Maintenance
Part of File Sharing Software: What to Test Before You Buy
File sharing security: questions to test in a trial
File sharing security checks help teams assess access history, external links, sync controls, recovery, retention, and access reviews in a trial.
What to take away
- Four questions help evaluate file sharing security in a business setup.
- Historical access is the hard onewho could reach a file on a past date, and since when.
- Audit logs have retention limits. Check each plan's window and whether it supports export.
- External links accumulate because no step closes them. Expiry by default is the fix.
- A restore you have not timed is a plan, not a control.
The four questions to test in a trial account
Run all four against one folder, with two members and one external guest, before you sign.
| Question | What a good answer looks like |
|---|---|
| Who can reach this file today? | Membership now, including groups and guests |
| Who opened, downloaded or deleted it? | Per-user activity with exportable timestamps |
| Which link carried that access? | Link creator, creation date, expiry date |
| Who had access last March, and since when? | Point-in-time membership history |
Questions one to three test current access and activity. Question four tests whether the product can reconstruct past access, which matters when resolving disputes.
If a product shows only today's state, your evidence about last year is whatever your team wrote down at the time.
Permission models differ between products, which is why the file sharing review belongs beside this page when you compare tiers.
Four questions to test
- Who has access, including groups and links?
- When was each grant made, and by whom?
- Who opened this file, and how far back?
- What did access look like six months ago?
External links outlive the reason they were made
External shares accumulate over time, one reasonable decision at a time.
- A link outlives its reason, because no step in the workflow closes it.
- A link is transferable, so the holder may not be the person it was made for.
- A folder-level link covers everything added later, including files that did not exist when it was created.
The countermeasures are dull and they work. Outside shares expire by default. The owning group reviews a listing of active links each quarter. Outside access is granted at file level unless somebody explains why a folder is needed.
Retention windows, defaults and export
Audit records expire, and retention varies by product and tier. Check the default window and whether you can export the log on demand.
Google Vault and Microsoft Purview eDiscovery are the best known add-on tiers built for retention, hold and export. Base business tiers of the same products, and the consumer sync tools many teams start with, generally show current membership and a short activity feed.
Check two things in a trial: the default window, and whether you can export the log on demand. For plan details, see file-sharing pricing.
An audit trail earns its keep only if it names the user, the action, the timestamp and the address the action came from.
Decide the lookback you would need after a client complaint, then buy retention that covers it.
Sync clients put copies on laptops
Sync makes these products pleasant. It also puts a full copy of a shared folder on a device you may not control. Three decisions follow.
- Which devices may sync, and whether policy enforces that or only states it.
- Whether selective sync is on by default, so a new member does not pull a large space by opening one file.
- What remote removal does to a lost device, and how quickly it takes effect.
NIST SP 800-171 is written for a narrower context than most readers occupy, but its requirement holds: limit access to authorized users and to the actions they may take.
Example: a restore test in a trial folder
Encrypted files sync. A client replicates the damage to the shared copy and to every other copy, at speed. Test the path before you need it.
- Add a set of files to a trial folder and give two members access.
- Encrypt or corrupt the test files.
- Restore from the vendor's rewind or version history, and start a timer.
- Record whether sharing settings returned, or whether the restored files arrived owned by whoever ran the restore.
Restore time depends on the recovery process and the amount of data. Measure it in your own environment rather than relying on a general benchmark.
The CISA ransomware guidance covers the wider problem. The number you need is your own restore time on your own estate.
Restoring is a permissions problem as much as a copying one, which is the argument in migration is a permissions problem.
Four things to test in restore
- Windowhow far back, admin vs user?
- Granularitywhole folder or file by file?
- Speedfour days or two hours?
- Statesharing settings kept or lost?
Two classification tiers you will maintain
Elaborate schemes fail the way elaborate naming schemes fail. People stop applying them, and a half-applied scheme produces false confidence.
Two tiers survive: ordinary and restricted. Restricted means a short written list, such as material holding personal data or material under a confidentiality obligation.
Apply exactly one rule to restricted that applies nowhere else. Banning external links is the usual choice.
Most enforcement stories involve ordinary lapses rather than sophisticated attacks, which is the point of the FTC data security guidance.
Who owns the twice-yearly access review
Twice a year, the owning group of each top level space receives a list of who has access and every active external share. They confirm or remove each entry. No forms.
The value is not the removals. It is that somebody looks.
Departures need their own check, because content owned by a leaver is the commonest way material becomes unreachable. That check belongs with the joiner and leaver process, and it works only if spaces were assigned to groups in the first place.
