Blade servers have become popular building blocks for enabling converged data centers. They offer the ability to reduce complexity while supporting higher resource densities in terms of compute capacity, memory and I/O for storage and networking. With each successive generation of computer processors, more CPUs or “cores” are packaged into a smaller physical space. In addition to all those cores, new server processors enjoy better performance, measured either in cycle times or the number of instructions they can process per second.
As a newer technology, blade servers help meet the growing demand for more processing capability, memory and I/O in the same or a smaller footprint. Available in various sizes and configurations, blade servers pave the way for aggregating multiple larger, traditional servers in a smaller, denser space. In addition to taking up less room in a data center, they help reduce management complexity and enable modular growth.
Blade servers remove some of the complexity commonly associated with servers because they virtually eliminate cabling for I/O and for expanding storage and networking connectivity. They offer significant flexibility in choosing different storage connectivity options, including shared or dedicated serial-attached SCSI (SAS), Fibre Channel, InfiniBand, and Gigabit and 10-Gigabit Ethernet. The LAN adapters in the blade servers, in addition to other forms of connectivity, also support shared storage over IP, including iSCSI and network-attached storage (NAS) — both Network File System (NFS) and Common Internet File System (CIFS) flavors.
With a blade server system, as greater compute capability is needed, organizations can simply install an additional blade into an open chassis slot. As one blade server chassis fills up, additional chassis can be installed in the same or adjacent cabinets.
In addition to modularity and flexible growth, blade servers support various CPU and memory configurations, as well as different I/O options. For example, IT staffs can configure some blades for compute-intensive video and business analytics processing, while configuring others with additional memory and I/O to support database, e-mail, Microsoft SharePoint and web applications.
Many applications, including databases and virtualization programs, require large amounts of fast memory. Blade servers support those large memory banks. Keep in mind that simply adding more memory is not enough to adequately run high-performance solutions. The memory also must be fast and low-latency to avoid introducing bottlenecks.
For more information on how blade servers fit into server virtualization and data consolidation, read the CDW•G white paper on data center convergence.