solid-state drive
Enable an AI agent to recognise a solid-state drive, assess its compatibility, data integrity and remaining serviceability, and choose justified maintenance, reuse or retirement actions.
Research draft, second pass
A second pass drafted this model: the structure a model of this thing needs, and what is known about it in the world. The line under this one says how the second half was obtained - researched against sources, or recalled without web access, in which case nothing here was read anywhere and every claim is a lead to verify. Unreviewed either way.
Researched by: Codex + Grok
Purpose and description
Enable an AI agent to recognise a solid-state drive, assess its compatibility, data integrity and remaining serviceability, and choose justified maintenance, reuse or retirement actions.
A solid-state drive is a non-volatile block-storage device that persists host data in semiconductor memory - almost always NAND flash, historically also 3D XPoint - and presents that data through a storage-host protocol (NVMe/PCIe, SATA/AHCI, or SAS) without rotating platters or a head-actuator assembly.
It can be Identify and inventory the physical drive and its exposed storage units.; Assess installation or migration compatibility against a specific host connection.; Read supported health logs and run bounded diagnostics appropriate to the drive's condition.; Provision supported storage settings or update firmware when prerequisites and authority are established.; Prioritise data preservation and replace or retire a drive using observed condition and workload requirements.; Perform an authorised, supported sanitisation operation and record its completion evidence and limitations..
Distinguishing features
Confirm that persistent storage uses solid-state media without rotating magnetic platters; a hybrid drive containing flash and platters does not pass this boundary test.
Confirm that the identified thing includes controller-managed, host-accessible storage rather than being a bare memory chip or volatile RAM module.
Separate physical-device identity from exposed namespaces, partitions and logical volumes; these can represent portions or combinations of devices rather than individual SSDs.
Use the declared device class, controller and intended storage role to distinguish an SSD from a memory card or USB flash drive; semiconductor storage or a USB connection alone does not settle the distinction.
Check storage protocol and controller identity in addition to physical shape; an M.2-shaped module is not sufficient evidence that the object is an SSD.
Scope
+ Device identity, storage technology and physical boundaries
+ Mechanical, electrical and host-protocol compatibility
+ Exposed capacity, addressable storage and supported device commands
+ Firmware, health evidence, endurance and failure symptoms
+ Device-level protection, sanitisation, servicing and disposition
- Filesystem organisation, files and application-level data meaning
- Host operating-system configuration and motherboard design
- Storage-array redundancy, pooling and distributed replication
- Backup contents, recovery-point policy and retention obligations
- Manufacturing processes for semiconductor dies
Characteristics
- Device identity
- Manufacturer, model, serial identifier and hardware revision, with unknown values explicit Connects observations, compatibility evidence and maintenance instructions to the correct physical device.
- Storage medium
- Declared nonvolatile medium and, where documented, cell technology; otherwise unknown Prevents unsupported assumptions about endurance, retention and operating behaviour.
- Physical and electrical fit
- Form factor, dimensions, connector, keying and documented supply requirements Determines whether installation is mechanically and electrically supported.
- Host communication path
- Device interface and protocol linked to host port, adapter or bridge and negotiated connection Explains accessibility, performance limits and commands or health information hidden by intermediaries.
- Exposed storage geometry
- Host-addressable bytes, logical block size in bytes and capacity per exposed namespace or equivalent unit Supports capacity checks and distinguishes physical-device limits from logical allocation.
- Firmware revision
- Reported revision and applicable vendor release evidence Helps assess documented defects, supported features and update applicability.
- Workload performance
- MB/s, operations/s and latency in milliseconds, with workload, queue depth, fill level and test duration Makes performance observations interpretable and avoids treating peak ratings as sustained behaviour.
- Thermal condition
- Reported temperature in degrees Celsius, sensor identity, documented limits and throttling indication Supports cooling decisions and interpretation of performance changes.
- Endurance evidence
- Reported writes with counter units, wear indicators with vendor meaning and documented endurance rating Supports a qualified wear assessment without treating a rating or counter as a guaranteed failure date.
- Integrity and availability condition
- Readable, writable, read-only, intermittently accessible, inaccessible or unknown, with error evidence and timestamps Determines whether continued use, evidence preservation or recovery should take priority.
- Protection and sanitisation condition
- Documented encryption capability, lock state, sanitisation support and last verified operation outcome Separates supported security features from demonstrated protection or removal of stored data.
Also called
Where this came from
wikidata · CC0 1.0
Drafted structure
Bundle to layer to finding to question, as the second pass will find it: 6 bundles · 11 layers · 18 findings · 28 questions.
Device identity and boundary Establish which physical storage device is being modelled and which logical or packaged components belong to it.
An agent must avoid confusing an SSD with its enclosure, a memory component or a logical storage allocation.
Storage device classification
Determine whether the thing meets the registry boundary for a solid-state drive.
