benedetti68 - Fotolia
If you adopt a blade server architecture, you may grapple with a wide range of storage options and alternatives. You should first consider if you want to include disks on each blade.
When storage is included per-blade, you can distribute storage across each blade in the chassis. Most modern blade servers support at least two disk products. For example, a Dell PowerEdge M640 for M1000e can hold up to two physical drives per blade, which supports up to 3.84 TB. By comparison, an HPE ProLiant BL660c Gen9 E5-4610v4 server ships with no drives installed, but supports up to four small form factor hot-plug drives per blade.
Integrated blade storage is a good choice when enterprise workloads demand the shortest possible latency. Because blades can support up to two disks, it is possible to employ RAID -- such as RAID 1 mirroring -- on the blade for storage redundancy.
Maintaining a blade server architecture
Storage challenges occur when blades fail and need replacement. One of the basic premises of blades is that their modular design eases replacement -- pull one blade out and put another in its place. But if there is unique data stored on disks on the blade, you must consider how to manage, move or replicate that disk data to the replacement blade.
Another option is to concentrate storage disks on dedicated modules or blades. This hardware is an option for centralized or shared data. For example, locating a database on a storage blade allows workloads running on different compute blades to access the common data store. Storage blades are also worthwhile when creating a software-defined storage infrastructure for platforms such as private clouds.
It's important to note that the choice between local and centralized storage is not mutually exclusive, and the two can coexist to support a multitude of enterprise workloads.
Your second consideration is where to locate disks on blades. You can populate the blades with a wide range of disk products. Most blades can support conventional, high-performance serial-attached SCSI (SAS) HDDs.
The bulk of the drive support comes from solid-state drive (SSD) products using conventional Serial Advanced Technology Attachment or SAS interfaces, such as 400 GB SSD SAS drives. The ultimate choice of disk type, interface and capacity depends on the needs and budget of the associated workloads. For example, high-performance workloads usually benefit from SSD devices.
Other informative classifications for an SSD blade server architecture are single-level cell, multi-level cell or triple-level cell. These designations refer to the construction and durability of the non-volatile memory devices used to construct the SSD.
Some SSDs are optimized for read/write cycles and can match the intended workload. For example, if a workload requires a high level of storage read performance, you can choose SSDs with read optimization to further enhance the storage performance.
Dig Deeper on Server hardware strategy
Related Q&A from Stephen J. Bigelow
Regression tests and UAT ensure software quality and both require a sizeable investment. Learn when and how to perform each one, and some tips to get... Continue Reading
Learn the meaning of functional vs. nonfunctional requirements in software engineering, with helpful examples. Then, see how to write both and build ... Continue Reading
Just because software passes functional tests doesn't mean it works. Dig into stress, load, endurance and other performance tests, and their ... Continue Reading