Today’s enterprise-level virtualization and database servers need to support massive level of scalability to run todays resource intensive OTLP, data warehousing and big data applications. However, it’s not just about scalability. In fact, while scalability is undoubtedly important it’s not as important as reliability and availability. The components needed to increase uptime and availability are even more important than the overall scalability of the system. Sure your server needs to handle your workloads but for mission critical workloads it also needs the uptime to consistently handle these workloads and deliver the high level of availability that your mission critical applications like SQL Server 2016 require.
The HPE Superdome X is definitely a highly scalable system. It can scale from 1-8 blades, 2-16 sockets, 8-288 cores and can support up to 24 TB of RAM. However, it’s also designed for extreme availability. At the hardware level, the Superdome X is designed to be highly serviceable and reliable. The key Superdome X components are redundant and hot-swappable. This includes components like power supplies, fans, and I/O switches. The Superdome X supports multiple hardware partitions, nPARs, where each partition can be electrically isolated for the other partitions. The Superdome V utilizes a high bandwidth Crossbar Fabric that provides a low latency between all the processors within each nPar. The Crossbar Fabric is completely fault tolerant and is able to survive a crossbar failure. It can automatically reroute data and recover immediately. This can avoid critical memory failures, which is essential for in-memory databases.
Beyond this, the HPE Superdome X provides a number of built-in intelligent management features that push its availability capabilities to the enterprise level. In Figure 1, you can see an overview off the Superdome X’s built-in management capabilities that enable enterprise-level reliability.
Figure 1 – HPE Superdome X’s Integrated On-board Management Capabilities
You manage the HPE Superdome X using the built-in Onboard Administrator (OA). The OA communicates with the HPE Integrated Lights Out (iLO) out-of-band management feature that’s built-into each blade. The OA provides a single management point to perform essential configuration and monitoring tasks for the compute enclosure, server blades, and I/O interconnects.
The HPE Superdome X uses a firmware first approach to identify and contain errors at the firmware level – before these errors can be surfaced to the OS. This helps prevent any possible data corruption.
The built-in Error Analysis Engine (EAE) continually analyzes the hardware for any possible faults. The EAE is embedded in the Superdome X’s hardware and firmware. The EAE can predict failures as well as initiate automatic recovery actions – all totally automated without any operator interaction. The EAE can also send notifications to IT administrators and HPE Insight Remote Support.
HPE Insight Remote Support connects to the Superdome X’s OA to remotely monitor for problems and to perform troubleshooting. It can connect to HPE support for automatic notification to of any system problem. The HPE Support Center portal enables your system administrators to track service events and support cases as well as view device configurations and reports provided by HPE Proactive Care services.
Sponsored by HPE and Microsoft
