The SM10000 High Density, Low Power Server

Download the PDF

The SM10000 Uses 1/4 the Power and Takes 1/4 the Space of Today’s Best in Class Volume Servers and Requires No Changes to Software

The SeaMicro SM10000 is the First Server Purpose Built for Scale Out Workloads

Designed to replace 40 1 RU Dual Socket Quad Core servers, the SM10000 integrates 512 Intel Atom low power processors, top of rack Ethernet switching, server management, and application load balancing in a single 10 RU “plug and play” standards-based server. The SM10000 uses 1/4 the power and takes 1/4 the space of the best in class volume server without requiring any modifications to existing software.

System Highlights:

• Dramatic reductions in TCO: The SM10000 uses 1/4 the power and takes 1/4 the space of the best in class volume server

• 512 1.6 GHz Intel Atom CPUs in 10 RU (2,048 CPUs per rack)

• Drop in adoption: Runs off the shelf OS and applications

• Reduces capital expense by consolidating layers of expensive networking infrastructure

• Provides the unique ability to guarantee performance and security for cloud deployments

• The first server architecture to support all CPU instruction sets

The Industry's First “Rack in a Box”

As the industry's first Rack in a Box™, the SM10000 server brings together in a single system the computational capability of forty dual socket quad core 1 RU servers, redundant Ethernet rack switches, terminal servers and a load balancer. The SM10000 eliminates both the cost of discreet networking devices and the operational complexity of deploying and managing them.

Redundancy and Reliability

The SeaMicro system implements redundancy in both hardware and software. At the hardware level, all major subsystems are redundant and hot swappable, including compute cards, disk, network interface cards, power supplies, and fans. At the software layer, customers can configure the system to run active/standby software on two separate management cards. In the event of a failure, standby software will assume the responsibility of managing the system without any manual intervention.  SeaMicro software also manages the internal Terabit fabric and has the intelligence to configure the hardware to route traffic around failure using multiple alternative fabric paths (path redundancy).

Seamicro’s modular management software provides process isolation and modularity with each major process operating in its own address space – thereby increasing system availability and reliability.

Benefits

Dramatic reduction in TCO through 75 percent less power usage: Power is the single largest operating expense in the data center center, often in excess of 30 percent of total OP EX. Reports from Google show that if current power trends continue, the cost of energy consumed by a server during its lifetime could surpass the initial purchase cost. (Power Provisioning for a Warehouse-sized Computer) The SM10000 reduces power consumption and its associated costs by 75 percent in comparison to the best in class servers available today.

Massive density lowers TCO by reducing server space requirements by 75 percent: The second leading operating expense in the data center is space. Through its revolutionary architecture, SeaMicro is able to pack 512 CPUs in a 10 RU system (2048 CPUs in a rack). This industry leading density reduces space and the associated costs of space by 75 percent.

Consolidates functionality reducing CAP EX: The SM10000 consolidates 512 Atom CPUs, the top of rack switch, the terminal server and the load balancer into a single 10 RU system. This dramatically reduces Capital Expense by eliminating the need for expensive discreet devices.

Guarantees secure performance in Cloud Deployments: One of the powerful differentiators that SeaMicro brings to cloud and managed hosting is in the SM10000's ability to secure traffic and to guarantee performance.   In today’s cloud environment CPUs are shared across multiple users, which can lead to contention and uncertain performance as well as questionable security. SeaMicro’s system provides dedicated, right-sized compute units eliminating the sharing of CPU resources. Further, with the unique vFabric architecture, Ethernet interfaces and disk interfaces can be securely tied to a particular compute unit enabling the cloud provider to guarantee performance and security.

Flexibility in the ratio of compute to IO: The SeaMicro architecture allows customers to choose the ratio of compute to I/O bandwidth allowing systems to be tailored for specific applications. For example, a Web server with backend data store could be configured with zero disk space, while a search system requiring high disk capacity and IOPS could be configured with dedicated high capacity disks and SSDs.

Flexibility – Dynamically modifies the ratio of compute to storage: Any CPU can be configured with multiple virtual disks without requiring any hardware/infrastructure changes. A virtual disk can also be shared across multiple CPUs, providing a large shared data cache amongst 512 CPUs. Sharing of a virtual disk enables users to store/update common data, such as operating systems, application software, and data cache once for an entire system.

Accelerated Deployment: Tight packaging of CPUs in a 10 RU appliance allows for simple installation. The steps needed to install a SeaMicro system with more than 800GHz of compute are as follows: 1) Install the SeaMicro hardware into the rack, 2) Connect up to 4 power supply cables to a power supply source, 3) Connect uplink cables from the SeaMicro appliance to the core switch, 4) Connect the management Ethernet cable to a management switch, and 5) Configure the management boot parameters for all servers just one time using the SeaMicro management user interface.

Simplified Operations: All of the 512 CPUs in a SeaMicro SM10000 can be managed remotely using SeaMicro’s redundant management infrastructure. CPU reset and power on/off, installation of new operating system and application software, dynamic modification of load balancer capacity, and system performance monitoring and troubleshooting can be done using a single management API. The management API is built to be integrated into existing operational service systems with minimal effort.