Frequently Asked Questions
The most-asked questions about computer storage, part 3.
What topics does Part 3 of the encyclopedia actually cover?
Part 3 focuses on advanced and enterprise-level storage architectures, including NVMe-over-Fabrics, CXL memory pooling, erasure-coded object storage, and tape-based archival systems. It also dives into storage virtualization layers, ZFS/Btrfs filesystem internals, and the transition from block-level to object-level data management.
Who are the main 'characters' or central technologies readers should know in this section?
The recurring pillars are NVMe (as the dominant high-performance interface), CXL (as the emerging memory-expansion fabric), and erasure coding (as the backbone of modern object stores like Ceph and MinIO). SAS and Fibre Channel appear as the enterprise workhorses that NVMe-oF is gradually displacing.
Where should a complete newcomer begin reading within Part 3?
Start with the 'NVMe vs. SATA: What Actually Changed' entry, since it bridges the consumer-level knowledge from Part 2 into the protocol-level thinking Part 3 demands. The 'Erasure Coding in Plain English' sidebar next to the object-storage chapter is also a recommended first read before tackling Ceph or MinIO articles.
What is the single most-searched entry in Part 3?
The 'TRIM, Garbage Collection, and Why Your SSD Slows Down' article consistently tops search queries, especially around the 70%-full threshold where write amplification becomes noticeable. Readers frequently cross-reference it with the ZFS ARC sizing guide in the same section.
What key factual detail does Part 3 emphasize that Parts 1 and 2 glossed over?
Part 3 stresses that storage latency is not a single number but a distribution: queue depth, I/O size, and the specific controller's command pipeline all shift the p99 versus p50 figures dramatically. The 'Latency Is a Liar' entry walks through why a spec-sheet '7 µs random read' can hide a 200 µs tail under mixed workloads.
Which 'notable moment' or breakthrough gets the most fan-encyclopedia attention in Part 3?
The 2019–2021 shift where hyperscalers moved primary workloads from spinning-disk SANs to NVMe flash arrays with erasure coding is treated as the section's central turning point. The 'Death of the Spinning Disk in the Data Center' timeline entry traces the economic crossover in detail.
How does Part 3 connect back to the consumer-level material in Parts 1 and 2?
Several entries explicitly map consumer concepts to their enterprise counterparts, such as how USB 3.x/4 link-layer retransmission mirrors Fibre Channel's credit-based flow control. The 'From USB-C to CXL: One Cable, Three Eras' entry is the most-cited bridge between the two halves of the encyclopedia.
What is the most debated or controversial topic among the fan community in Part 3?
Whether CXL 2.0/3.0 will truly replace traditional DIMM slots or merely supplement them sparks the longest comment threads. The 'CXL Skeptic's Checklist' entry, written as a counter-argument, is as popular as the optimistic 'CXL Will Reshape Server Memory' piece it responds to.
Are there any lesser-known 'hidden gem' entries worth finding?
The 'LTO-9 vs. LTO-10: A Tape Engineer's Field Notes' piece, contributed by a former IBM tape-line technician, is frequently cited as the most human and detailed archival-storage write-up in the entire encyclopedia. The 'Why ZFS Checksums Made Me Sleep Better' personal essay also draws strong reader praise.
What reading order do veteran fans recommend for Part 3?
The community consensus is: NVMe protocol basics → erasure coding fundamentals → object-storage architecture (Ceph/MinIO) → CXL and memory pooling → tape and optical archival as a capstone. Skipping the erasure-coding chapter before the object-storage entries is the most common mistake new readers make.