2 Moving cold data off NAS filers is one of the fastest ways to cut storage cost and delay an expensive refresh. Enterprise scale-out NAS clusters and filer systems often hold petabytes of file data, and analysis typically shows that much of it has not been accessed in a year or more. Keeping that data on primary NAS means paying premium prices for capacity, snapshots, replication and backup of files nobody uses. Moving it to S3-compatible object storage, while keeping it findable and recoverable, frees capacity for active work and shrinks the next refresh. This article explains how to identify cold data, the main methods for moving it, how stubs and recall work, what happens to backup and permissions and how to plan the project. It is vendor-neutral and applies to most enterprise NAS platforms. For the strategic picture, see our hub on when object storage can replace NAS. Step 1: Find the cold data Start with file analysis. Tools that scan file systems or read filer metadata can report: Age: last access and last modification times by share and folder. Size distribution: where large files concentrate capacity. File types: media, archives, backups, logs and office documents. Owners: departed users, completed projects and orphaned shares. Duplicates: identical files stored in multiple places. Be aware that some filers or configurations do not update access times reliably, so combine access time with modification time and business knowledge. The results usually reveal a large share of capacity in data untouched for a year or more. Step 2: Decide the policy Define what counts as cold and where it should go: Age thresholds, such as files not accessed or modified in 12 or 24 months. Exclusions for active projects, databases, application data and files that must stay on NAS. Whole-share moves for completed projects or departed users, versus file-level tiering within active shares. Retention for archived data, including legal holds and deletion schedules. Agree the policy with data owners, legal and records management where needed. Step 3: Choose a method Native filer tiering features Many NAS platforms include features that tier cold data to object storage. Some tier at the file level, leaving stub files that point to data in the object store; others tier at the block level beneath the file system. These features are convenient and transparent to users, but data is typically stored in the filer’s own format, so it cannot be read directly as files from object storage, and it remains tied to that filer platform. Check whether the feature supports your S3-compatible target and what happens if you later retire the filer. Third-party data movers with stubs Data management tools can scan NAS shares, move cold files to object storage and leave stubs or links so users can still open them. Some store each file as a normal object with its original name and metadata, making data readable over S3 independently of the tool. Recall happens when users open a stubbed file. Copy and remove without stubs For whole shares or project folders, the simplest method is copying data to object storage and removing it from the NAS, with users accessing archived data through a portal, a gateway or S3 tools. This gives the cleanest result and no stub dependencies, but changes how users find data. Gateway-based access Archived data can be presented back to users through a file gateway, giving read access via familiar shares without stubs on the original NAS. See what a cloud file gateway is. Stubs and recall: the trade-offs Stubs keep paths unchanged, which minimizes user disruption. But they introduce dependencies: Recall performance depends on network and object storage speed. Backup software may recall every stubbed file if not configured correctly, or may back up only stubs. Antivirus and indexing tools can trigger mass recalls. Migration complexity increases when retiring the NAS, since stubs must be resolved or moved. Native object formats, where each file becomes a readable object, reduce long-term lock-in. Weigh convenience now against flexibility later. Permissions and metadata Files carry permissions, ownership and timestamps that matter for security and compliance. When moving data: Preserve ACLs and ownership where users will access data through file interfaces. Map permissions to bucket policies or gateway shares where access is via S3 or a portal. Record original paths and metadata with each object, for findability and audit. What happens to backup Moving cold data off NAS usually shrinks backup windows dramatically. Archived data on object storage can be protected with erasure coding, replication to a second site, versioning and object lock instead of repeated backups. Make the new protection model explicit and ensure backup tools handle stubs correctly. See S3 object lock: immutability and WORM. Planning and executing the project Pilot with one share or department to validate the method, recall behavior and user experience. Communicate with users about what is moving and how to access it. Move in waves, starting with the largest and coldest data. Throttle data movement to avoid impacting NAS performance during business hours. Verify moved data with counts and checksums before removing it from NAS. Report capacity reclaimed to show progress and justify the next phase. Estimating the savings Build a simple model before starting. If a filer holds 1.5 PB and analysis shows 60 percent has not been accessed in over a year, moving that 900 TB frees most of the capacity that would otherwise drive the next expansion or refresh. Savings come from several places: avoided NAS capacity and support costs, reduced snapshot and replication capacity on the NAS, smaller backup licenses and infrastructure and lower power and rack space. Against that, add the cost of object storage for the archived data, data movement tools and staff time. In most cases the cost per terabyte on object storage is a fraction of primary NAS, and backup savings alone can be significant. See storage cost per terabyte. Keeping NAS from filling up again Moving cold data once helps; keeping it moving helps more. Schedule regular policy runs so newly cold data continues to flow to object storage, set quotas on shares, archive departed users’ data as part of offboarding and close completed project shares on a defined schedule. Report NAS growth and archive volumes to data owners so they see the effect of their choices. Sizing the object storage target Size for the cold data you will move, growth from future archiving, protection overhead and headroom. Throughput must handle bulk moves without excessive duration. On-premises object storage keeps data local and avoids cloud retrieval fees when users recall files. See storage capacity planning. Regional considerations File data in Europe, Japan and the Middle East often includes personal or sensitive data subject to location requirements. Moving cold data to on-premises object storage keeps it within the organization’s facilities and country, while public cloud tiering may raise data transfer questions. See data sovereignty vs data residency. Ransomware benefits Archived data on immutable object storage is out of reach of ransomware that encrypts mapped file shares. Moving cold data reduces what attackers can damage on primary NAS and guarantees a clean copy of historical files. Checklist: move cold data off NAS Run file analysis by age, size, type, owner and duplicates. Define cold data policies and exclusions with data owners. Choose native tiering, data movers, copy-and-remove or gateway access. Understand stub behavior, recall performance and lock-in. Preserve permissions and record metadata. Configure backup and security tools for stubs. Protect archived data with replication, versioning and object lock. Pilot, communicate, move in waves and verify. Size object storage for capacity and throughput. Putting it together Moving cold data off NAS filers frees premium capacity, shortens backups and reduces the size and cost of the next refresh. Start with file analysis, agree clear policies, then choose a method that balances user convenience against long-term flexibility, with native object formats offering the most freedom. Protect archived data on object storage with replication and immutability, verify every move and report the savings as you go. Frequently asked questions How much NAS data is typically cold? File analyses commonly find that a large share of capacity has not been accessed in a year or more, though it varies by organization. What is a stub file? A small placeholder left on the NAS that points to data moved to another tier, allowing users to open the file as if it were still there. Can native NAS tiering features use any object storage? Many support S3-compatible targets, but check compatibility and note that data is usually stored in the filer’s own format. Will moving cold data affect backups? It usually shortens them significantly, but backup tools must be configured to handle stubs correctly. How do users access archived data without stubs? Through gateways, portals or S3 tools, depending on the access model chosen. Further reading See when object storage can replace NAS, cloud file gateways, NFS and SMB access to object storage, distributed office file sharing and cold data migration strategy.
