Examining Solid State Drives as Part of Modern Storage Stacks and Large-Scale Enterprise Storage Systems
Bibliographic record
Abstract
System reliability is arguably one of the most important aspects of a storage system and as such, a large body of work exists on the topic of storage device reliability. Considering that more data is being stored on solid state drives (SSDs), we study the reliability characteristics of NAND-based SSDs deployed in enterprise storage systems, including several factors that were not studied before. The findings of our study contradict common expectations, highlighting the importance of constantly evaluating the reliability of storage systems as new technologies and devices emerge. Furthermore, as we increasingly rely on SSDs for data storage, it is also important to understand what their operational characteristics look like in the field. Compared to HDDs, the performance and expected lifespan of an SSD are affected by operational characteristics in fundamentally different ways. For instance, SSDs require background work, such as garbage collection, which generates write amplification, thereby affecting both a drive's performance and lifespan. We present the first large-scale field study of key operational characteristics of SSDs in enterprise storage systems, while focusing on the drives' write rates, level of write amplification and how it is affected by various factors, along with the effectiveness of wear leveling. Next, we study the reliability characteristics of modern file systems, running on top of flash-based SSDs. Given the growing adoption of SSDs as a form of secondary storage medium, storage reliability also depends on the ability of higher levels within the storage stack to handle any errors these devices might generate. The results of our analysis indicate that file systems frequently cannot be mounted or even repaired. Based on our findings, we file several bug reports and also recommend several design guidelines tailored for file systems running on top of SSDs. In the last part of this thesis, we focus on system configuration, as the overall performance and reliability of a storage system are both affected by its underlying configuration. We present a detailed study on system configurations in production environments, while focusing on the characteristics of RAID groups, along with the number of total spare SSDs available. Finally, considering that systems need to adapt to increasing capacity needs during their lifetime, we additionally present the frequency and characteristics of system expansions and reconfigurations, as observed for real-world production systems.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; both teacher heads agree on what is shown here.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".