Samsung announced it has begun mass production of 25GB and 50GB solid state drives, these new models feature faster single-level-cell (SLC) flash memory that delivers about twice the random write performance of the firm's standard 32GB and 64GB SLC SSDs. Samsung says the drives have a sequential read speeds of 100MB/s and write speeds of 80MB/s. The drives are marketed for performance-optimized server applications such as video on demand, web serving and online transaction processing.
Designed specifically for the enterprise market, the new Samsung single-level-cell (SLC) SSDs have about twice the random write performance of its standard 32GB and 64GB SLC SSDs.
Delivering the “greenest” storage solution for data centers today, Samsung’s enterprises SSDs use 1.25 watts of power in active mode and 0.3 watts in idle mode. They consume less than 25 percent of the power of a 2.5-inch 15K SAS HDD, and provide the lowest possible heat load on a data center air conditioning system. In recent ‘green computing’ campaigns, servers with high-rpm HDD solutions have come under scrutiny for heavy power consumption.
“Our SSDs give IT managers the best in high-performance, high-endurance storage for servers, with markedly less energy consumption,” said Jim Elliott, vice president, memory marketing, Samsung Semiconductor, Inc. “Now being considered by virtually all major PC OEMs, the proven technology of enterprise SSDs provides a compelling combination of price, performance and longevity for many medium-sized businesses as well as large corporations,” he added.
Available in a 2.5-inch interface, Samsung’s Enterprise SSDs provide significantly higher performance per watt than conventional 10K and 15K rpm HDDs. The drives process 100 times the number of IOPS per watt as a typical 15K 2.5-inch SAS HDD and five times the number of IOPS per dollar as a 10K 2.5-inch SCSI HDD.
The new performance-optimized Samsung Enterprise SSDs can process IOPS (input/output per second) up to 10 times faster than the fastest 15K rpm SAS HDD available for transactional data workloads, with the precise amount of performance improvement depending upon the specific application.