The ultra low-latency SSD

The Samsung NVMe 983 ZET SSD sustains high-performance computing applications at incredible speeds. Boosting performance and QoS with low-latency V-NAND technology, the 983 ZET delivers high throughput for state-of-the-art data analytics and processing, ideal for taking business to the next level.

* The NVM Express® design mark is a registered trademark of NVM Express, Inc.
.

Engineered for fast responsiveness

Bridging the gap between system memory and storage, the 983 ZET is powered by cutting-edge low-latency V-NAND technology to retrieve data quickly when needed. Its impressive Quality of Service (QoS) reads and writes up to 0.03 ms, with a minimal latency of up to 20 μs.*

* The random read/write latency is measured by using FIO in CentOS7.0 and 4KB transfer size with queue depth 1 on a random workload of sustained state.

QoS is measured using FIO (99.99 %) with queue depth 1 on 4 KB random read and write.

Performance for complex processing

Experience a new level of performance. The NVMe™ interface empowers the 983 ZET to reach sequential read/write speeds up to 3,400/3,000 MB/s and random read/write speeds up to 750K/55K IOPS, for the seamless processing of even complex computations.*

* Actual performance may vary depending on use conditions and environment.

1) Sequential performance was measured by using FIO in CentOS7.0 with queue depth 32 by 1 worker.

2) Random performance was measured by using FIO in CentOS7.0 with queue depth 32 by 4 workers.

Secure critical data

The 983 ZET features end-to-end data protection to ensure consistency over the entire data transfer path and power loss protection to prevent data corruption in case of power failure.

.

Samsung Quality and Reliability

Experience the superior SSD quality and reliability of in-house production using Samsung built components. Empower your 24/7 business to run faster and more efficiently with world-class dependability with a limited 5-year warranty or up to 10 DWPD.

.

SPECIFICATIONS

Form Factor

HHHL

Capacity

480 GB, 960 GB

Sequential Read Speed

Max. 3,400 MB/s

Sequential Write Speed

Max. 3,000 MB/s

Show More Specs layer open

MORE SPECS

Model Code (Capacity)1)

  • MZ-PZA480BW (480 GB)

  • MZ-PZA960BW (960 GB)

General Feature

  • APPLICATION

    Data Center

  • FORM FACTOR

    HHHL

  • INTERFACE

    PCI Express Gen3 x4, NVMe 1.2

  • DIMENSION (WxHxD)

    Max. 167.7 x 69.9 x 18.8 (mm)

  • WEIGHT

    Max. 330g

  • NAND TYPE

    Samsung Low Latency V-NAND

Special Feature

  • ENCRYPTION SUPPORT

    AES 256-bit

Performance2)

  • SEQUENTIAL READ

    Up to 3,400 MB/s

  • SEQUENTIAL WRITE

    Up to 3,000 MB/s

  • RANDOM READ (4KB, QD32)

    Up to 750,000 IOPS

  • RANDOM WRITE (4KB, QD32)

    Up to 60,000 IOPS:
    MZ-PZA480BW (480 GB)


    Up to 75,000 IOPS:
    MZ-PZA960BW (960 GB)

  • QoS READ (99.99%, 4KB, QD1)

    Up to 0.03 ms

  • QoS WRITE (99.99%, 4KB, QD1)

    Up to 0.03 ms

Power Consumption3)

  • ACTIVE READ

    Typ. Up to 8.5 W

  • ACTIVE WRITE

    Typ. Up to 9.0 W

  • IDLE

    Up to 5.5 W

Environment

  • ALLOWABLE VOLTAGE

    12.0V ± 10%

  • RELIABILITY (MTBF)4)4)

    2.0 Million Hours

  • UBER5)5)

    1 sector per 1017 bits read

  • OPERATING TEMPERATURE

    0 - 55 ℃

  • SHOCK

    1500G, duration 0.5 ms, Half Sine Wave

Warranty

  • MZ-PZA480BW (480 GB)

    5 Year or 8.5 DWPD

  • MZ-PZA960BW (960 GB)

    5 Year or 10 DWPD

  • 1) 1 GB = 1 Billion bytes by IDEMA. Actual usable capacity may be less (due to formatting, partitioning, operating system, applications or otherwise).
  • 2) Actual performance may vary depending on use conditions and environment.
    1) Sequential performance was measured by using FIO in CentOS7.0 with queue depth 32 by 1 worker. 1MB/sec = 1,000,000 bytes/sec was used in sequential performance.
    2) Random performance was measured by using FIO in CentOS7.0 with queue depth 32 by 4 workers. Measurements were performed on a full Logical Block Address (LBA) span of the drive in sustained state.
  • 3) Actual power consumption may vary depending on system hardware & configuration. Active power is measured using IOMeter2006 on Windows Server 2012 R2.
  • 4) MTBF is Mean Time Between Failure. By definition, Mean Time between Failures (MTBF) is the estimated time between failures occurring during SSD operation.
  • 5) Uncorrectable Bit Error Rate (UBER) is a metric for the rate of occurrence of data errors, equal to the number of data errors per bits read as specified in the JESD218 document of JEDEC standard.
    For enterprise applications, JEDEC recommends that UBER should be below 10-16.

DOWNLOAD FILES