As enterprises generate more and more data, one question increasingly haunts IT and data governance teams: What should we do with dark data? Is it better to delete it outright to reduce risk and costs, or is it wiser to tier it into cheaper storage, balancing accessibility and expense? This decision is anything but trivial — especially when dealing with unstructured data spread across NAS systems and object storage.
What Is Dark Data and Why Does It Persist?
Dark data refers to data that organizations collect, process, and store—but do not actively use for business purposes. Think of files left on a shared NAS folder, dormant backups, old email attachments, or compliance archives piling up over the years. It’s “dark” because it’s invisible or inaccessible to typical governance and analytics efforts.
Why does dark data stick around so stubbornly?
- Fear of Deletion: Nobody wants to be the person who deletes “something important.” Without clear ownership (“Who owns this folder?”), data accumulates by default. Lack of Visibility: Unstructured data folders often lack proper metadata or tagging, making it hard to identify what’s truly needed. Compliance and Legal Concerns: Some data must be retained to meet regulations, but precisely which data can be murky. Backup Sprawl: Backup policies multiply data copies, further obscuring “live” versus obsolete data.
The Visibility Problem with Unstructured Data
Structured data in databases is relatively straightforward to manage, but unstructured data—documents, images, videos, and logs—are much harder to catalog and analyze at scale.
Most organizations rely heavily on NAS (Network Attached Storage) solutions for unstructured file shares and increasingly supplement with object storage platforms for scalability and cloud integration. However, these systems typically lack embedded intelligence to distinguish "dark" from "active" data, causing a “data swamp” effect:
- NAS shares accumulating gigabytes to terabytes of untouched files, increasing storage lensblindness. Object storage platforms becoming inexpensive dumping grounds but with unclear data hygiene, causing hidden costs. Invisible dark data growing unnoticed due to lack of searchable indexes or data classification tools.
The Human Factor: Ownership and Governance
Before weighing tier vs delete, answering “Who owns this data?” is critical. Without clear folder or project owners driving decisions, data remains stagnant because no one feels responsible for cleanup or classification.
Tier vs Delete: Cost Savings and Risk Reduction
When tackling dark data, organizations face a fundamental choice:
Sound simple? It isn't. Here are key considerations guiding the Article source decision:
1. Cost Implications
Storage isn’t free—especially when you factor in operational overhead and backup multiply effects.
Storage Type Typical Cost/GB/Month Backup Multiplier Impact Hidden Costs Primary NAS (active tier) $0.02 - $0.10 3-5x (snapshots, replication) High power, cooling, management Cold NAS Tier (archive tier) $0.005 - $0.02 1-2x Lower power, lower performance Object Storage (cloud or on-prem) $0.003 - $0.01 1-2x (mostly snapshots and versioning) Egress fees with retrievalBack-of-the-napkin math shows why deletion is tempting. If you delete 1TB of dark data that’s replicated and snapshotted 4 times, you're truly removing up to 4-5 TB of storage burden — instantly saving on disk costs, backup windows, and slowdown impacts.
Conversely, tiering reduces costs but does not eliminate them. You pay less to store data on cold tiers or object storage. But watch out: egress fees and longer retrieval times can add operational headaches, especially if high-speed restores or ransomware recovery are needed.
2. Risk Reduction and Ransomware Exposure
Dark data sitting in primary NAS shares is an easy target for ransomware attacks, which often focus on live file systems ripe for encryption.
Tiering data to isolated archive tiers or immutable object storage can reduce exposure—effectively creating air-gaps and write-once storage classes. However, if tiering isn’t implemented with strong access controls or immutability, it merely delays the risk.
Deleting unneeded data outright removes the attack surface completely. No copies mean fewer targets. But be cautious—you must be certain data is truly obsolete.
3. Recovery Time Objectives (RTO)
Tiered data typically takes longer to recover:
- Cold NAS tiers have slower disk speeds optimized for capacity, not IOPS. Object storage often involves data retrieval delays and network egress.
If you keep dark data on less expensive tiers, recovery in incident response scenarios (ransomware recovery, audits) may extend from minutes to hours or even days, increasing operational risk and downstream costs.
4. Compliance and Legal Holds
Some dark data cannot be deleted due to regulatory hold requirements or litigation. In such cases, tiering into immutable storage is preferable and often mandatory.
Practical Guidance: Who Owns This Folder?
Before tooling or tech decisions, ask the simple https://technivorz.com/why-does-dark-data-matter-for-ai-projects/ but powerful question: “Who owns this folder?”
In my experience, once a data owner is identified, clear decisions can be taken on whether to delete, retain, or tier data—rather than letting it fester ungoverned. This makes governance discussions less about vague buzzwords like "AI-ready in minutes" and more about concrete risk and cost tradeoffs.
Tooling Considerations: NAS and Object Storage
Modern NAS solutions often support hierarchical tiering, moving cold files automatically to cheaper storage, including object storage targets. This makes tiering seamless and transparent at the user level.
Object storage, whether on-prem or cloud-based, excels for archiving due to low cost and scalability but demands careful planning around:
- Access policies and encryption Immutable versions and retention locks Retrieval cost and recovery SLAs
Backup tools must also be reviewed—if backups keep multiplying copies of dark data, you may increase costs and extend recovery times unintentionally.
Summary: To Tier or Delete?
Prioritize identifying data owners before deciding the fate of dark data. Delete permanently data that no longer has business value, compliance need, or legal hold. This yields maximum cost savings and risk reduction. Tier to cold NAS or object storage when data must be retained but accessed infrequently to save costs while preserving availability. Build immutable tiers or air-gapped storage in your tiering strategy to reduce ransomware and insider threat risk. Beware backup sprawl and test recovery times — to avoid hidden costs and extended outages.Dark data is a hidden drag on your IT environment in terms of cost, security, and operational complexity. Thoughtful tier vs delete policies, backed by strong data ownership, will unlock real risk reduction and cost savings.
Remember: Don’t just dump dark data into cheaper tiers and call it done—know who owns it, understand when and how it will be used or recovered, and always do the back-of-napkin math on total cost of storage and recovery impacts.