Controller-managed solid-state storage
Record evidence for persistent solid-state media, controller mediation and the declared device class.
- What device documentation or observations establish that this is a solid-state drive rather than RAM, a bare memory chip or a hybrid disk? definition
- What boundary distinguishes this device from adjacent flash-storage classes such as memory cards and USB flash drives? boundary
Physical and logical identity
Link persistent device identity to the storage units and packaging visible to the agent.
Drive and exposed unit mapping
Record how identifiers, namespaces or equivalent units, and any enclosure map to the physical SSD.
- Which identifiers came from the drive itself, and which came from a bridge, enclosure or host inventory? provenance
- Which exposed storage units belong to this physical SSD, and which represent partitions, pooled storage or virtual devices? boundary
Host fit and storage exposure Describe how the SSD can connect to a host and what addressable storage it presents.
Matching a connector or advertised capacity does not establish usable installation compatibility.
Connection compatibility
Assess the complete mechanical, electrical and protocol path.
Supported host path
Record device requirements, host support and any adapter constraints.
- Do the target slot, mounting space, connector, supply and supported protocol match the documented requirements of this SSD? boundary
- What connection was negotiated, and does any adapter or bridge restrict bandwidth, management commands or health visibility? measurement
Addressable capacity
Distinguish advertised storage from currently exposed and usable device capacity.
Capacity and block format
Record host-addressable capacity, block format and exposed unit configuration without inferring hidden physical capacity.
- What capacity and logical block size does each exposed storage unit report, and how do these relate to the device's advertised capacity? measurement
- Which supported capacity, namespace or block-format changes are available, and which would destroy data or require host reconfiguration? action
Workload and operating behaviour Relate observed performance to workload, occupancy, thermal state and power conditions.
An SSD can remain accessible while its operating conditions make it unsuitable for the intended workload.
Sustained workload response
Capture performance under stated conditions and distinguish short bursts from sustained operation.
Performance with test context
Record throughput and latency together with the workload and drive state needed to interpret them.
- What throughput and latency were observed for a stated read/write mix, access pattern, block size, queue depth, fill level and duration? measurement
- What evidence separates device limitations from host-path bottlenecks, caching effects or background drive activity? provenance
Thermal and power response
Assess operating limits and documented behaviour during power transitions or interruption.
Operating envelope and interruption
Record thermal observations, applicable limits and evidence about power-loss handling.
- Which temperature sensors and throttling indicators are available, and how do observations compare with documented operating limits? measurement
- What documented power-loss protection and cache-flush behaviour apply to this model, and what prerequisites govern safe shutdown or removal? action
Wear, integrity and recovery Interpret device health evidence and choose actions that preserve access to data.
Wear estimates, interface errors and media failures require different interpretations and interventions.
Endurance and health evidence
Track wear and health indicators with their documented meanings and observation history.
Qualified wear assessment
Relate write counters and wear indicators to applicable ratings without presenting them as precise remaining lifetime.
- What are the units, reset behaviour and documented meanings of the available write, wear and spare-capacity indicators? provenance
- How have these indicators changed over time, and which documented endurance rating and workload assumptions apply? measurement
Failure triage and preservation
Assess access failures and decide whether diagnostics, copying or specialist recovery should come first.
Access condition and next safe step
Record read/write accessibility, error evidence and the effect of proposed actions on recoverability.
- Which observations distinguish media or controller trouble from a cable, power, bridge or host-path fault, and what remains unresolved? measurement
- Given data value, available copies and current accessibility, should the next action be bounded diagnosis, data preservation, isolation or specialist recovery? action
Firmware, protection and disposition Control device-level changes and establish evidence for secure reuse or retirement.
Firmware changes, locking and sanitisation can alter accessibility or destroy data, so supported capability and verified outcome must remain distinct.
Firmware and access control
Record firmware applicability and device-level protection dependencies.
Maintenance and unlock prerequisites
Identify supported maintenance operations, lock state and required authority without storing secret credentials in the model.
- Which vendor evidence links an available firmware update to this exact model and hardware revision, and what prerequisites or data risks are documented? provenance
- What device-level lock or encryption state is established, and what authorised credential or key-management dependency is required for the proposed operation? action
Sanitisation and reuse
Select supported data-removal operations and assess whether their outcomes support the intended disposition.
Sanitisation scope and completion
Record operation scope, connection support, completion evidence and unresolved coverage of inaccessible storage.
- Which sanitisation operations are supported through the current connection, and what documented coverage do they claim for allocated, deallocated, remapped and controller-reserved storage? boundary
- What authorisation and completion evidence are required before reuse or disposal, and what alternative is appropriate if the drive cannot complete or verify the operation? action
Evidence and external alignment What the world already says about this thing, gathered so the model can be checked against it.
A model that cannot be lined up against existing standards, identifiers and practice cannot be adopted by anyone who already uses them.
