Data rarely sits still. A data science team wants last night's production share in the cloud by morning. QA needs a fresh test/dev copy every sprint. A new project spins up and needs a working set of files in its own environment. And every few years, somebody has to retire a filer.
Moving files sounds simple until the file count hits the billions. Then you are babysitting a copy that hammers the source, can't tell you what it already moved, and falls over at 2 a.m. on a Sunday.
NAS Cloud Direct (NAS CD) Copy is a policy-driven engine that syncs file and object data from a source to an on-prem or cloud target. You manage it from Rubrik Security Cloud (RSC) so schedules, assignments, and monitoring all live in one place. The heavy lifting of moving the data is done by Rubrik's own tooling, through a stateless VM that sits right next to your data.
What Copy Is (and Is Not)
Copy is a one-way sync from a source to a target, on-premises or in the cloud. It keeps a 1:1 replica of the current source state. It does not keep versions or point-in-time snapshots, so it is not a backup and not a recovery tool.
File-to-file copies use the same protocol on both ends (NFSv3 to NFSv3, or SMB to SMB). Cloud targets include S3 and S3-compatible storage, Azure Blob, and Google Cloud Storage, in any tier.
Key Use Cases
Copy moves data to another workspace or environment, and the most frequent reasons look nothing like a migration.
Refresh test/dev shares: Give dev and QA teams a current copy of production data on a schedule, without a manual pull each sprint.
Stand up new projects: Move a working set of files into a new environment so a team can start on real data.
Move to the cloud for analytics and data science: Land file and object data where cloud analytics and data science tooling can reach it.
Migrate between NAS platforms: The least frequent use case, and the one with the highest stakes. Incremental runs keep the target current until cutover.
Replicate to a second site or archive cold data: Keep a second copy elsewhere, or push cold data to lower-cost cloud tiers such as S3 Glacier, Azure Archive, or GCS Coldline and Archive.
Copy is not a backup. You pair it with NAS CD backup when you also need point-in-time recovery.
Why Copy
Copying files is the easy part. The value is in making the copy fast, safe, and governed once the data gets big. Here is what Copy delivers:
Incrementals: After the first copy, each run diffs the source against a catalog and moves only what was added, changed, or deleted.
Built for scale: Scanning, indexing, and copying run in parallel and in batches on the same engine NAS CD uses for backup, which discovers and indexes hundreds of thousands of files per second.
Protocol aware: SMB stays SMB and NFS stays NFS. Rubrik's own NFS, SMB, and S3 clients read the source and write the target, so the data keeps its native protocol behavior, and NTFS ACLs and hard links come along on SMB when the target supports them.
Checkpointing: Metadata is committed to the catalog only after the target acknowledges the writes. Because the catalog lives in the control plane and not on the VM, you can pull the VM out and re-attach it, and the copy resumes where it left off.
Built-in source throttling: The VM measures latency on your source and adjusts the number of connections, so production gets priority over transfer speed.
Bulk actions: Assign a schedule to up to 100 shares in one request and get a result for each share.
Policy-based management: A schedule is a name, a frequency, and a target, managed from RSC with RBAC, quorum approval, and audit.
One Place to Run It: Rubrik Security Cloud
A copy schedule lives on an NCD cluster, and many source shares can share one schedule. From RSC you can:
Create, edit, pause, resume, and delete schedules
Assign a schedule to shares and get a result for each one
Open a share's Copies tab, next to Snapshots, to see where its data goes
Follow copy jobs in Events and task monitoring
What Happens on Every Run
Three pieces do the work:
The stateless VM sits near your source (VMware, Nutanix AHV, Hyper-V, KVM, AWS EC2, or Azure VM) and handles the scanning and copying. Data passes through it but is never stored on it.
The NCD Control Plane is a Rubrik-managed tenant isolated to you, holding the scheduler, the Cloud Slab, and the Index DB, which is the catalog of everything already copied.
Rubrik Security Cloud is where you tell it what to do. The VM only ever calls out to the control plane, over HTTPS on port 443 with TLS 1.3. There are no inbound connections.
Here’s how the run itself works:
Scan: The VM reads the source over NFS, SMB, or S3 and builds an index of the metadata.
Decide what to compare: The first copy has nothing to compare against, so it skips ahead and copies everything. Every later run has the Lister pull the previous scan from the Index DB and line it up against what the source looks like right now.
Work out the diff: New scan plus old index gives a list of what was added, changed, or deleted. Nothing else gets touched.
Read only what changed: The differential data is pulled from the source.
Write it out, politely: A multi-threaded copy job writes to the target and sets permissions on the target array. If production slows, Copy backs off.
Commit: Only after the target confirms the writes does the VM commit metadata to the Index DB. If a run dies halfway, the catalog never claims something was copied when it wasn't.
The Short Version
Moving data is easy to describe and hard to do well. Copy is built around the places where copies usually go wrong:
It diffs against a catalog, so every run moves only what changed.
It watches your source and backs off when production slows down.
It commits to the catalog only after the target confirms the write, so the record is always honest.
It runs from one place in RSC, with RBAC, quorum approval, and audit built in.
Whether you are refreshing test/dev, feeding a data science project, or retiring a filer, the result is the same: a copy you can schedule, watch, and trust.
See It In Action
Want to see Copy in action? Walk through the NAS CD clickthrough demo to see how schedules are created, assigned, and monitored.
For more details on how Copy works, reach out to sales@rubrik.com or your Rubrik representative.