Welcome back to Office Hours. Session One covered data tiering and the storage price spike. Session Two covered AI data workflows and KAPPA data services. This is session three, on unstructured data migration, with Benjamin Henry, Field CTO, Komprise.
Watch the recordings and register for live sessions at komprise.com/office-hours.
The Pain: Migrations Get the Knot in Your Stomach for a Reason
The same storage price spike driving the tiering conversation is pushing more migrations to the cloud right now, where capacity and pricing are not the constraint they are on-premises. But most migration anxiety, and most cost overrun, traces back to one habit: skipping the assessment and diving straight into moving data. Ben puts it simply: 80% of a successful migration is the analysis. Get that right, and the actual data movement is the easy part.
Step One: Assess Before You Move Anything
Run ACE first. Komprise Assessment of Customer Environment (ACE) operates outside Komprise entirely, testing hypervisor performance, network and storage (NAS, NFS, or SMB), so you know your expected throughput before modeling anything.
Add a data store once, at the cluster level. Point Komprise at a Dell PowerScale (Isilon) or NetApp cluster and credentials, and it auto-discovers every access zone, share, volume, and export protocol underneath: no manual entry per share.
Enable analysis in bulk or one at a time, then customize your columns per session. For migration planning, order by share path (the real relative path on the cluster, not just the share name), directories, items, protocol, and percentage of cold data (default threshold: 1 year, fully adjustable).
Export the same screen as your migration worksheet. Add two columns, destination share and wave (your cutover grouping), and the assessment report becomes your change-control document.
Step Two: Decide Lift-and-Shift vs. Smart Migration
The same heat map used for tiering also models a migration. Before moving anything, ask whether it makes sense to bring cold data along at all. Slide the cold-data threshold and watch the model update live. Moving the age cutoff from one year to three might show 40% of a share qualifying as cold; tightening it to three months might show 54%, each with its own capacity, file-count breakdown, and three-year cost projection.
A smart migration tiers cold data first, then migrates only what’s active. A lift-and-shift drags everything along, including data nobody has touched in a year or more, at today’s storage prices.
- Check the usage tab for what will slow you down.
- View by file count or size.
- Large files move fast; a large population of small files is “the disruptor,” and worth knowing before you commit to a timeline.
- Multi-byte file and folder names are supported natively, unlike platforms that only handle single-byte characters.
- Watch latency while analysis runs. High latency between source and destination flags a bottleneck to fix before migration, not during it, which is exactly what ACE is designed to catch up front.
Step Three: Run and Audit the Migration
Point, pick a scope, go. Choose source and destination. Komprise can auto-create destination shares then an entire share or a subdirectory on either end.
Pick up where another tool left off. If a share already has partial data from rsync, robocopy, or any other tool, Komprise detects it and continues instead of recopying the whole data set: useful for WAN and cloud migrations.
Keep latest version for near-zero-downtime cutovers. Common in healthcare EHR or media environments where writes never stop: whichever side has the newer file version wins, so nothing gets overwritten by an older copy.
Map security identifiers (SIDs) in flight for SMB migrations. Useful in M&A and Active Directory consolidation, mapping old SIDs to new ones during the copy itself avoids a separate remediation pass.
Schedule by interval, time of day or manually, and run multiple migrations at once from a template.
Audit logs record everything by default: every action, every iteration, plus file-level validation (MD5 for NFS, hash value for SMB) on every copy. Centralize them off the destination so auditors, legal, or migration partners get access without asking.
Chain of custody is a rerun, not a replacement. It reruns full validation at the final iteration for compliance sign-off, extending the cutover window. Most teams rely on the standing audit logs and add chain of custody only when a specific requirement calls for it.
Why This Holds Up at Scale
Komprise recently patented Elastic Shares, a dynamic load-balancing technology that continuously redistributes work across observers so that no machine sits idle while another churns through a dense branch of the file tree. Elastic Data Migration runs 27 times faster than native tools for NFS files, and Hypertransfer runs 25 times faster for SMB files.
A migration assessment surfaces more than a moving plan. A Deep Analytics query for an undefined file owner turns up orphaned data, which SID mapping can then reassign instead of migrating ownerless files. The same Showback report used for tiering works here too: it prices out what a share costs annually whether or not you migrate it.
Beyond migration, many Komprise customers leverage the full Intelligent Data Management platform for ongoing storage refresh, tiering, ROT clean up, sensitive data detection and mitigation, AI data curation and workflows.
Key Takeaways
- Spend the bulk of your time on assessment and the Data Stores page to model performance and cost before a single file moves.
- Decide lift-and-shift versus smart migration with real numbers. The cold-data model shows exactly what you’d save by tiering first.
- Audit logs, not just chain of custody, to cover most compliance needs. Reserve the full revalidation rerun for when a specific requirement calls for it.
- Elastic Shares and Hypertransfer are why this scales. Dynamic load balancing beats static partitioning on large, uneven data sets.
Watch the Webinar
Watch all of the recordings and register for the next one at komprise.com/office-hours.
