If you’ve inherited decades of records, the storage decision on your desk is bigger than where the servers go. Some of what you manage is paper in boxes, some might be a scanning project that stalled halfway, or digital files spread across systems that predate you. All of it carries retention periods, and any of it could end up under a legal hold.
Most cloud vs on-premise comparisons treat storage as an infrastructure question about backups, file shares, and ransomware, and they leave out the paper, the backlog, and the retention clock. This records storage decision framework takes cloud, on-premise, and hybrid cloud storage through four factors, so you end up with a choice you can explain to finance, legal, and your auditors.
Three Ways to Store Records
- Cloud storage. Your records sit on infrastructure a cloud provider owns and runs, and you pay for capacity and usage as you go. The provider supplies the hardware, redundancy, and security tools. Deciding who can open what stays with your team.
- On-premise storage. You run the servers in your own building with your own staff. You control every layer, from the hardware to the backup schedule, and you carry every cost of keeping it running.
- Hybrid cloud storage. A mix of the two, with each group of records assigned to the tier that fits it. For records, a hybrid model usually adds a third tier: paper archives held by an offsite storage provider, indexed so any box or folder can be pulled, and often scanned, on request.
On-premise storage still has a place. Air-gapped systems, which are cut off from the internet and other outside networks, remain the right call for some highly sensitive data, including parts of government and defense work. A cloud-first build can keep one or two on-premise records management systems where the data calls for it.
The Hybrid Cloud Storage Decision Framework
Start by sorting what you hold into record classes, groups that share an access pattern and a retention rule. Three classes cover most inherited archives:
- Active records: files people open daily or weekly, such as current contracts, open cases, and this year’s accounting
- Semi-active records: closed files still inside their retention period that someone requests now and then, for an audit, a dispute, or a former employee’s question
- Archival records: files kept only because a retention schedule or a legal hold says so, which is where most decades-old paper sits
Then score each class against four factors. The table shows how each architecture tends to perform, and the sections after it explain why.
| Factor | Cloud | On-premise | Hybrid |
|---|---|---|---|
| Total cost of ownership | Low upfront cost, ongoing usage fees, egress and retrieval charges | High upfront cost, hardware refresh, staff, power, and floor space | Each record class priced on the tier it actually needs |
| Recovery time and recovery point | Fast on standard tiers; archive tiers can take hours to restore | As fast as your own backups and staff | Short targets for active records, longer ones accepted for archives |
| Compliance and retention | Vendor attestations under a shared-responsibility model | Full control and full burden; fits air-gapped and tightly regulated systems | Two tiers to govern, so one retention schedule and hold process has to span both |
| Paper and scanning backlog | Holds only what’s been scanned | Paper usually sits on-site | Paper indexed offsite and scanned on demand |
Total Cost of Ownership and Egress In Hybrid Cloud Storage
Total cost of ownership (TCO) is everything a system costs over its life, not only the purchase price or the monthly bill. Each architecture hides different costs.
On-premise TCO includes servers and storage arrays, the refresh every few years, backup media, power and cooling, floor space, software licenses, and the staff time to run it all. Cloud TCO includes storage fees, charges per request, and egress fees, which are what a provider bills to move data out of its cloud. The UK’s Competition and Markets Authority concluded in its cloud services investigation that egress fees restrict customers from switching providers or using more than one.
AWS, Google Cloud, and Microsoft Azure now waive egress fees, on request and with conditions, for customers moving their data off the platform. The waivers apply when you leave, though, not when you pull files back for everyday work.
Cloud archive tiers add two more line items. Retrieval is billed separately, and it isn’t quick. A standard retrieval from Amazon S3 Glacier Deep Archive typically finishes within 12 hours. The deepest archive tiers at both AWS and Azure also carry a 180-day minimum storage period, with a charge for deleting data early.
Paper has its own TCO. Shelves in a basement cost floor space, plus staff time every time someone goes looking for a box, while offsite storage is usually billed per box each month, with pickup and rush or scheduled retrieval adding to the bill. It’s worth comparing the cost of document storage in-house against a professional facility before you decide where the boxes go.
Scanning a box is a one-time conversion cost, but the images then sit in storage and are billed every month for as long as you keep them.
A scan-on-demand archive avoids paying to store images nobody ever accesses. The paper stays in a box until someone needs a file, the provider scans only that file, and the image goes straight into your active tier. Nothing in the archive incurs egress, because it never entered the cloud.
Recovery Time and Recovery Point In Hybrid Cloud Storage
Two targets govern recovery for each record class. The National Institute of Standards and Technology (NIST) defines the recovery time objective (RTO) as how long a system can stay in recovery before the outage harms the organization’s operations. Its companion, the recovery point objective (RPO), is the point in time your data has to be restored to, which sets how much recent work you can afford to lose.
Active records need a short RTO and an RPO close to zero, since an hour of lost billing entries or patient notes is an hour someone has to re-create. That points to cloud storage with replication, or an on-premise system with frequent backups and staff to restore them.
