How many TB can a SSD write?

SSDs (solid-state drives) differ significantly from traditional HDDs (hard disk drives) when it comes to write endurance. Whereas HDDs rely on physical spinning platters and read/write heads, SSDs use NAND flash memory chips to store data. This fundamental difference in technology leads to major advantages for SSDs like higher performance, lower latency, and better shock resistance. However, it also results in a finite lifespan for SSDs, as each NAND flash memory cell can only withstand a certain number of write/erase cycles before failing.

According to this source, SSDs are well-suited for read-intensive workloads like databases thanks to their performance profile. But write-intensive workloads will wear down an SSD faster than an HDD. Careful workload monitoring and optimization is required to maximize SSD lifespan.

In the sections below, we’ll explore SSD endurance metrics, typical endurance figures, real-world endurance, and various technologies that aim to improve write endurance in SSDs.

What Determines SSD Endurance?

SSD endurance refers to how much data can be written to an SSD before it fails. There are several key factors that impact SSD endurance:

Type of NAND flash – SLC NAND can withstand 100,000 writes per cell, while MLC NAND is rated for 10,000 and TLC for 1,000. SLC SSDs have higher endurance than MLC and TLC 1.

Controller – The controller manages operations like wear leveling to distribute writes across all cells. More advanced controllers improve endurance 2.

Over-provisioning – SSDs reserve extra NAND capacity for replacing worn-out cells. More over-provisioning means higher endurance.

In general, enterprise SSDs designed for write-intensive workloads offer much higher endurance than consumer SSDs. But endurance ultimately depends on the combination of the factors above.

Endurance Metrics

SSD endurance is commonly measured in three main ways:

TBW (Terabytes Written) – This refers to the total amount of data that can be written to the drive before it can no longer reliably store data. It is measured in terabytes (TB) and indicates the overall endurance of the SSD.[1]

DWPD (Drive Writes Per Day) – This measures the number of times you could overwrite the drive’s entire capacity in a day. For example, a 1 DWPD drive with a 500GB capacity means you could write 500GB per day for the warranty period.[2]

P/E Cycles (Program/Erase Cycles) – This refers to the number of times each NAND flash memory cell in the SSD can be programmed and erased before wearing out. Higher quality cells have higher P/E cycle ratings.

TBW offers a simple overall endurance rating, while DWPD provides a normalized daily usage guideline. P/E cycles are more technical but indicate performance at the component level.

Typical Endurance Figures

SSD endurance is typically measured in total bytes written (TBW). This refers to the total amount of data that can be written to an SSD before it is likely to fail or lose significant performance. TBW ratings vary significantly between different classes of SSDs:

Consumer SSDs aimed at general desktop use often have TBW ratings between 150-600TB. Examples include the Crucial MX500 with 180 TBW or Samsung 870 EVO with 600 TBW. These consumer drives are designed for mainstream PC use like booting an OS, launching applications, and general productivity.[1]

Prosumer NVMe SSDs for enthusiasts step up the endurance with TBW ratings from 600-1800TB. The Corsair MP600 Pro XT has a 1400 TBW rating. While the Samsung 980 Pro offers 600TB. These are suitable for heavy gaming, creative workloads, and productivity.[2]

Enterprise SSDs are designed for 24/7 operation in data centers and servers. Models like the Samsung PM1735 and Intel DC P5510 offer 10-20 Drive Writes Per Day (DWPD) totaling 36,500-73,000 TBW over a 5 year period.

Real-World Endurance

In the real world, SSD endurance can vary significantly depending on different use cases and conditions. According to one analysis, consumer SSDs typically last between 3-5 years with normal everyday usage https://arstechnica.com/civis/threads/will-increased-ram-prolong-ssd-life.1190529/. However, heavy workloads like video editing or running database servers can wear out an SSD much faster.

One study looking at physics-based SSD reliability models found a wide variance in predicted lifespans, ranging from 1 to over 10 years under different conditions https://ieeexplore.ieee.org/document/6517673. Factors like write amplification, over-provisioning, and wear leveling all impact how evenly writes are distributed across flash cells.

In server environments, one analysis of Samsung datacenter SSDs showed they maintained over 60% of their endurance ratings after 3 years of intensive use https://www.anandtech.com/show/13322/samsung-launches-broad-range-of-datacenter-ssds. However, real-world endurance can be difficult to predict and varies based on individual workloads.

Improving Endurance

There are several techniques SSD manufacturers use to improve endurance ratings and prolong the lifespan of drives:

Over-provisioning reserves a portion of the NAND flash capacity solely for replacing worn-out cells, reducing write amplification and extending endurance. More over-provisioning means more spare area for the controller to utilize (Horizon Technology).

