HP Superdome 2 powers demanding, mission-critical environments for enterprises that require high availability and rich virtualization. It incorporates innovations, such as the fault-tolerant crossbar fabric and the Superdome 2 Analysis Engine.
Superdome 2 is based on the HP Converged Infrastructure strategy, which attacks IT sprawl with standards-based, modular architectures. This provides a common, bladed architecture, common components and unified management across all HP servers.
Superdome 2 uses an 18U rack and HP's c7000 enclosures. Expansion enclosures can be added to provide an additional 96 PCIe I/O slots if required. Its sx3000 chipset provides 128 MB level 4 cache per blade. As the memory controllers are inside the Itanium 2 processor 9300 series CPU socket, they provide double the bandwidth with less latency than earlier Superdome systems. Improvements in the hyper-threading have also boosted performance by up to a 25 percent. All together, this results in up to 816 GB/s of I/O bandwidth and 1248 GB/s of bandwidth. It scales up as high as 128 processor cores and 176 I/O.
To maintain reliability and availability, the Superdome 2 uses error correcting, self healing and diagnostics. In addition, the Superdome 2 Analysis Engine moves diagnostics into the firmware level. It can report failures, for example, when the OS cannot boot. nPartitions provide hardware isolation while virtual partitions (vPars) and HP Integrity Virtual Machines (HPVMs) further divide partitions for resource assignment. These servers run either Windows or HP-UX.
In summary, the Superdome 2 supports the mission-critical features of legacy Superdome
Systems while integrating the many features of HP c-Class BladeSystems. With its high-performance sx3000 chipset, it has enhanced the reliability, availability and serviceability (RAS) capabilities.