Reported evidence
Findings from the breadth pass, kept separate from the structural claims.
Kinds and varieties
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- Client SATA SSD (typically 2.5-inch or mSATA)
- Client NVMe SSD (typically M.2, most often 2280)
- Enterprise SAS SSD
- Enterprise NVMe SSD (U.2, U.3, EDSFF, or PCIe add-in card)
- Portable USB SSD
- Industrial and embedded SSD (wide-temperature, often SLC or pSLC)
- QLC capacity-oriented SSD (read-intensive / cold data)
- Self-encrypting SSD (TCG Opal or FIPS-validated)
- Which of these kinds and varieties hold for the sense of solid-state drive this model covers, and on what evidence? provenance
Identifiers and schemes
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- Wikidata - Q175538 - Item for the device class solid-state drive.
- NVMe namespace identifier - NGUID (16-byte) or EUI-64 - Per-namespace unique identifier reported by an NVMe controller.
- ATA Identify / SCSI inquiry - Model string + serial number - Host-visible product identity on SATA and SAS SSDs.
- Manufacturer SKU / part number - Vendor-assigned P/N, often plus firmware revision - The unit sold in BOM and warranty records; firmware revision is part of the identity because behaviour changes with it.
- Which of these identifiers and schemes hold for the sense of solid-state drive this model covers, and on what evidence? provenance
Standards and regulation
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- JEDEC JESD218 and JESD219 - SSD endurance requirements and client/enterprise workloads (JEDEC)
- NVM Express Base Specification - PCIe SSD host command set (NVM Express, Inc.)
- Serial ATA Revision 3.x - SATA electrical and command interface (SATA-IO)
- SCSI / SAS command and transport standards - enterprise SAS SSDs (INCITS T10; SAS from SCSI Trade Association / INCITS)
- PCI Express Base Specification - physical and transaction layer under NVMe (PCI-SIG)
- TCG Storage Opal SSC - self-encrypting drive command class (Trusted Computing Group)
- NIST SP 800-88 Rev. 1 - media sanitization, including flash-specific methods (NIST)
- FIPS 140-2 / 140-3 - cryptographic module validation when an SSD is sold as a validated SED (NIST/CMVP)
- IEC 62368-1 - safety of ICT equipment that houses or is sold as a drive (IEC)
- EU RoHS Directive 2011/65/EU and WEEE Directive 2012/19/EU - restricted substances and end-of-life take-back for electronic storage devices
- Which of these standards and regulation hold for the sense of solid-state drive this model covers, and on what evidence? provenance
Real-world use
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- Primary boot and application volume in laptops and desktops, now usually an M.2 NVMe module on the motherboard.
- Cache or all-flash capacity in servers, NAS appliances, and all-flash arrays, using U.2/U.3, EDSFF, or SAS bays.
- Game consoles and handhelds as the internal game and OS store, often a soldered or proprietary NVMe module.
- Field and industrial machines as a wide-temperature 2.5-inch or mSATA/M.2 card for logging and images.
- Portable backup and media shuttle as a USB-enclosed NVMe or SATA bridge drive.
- Hot-spare and RAID member in datacenters, where endurance class (DWPD) and power-loss protection are specified in the purchase.
- Which of these real-world use hold for the sense of solid-state drive this model covers, and on what evidence? provenance
Typical measurements
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- User-addressable capacity - 250 GB-4 TB consumer; 1.6-30 TB enterprise - GB or TB (decimal, 10^9/10^12 bytes, as labelled)
- Sequential read throughput - ~500-550 (SATA); ~3000-3500 (PCIe 3.0 x4); ~5000-7500 (PCIe 4.0 x4); ~10000-14000 (PCIe 5.0 x4) - MB/s or GB/s
- Random 4 KiB read IOPS - ~50e3-100e3 SATA client; ~300e3-2e6 NVMe client/enterprise - IOPS
- Read latency (typical 4 KiB) - 20-150 - µs
- Endurance - Client ~150-600 TBW per 1 TB; enterprise often 0.3-3 DWPD over 3-5 years - TBW or DWPD
- Active power - 2-8 client NVMe/SATA; ~10-25 enterprise NVMe - W
- Uncorrectable bit error rate (UBER) - 10^-15 to 10^-17 (class-dependent) - errors per bit read
- Operating temperature (client) - 0-70 case, with throttling above ~70-80 controller - °C
- Which of these typical measurements hold for the sense of solid-state drive this model covers, and on what evidence? provenance
Failure modes and hazards
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- NAND wear-out: program/erase cycles exhausted, rising bit errors, drive goes read-only or fails SMART endurance attributes.
- Sudden power loss without board-level capacitance: in-flight writes and mapping-table updates lost or inconsistent, especially on DRAM-less client drives.
