SSDs in servers deliver performance, whether measured in IOPS or GBps or latency, that is just plain impossible with disks. SSD capacity is now greater than that of performance disks. Price is the factor that still keeps many organizations away from deploying servers with SSD storage when the improved performance is not a significant benefit. The better reliability of SSD might be more important. Continue reading SSD vs HDD RAID in Servers and Storage
For years, ion has reported usable capacity of disk and SSD storage in GiB and TiB instead of GB and TB. Why? And what’s the difference? Basically, 1 GB = 1,000,000,000 bytes while 1 GiB = 1,073,741,824 bytes. That is about 7%! Look at ion‘s SR-71mach6 SpeedServer or PS StorageServer for examples of capacity reporting. You can learn more about the differences in the Continue reading Storage Capacity: TB vs TiB
Servers based on Intel® Xeon® Scalable Processors, the proccesor architecture known as “Skylake”, have arrived! ion is producing 1U and 2U rackmount servers along with pedestal/4U servers and high density half-U compute nodes. These servers feature much improved capability for NVMe SSDs and significantly higher memory bandwidth, along with options for more cores and higher clock speeds. Intel® Xeon® Scalable Processors include a number Continue reading ion P-series Servers with Intel Xeon Scalable Processors
Much of modern storage management focuses on efficient use and allocation of storage capacity. “Thin provisioning” is a primary mechanism for this, allocating just enough space to match each consumer‘s current needs, while promising more capacity when needed. Thin provisioning is an effective tool for allocation of storage capacity. When latency, bandwidth and IOPS are more important, thin provisioning makes these performance results nondeterministic. Continue reading Thin Provisioning: Nondeterministic Storage Feature #2
Every producer of storage software (including what is hidden inside of “hardware”) has spent years and many engineering hours developing their cache algorithms. Disks have cache; controllers have cache; subsystems have cache; operating systems have cache. The better the algorithm is, the better the ratio of cache hits to cache misses. Ultimately though, the decision of what to have in cache is a guess. Continue reading Cache: Nondeterministic Storage Feature #1
If you have spent any time with computer benchmarks, you have probably heard the expression “Lies, damned lies, and benchmarks!” Why do people feel that way about benchmarks? Well, it is because many benchmarks have been presented with little or no context about what that benchmark really meant. “What was the hardware configuration?” “Was it running special firmware?” “How was the software tuned and Continue reading A Few Thoughts on Benchmarks
The title of this series of posts is “Spinning Rust is Dead”. Well, actually it is “Spinning Rust is (almost) Dead”. Why, given all of the facts, do I say “almost“? Disks are really good at one thing: storing a LOT of data. So for backup, archival, or storing many terabytes of data that gets accessed only occasionally, disks are good. Data centers should Continue reading Spinning Rust is (almost) Dead, Part 5, Almost?
“But wait,” you say, “price and performance aren’t everything!” Really? You are correct. Reliability counts when your data is sitting on the drive. To start with, think about microscopic magnetic heads flying back and forth over platters spinning at 7,200 rpm to 15,000 rpm. Need I say more? We can compare manufacturer’s MTBF (Mean Time Between Failure) numbers, but nobody really knows how they Continue reading Spinning Rust is (almost) Dead, Part 4, SSD Power and Reliability