Caching writes in the DRAM buffer helps absorb write operations before they reach the NAND flash, reducing total writes to the cells. Larger DRAM caches improve endurance (ACM).

Hardware techniques like die-stacking NAND layers and using high-density 3D NAND technology allow SSDs to fit more storage capacity into a smaller space. This increases endurance by having more cells in parallel to distribute writes across (Tech4Gamers).

Overall, manufacturers leverage over-provisioning, caching algorithms, and hardware innovations to extend SSD lifespan and endurance.

Estimating Your Needs

When purchasing an SSD, it’s important to estimate your endurance needs based on your anticipated workload. This allows you to select a drive that can withstand its expected lifespan in your environment.

Endurance needs can be estimated by calculating your daily write workload and extrapolating that over the required useful life of the SSD. For example, if you write 84.935GB per day and need the SSD to last 3 years, you can use an online SSD endurance calculator to determine you’ll require around 1,100TBW for your workload.

Likewise, you can take the total capacity of your SSD and divide it by the drive’s DWPD rating to find out its approximate lifetime writes. So a 1TB drive with 1 DWPD could write 1TB x 365 days x 1 DWPD = 365TBW before wearing out.

When estimating endurance, also consider future growth in capacity needs and workload. An SSD sized appropriately today may fall short tomorrow if write demands increase substantially. Adding some headroom to your endurance calculations helps ensure your SSD will continue meeting performance and lifespan expectations.

Endurance Technologies

There are several technologies used in SSDs to improve endurance and lifespan:

One technology is SLC Caching, which uses faster and more durable SLC (single-level cell) NAND flash to boost write speeds and endurance. The SLC cache acts as a buffer to absorb a certain amount of writes before they reach the slower TLC or QLC NAND flash. This helps reduce write amplification and improves overall endurance (https://documents.westerndigital.com/content/dam/doc-library/en_us/assets/public/western-digital/collateral/tech-brief/tech-brief-slc-flash.pdf)

LDPC (Low Density Parity Check) is an error correcting code used in SSDs to detect and recover from data corruption. By reducing the number of uncorrectable read errors, LDPC helps improve endurance and data retention over the lifespan of the SSD (https://www.atpinc.com/blog/what-is-ldpc-how-does-it-relate-to-ssd-endurance).

The type of NAND flash used also impacts endurance. SLC offers the highest endurance, up to 100,000 P/E cycles, while MLC offers 3,000-10,000 cycles, TLC offers 1,000-3,000 cycles, and QLC offers only 100-1,000 cycles. However, higher density NAND like TLC and QLC are used to lower costs, while technologies like SLC caching help compensate for lower inherent endurance (https://www.westerndigital.com/en-asia/solutions/datacenter/data-center-ssd#0B12BF64AD1286095D819D8A).

Failure Models

SSDs can fail in several ways after exceeding their write endurance limits. Some of the most common failure modes include:

Total failure – the SSD stops working completely and can’t be detected or recovered. This tends to happen more often with older SSDs using Sandforce controllers (Source: https://www.reddit.com/r/DataHoarder/comments/4uezzb/what_ssd_failure_modes_have_you_observed/).

Performance degradation – write and read speeds will steadily drop as more NAND blocks wear out. The SSD may still work but will feel increasingly slow and unresponsive.

Bad blocks – portions of the NAND become unusable but the SSD remains functional by swapping in spare blocks. The available storage capacity decreases over time.

Corrupted data – with no bad blocks, stored data can still become unreadable or corrupted as cells wear out. Error-correction codes help avoid this.

Overall, endurance varies greatly between SSD models and workloads. But exceeding the write limits consistently will eventually cause irreversible failure or make the device unusable in practice.

Maximizing Lifespan

For power users with heavy workloads, there are some best practices to maximize the lifespan of your SSD:

  • Avoid completely filling up the drive – IDEAL utilization is 50-75% full. This allows the drive to properly distribute writes and manage bad blocks.

  • Enable the TRIM command if your OS supports it. TRIM frees up blocks no longer in use and improves performance.

  • Use the manufacturer’s SSD utility tool to monitor disk health and lifespan. Keep firmware up to date.

  • Maintain your system – defrag, clean temp files, close background apps. Reduce unnecessary writes.

  • Consider overprovisioning unallocated space to extend lifespan. But compare to cost of buying a larger drive.

  • Keep your SSD properly cooled for optimal performance. Heatsinks can help disperse heat.

  • For critical data, use enterprise SSDs designed for heavy workloads. Or use RAID mirroring as a backup.

With proper precautions, even heavy users can maximize lifespan and get full value from consumer SSDs before needing to replace them.