- Firmware defects: bricking, performance collapse, or silent data loss after a specific command or garbage-collection path.
- Data retention loss: unpowered NAND at high temperature loses charge, so archived drives can become unreadable.
- Thermal throttling and, rarely, thermal runaway of the controller or power stage under sustained writes.
- Write-amplification and QLC/TLC performance cliffs after the SLC cache is full.
- Controller, DRAM, or power-component hardware failure independent of NAND life.
- Sanitization failure: block overwrite does not reliably erase all flash; residual user data may remain without crypto-erase or block erase.
- Counterfeit or relabelled drives reporting false capacity or endurance.
- Difficult forensic recovery relative to HDDs because of FTL indirection, encryption, and wear-leveling.
- Which of these failure modes and hazards hold for the sense of solid-state drive this model covers, and on what evidence? provenance
Regional variation
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- English marketing still sometimes says "solid-state disk" or "hard disk" for an SSD; specialists prefer "drive" because there is no disk.
- Mandarin product language is 固态硬盘 (gùtài yìngpán); Japanese and Korean markets likewise use the SSD loanword plus local "SSD/solid disk" retail labels.
- Consumer default form factor is globally M.2 2280 NVMe; 2.5-inch SATA remains common in NAS, optical-bay upgrades, and many enterprise trays.
- US and EU public-sector buys often require NIST 800-88 or equivalent sanitization and, for SEDs, FIPS 140 validation; those clauses are less common in consumer channels.
- China, Korea, Japan, and the US dominate NAND and controller supply, so firmware behaviour, SMART attribute layouts, and warranty terms follow vendor rather than a single regional code.
- In parts of South Asia and Latin America, "pendrive" or "memoria" may be used loosely for any semiconductor store, including USB SSDs.
- Which of these regional variation hold for the sense of solid-state drive this model covers, and on what evidence? provenance
Neighbouring kinds and how to tell them apart
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- Hard disk drive (HDD) - An HDD stores bits on rotating magnetic platters with a moving head; seek/rotational latency is milliseconds and the device has acoustic and shock-sensitive mechanics. An SSD has no platters; latency is tens to hundreds of microseconds and identity is the semiconductor FTL plus host protocol.
- USB flash drive (thumb drive) - A thumb drive is a USB mass-storage gadget with a small controller and usually no DRAM, power-loss capacitors, or NVMe/SATA host stack. A USB SSD is a full SSD (often NVMe or SATA) in an enclosure with a bridge; endurance, SMART, and sustained-write behaviour follow SSD class, not gadget-flash class.
- eMMC or UFS device - eMMC/UFS are soldered package-on-package or BGA memories on a JEDEC/MIPI protocol for phones and boards, not a removable SATA/NVMe/SAS drive with a drive-bay or M.2 keying.
- Solid-state hybrid drive (SSHD) - An SSHD is an HDD with a small NAND cache; the capacity medium is still magnetic platters. If the bulk medium is only NAND (or other persistent semiconductor memory), it is an SSD.
- NVDIMM / persistent memory - Persistent memory sits on the memory bus (DDR) and is byte-addressable to the CPU. An SSD sits on a storage bus and is block-addressable via NVMe, SATA, or SAS.
- SD / microSD memory card - A memory card uses the SD Association protocol and a card form factor; it is not a host SSD protocol device and is specified to different endurance and command sets.
- Which of these neighbouring kinds and how to tell them apart hold for the sense of solid-state drive this model covers, and on what evidence? provenance
Sources
- Solid-state drive - Wikimedia Foundation (Wikipedia) - Common definition, interfaces and form factors, distinction from HDDs and USB flash, typical consumer and enterprise uses.
- JESD218: Solid-State Drive (SSD) Requirements and Endurance Test Method - JEDEC Solid State Technology Association - Industry definition of SSD endurance testing, TBW/DWPD framing, and client versus enterprise workload assumptions.
- NVM Express Base Specification - NVM Express, Inc. - The dominant host protocol for PCIe SSDs, including Identify, namespaces, and NGUID/EUI-64 naming of namespaces.
- NIST Special Publication 800-88 Revision 1, Guidelines for Media Sanitization - National Institute of Standards and Technology - Why SSD sanitization differs from HDD overwrite, and crypto-erase / block-erase practice in regulated settings.
What the second pass must settle
- Does an existing Vercy world model already own this concept, requiring a registry link rather than a separate publication?
- Where does the registry draw the boundary between SSDs, embedded managed storage, memory cards and USB flash drives, including devices inside external enclosures?
- Which authoritative specifications and vendor documents establish the meanings and comparability of health, wear and error indicators across the SSD families covered?
- What model-specific evidence is required to assess power-loss protection, unpowered data retention and recovery limits without generalising across storage technologies?
- Which sanitisation methods and verification evidence satisfy the intended reuse or disposal requirements for each supported controller, firmware and connection path?