issue on the east coast” where many financial services organizations are headquartered, Ogrey adds. “Financial services has a variety of needs. They need high perfrrmance computing for systems that are speed sensitive, such as simulations” or trading. “There are also varying needs for storage,” both long and short term storage, a well as for cloud infrastructure, Ogrey continues. With AMD Open 3.0, “they have a standard platform and they can apply it to their different needs. It will simplify server management and it will reduce the number of different server management infrastructures. The IT costs will go way down.”

[For more on the Open Compute Project, read Open Compute Project Aims to Bring Open Source Standards to Data Center Technology.]

Today’s servers are designed with a “one size fits most” approach incorporating many features and capabilities that inefficiently utilize space and power, increasing cost. Mega data centers, such as those run by Google or Facebook, have engineers developing optimized platforms with the minimum set of components for specific workloads, but enterprise users do not usually build their own servers. As a result, many “one size fits all” servers running in financial services data centers waste energy and increase costs. With AMD Open 3.0, the result is a solution with a tailored combination of power, space and cost. The AMD Open 3.0 platform is designed to easily enable IT professionals to “right size” the server to meet specific compute requirements, according to AMD.

“I am a platform guy,” Ogrey said. “I believe this is the beginning of change in the way platform and servers are deployed. It provides correct features, scale, simplifies data center management and reduces cost and OPEX [operational expenditure]. This will be the future of computing service environments going forward. Everyone’s requirements are different, but having a flexible and modular approach is important” for today’s enterprise users, Ogrey added.

“This is a realization of the Open Compute Project’s mission of ‘hacking conventional computing infrastructure,'” said Frank Frankovsky, Chairman of the Open Compute Foundation and VP of Hardware Design and Supply Chain at Facebook, in a release. “What’s really exciting for me here is the way the Open Compute Project inspired AMD and specific consumers to collaboratively bring our ‘vanity-free’ design philosophy to a motherboard that suited their exact needs.”

Technical Specifications

According to AMD, Open 3.0 is powered by the AMD Opteron 6300 Series processor and can be installed in all standard 19″ rack environments without modification. The motherboard is a 16″ x 16.5″ board designed to fit into 1U, 2U or 3U rack height servers. It features two AMD Opteron 6300 Series processors, each with 12 memory sockets (4 channels with 3 DIMMs each), 6 Serial ATA (SATA) connections per board, a 1 dual channel gigabit Ethernet NIC with integrated management, up to 4 PCI Express expansion slots, a mezzanine connector for custom module solutions, 2 serial ports and 2 USB ports. Specific PCI Express card support is dependent on usage case and chassis height.

There are also three recommended system configurations for high-performance computing (HPC), general purpose and storage.

HPC

• 1 DIMM per channel (U/RDDR3 1600MHz, 1866 MHz Stretch Goal)
• Fits into the 1U chassis
• Cooling and Power for SE 140W parts
• 1 SR5670 tunnel
• Supports 6 SATA drives natively off of the “Southbridge”
• Supports up to ten 2.5″ total drives with add-in card (Full details in section 7.2)
• Supports up to 2 low-profile PCIe cards or 1
• Single 1Gb on-board controller via BCM5725
• 10Gb solution via add-in mezzanine card

General Purpose

• 3 DIMM per channel (Up to 1600 MHz support)
• Fits into 2U chassis
• Cooling and Power for SE 140W Parts
• Support for twenty five 2.5″ SATA/SAS drives
• Supports up to 1 standard-height and 1 low-profile PCIe cards (2 cards total)
• Single 1Gb on-board controller via BCM5725
• 10Gb solution via add-in mezzanine card

Storage

• 3 DIMM per channel (Up to 1600 MHz support)
• Fits into 3U chassis
• Cooling and Power for SE 140W Parts
• Support for thirty five 2.5″ SATA /SAS drives
• Supports up to 4 full-height, short PCIe cards
• 10Gb solution via add-in mezzanine card