Archival records can usually tolerate a next-business-day RTO, which is roughly the retrieval turnaround offsite storage providers offer. Paper raises a different RPO question. The original is often the only copy, so the risk is losing it to fire or water, and that’s an argument for a storage facility with climate control and fire suppression over a basement.
For on-premise digital records, keep at least one backup offline. The Cybersecurity and Infrastructure Security Agency (CISA) recommends offline, encrypted backups, because many ransomware variants hunt for reachable backups and encrypt or delete them.
Compliance And Retention In Hybrid Cloud Storage
The first document to request from a cloud vendor is its SOC 2 report, an independent auditor’s examination of the vendor’s controls for security, availability, processing integrity, confidentiality, or privacy, under the American Institute of Certified Public Accountants’ (AICPA) SOC 2 reporting guide. A Type 2 report is the one to ask for, since it tests whether those controls worked over a review period instead of on a single date.
Encryption at rest protects stored files, and encryption in transit protects them as they move between your network, the cloud, and any provider. Both should be on by default for any tier holding personal or regulated information, and a provider should be able to explain how it handles the security of cloud document storage beyond encryption.
Ask, too, whether the encryption modules are validated under Federal Information Processing Standard (FIPS) 140-3. NIST’s validation program moved FIPS 140-2 modules to its historical list after September 21, 2026.
Some records come with format rules of their own. Financial firms that operate as broker-dealers keep certain electronic records under SEC Rule 17a-4, which since May 2023 accepts either a non-rewriteable, non-erasable format or a complete audit trail. Check whether any class in your inventory has a regulator-specified format before you assign it a tier.
The retention schedule doesn’t change with the medium. A personnel file has the same retention period whether it’s a scanned image or a folder in a box, and record retention guidelines by state apply to both. A hybrid build needs one schedule that covers both tiers and one index that shows where each record sits.
Legal holds need the same reach. Once litigation is reasonably foreseeable, courts expect you to preserve relevant records in any format. For electronic records, the federal rules let a court step in when a party’s failure to take reasonable steps to preserve them causes a loss that can’t be made up. A hold that freezes cloud deletion but misses the boxes scheduled for destruction next month is incomplete.
Paper Records and the Scanning Backlog
Physical records belong in a hybrid model, and in many inherited archives they’re the largest tier. The usual approach to a hybrid records system is to index every box, scan the active classes in full, and keep everything else offsite until someone asks for it.
A stalled backlog gets the same treatment. Day-forward scanning, which digitizes everything that arrives after a set date, stops the backlog from growing. The older paper then moves offsite, indexed at the box or folder level, where many providers in our network offer scan on demand. If the framework is pointing you toward a split like this, compare offsite storage options for the archive tier before you commit budget to scanning the full archive.
Running The Hybrid Cloud Storage Framework
Here’s how the framework tends to play out for an IT team that has inherited a mixed archive and no written storage policy.
- Inventory and classify. List every system, share, and storage room, then sort the contents into active, semi-active, and archival classes, noting any legal holds.
- Set targets. Assign an RTO and RPO to each class, with input from the departments that use the records.
- Price the cost categories. Put each class on each architecture and compare the line items, including egress, archive retrieval, minimum storage periods, and staff time.
- Map compliance. Note which classes need SOC 2-attested vendors, encryption, a regulator-specified format, or an air-gapped system.
- Assign tiers. Put each class on the tier that meets its targets at the lowest total cost, and write the result into your records policy.
For most mixed archives, that process lands on a hybrid build. Active records go to the cloud, usually managed through a document management system that applies the retention schedule. Any system that regulation keeps air-gapped remains on-premise. Decades of archival paper move to offsite storage, indexed and available for scan on demand.
What Changes After Go-Live
Retrieval time is the most visible change. Before the build, finding a file from 2009 might mean someone walking to a storage room and opening boxes until it turns up. After it, active records come up in a search, and the provider delivers an archived file as a scan, often the same or next business day. With one retention schedule and one index across both tiers, a legal hold or a disposal date applies everywhere the record exists.
When you’re ready to price the build, you can get your free quote on cloud records storage and compare providers tier by tier.
FAQs on Hybrid Cloud Storage
Hybrid storage fits when your records have very different access patterns, as most archives more than a few years old do. Active files benefit from cloud speed, while archival paper usually costs less to keep in boxes than to scan and store as images nobody may ever open.
It depends on how you use the data. Cloud usually costs less for active records with unpredictable growth, while on-premise can cost less for large, stable data sets you already have the hardware and staff to run. Egress and archive-retrieval fees can shift that math, so price them before you decide.
Look for a current SOC 2 Type 2 report, encryption at rest and in transit, access controls with audit logs, and the ability to apply retention rules and legal holds. For physical storage, look for a provider certified under PRISM Privacy+, the International Secure Information Governance & Management Association’s (i-SIGMA) standard for records and information management.
Yes. In most hybrid builds for records, paper is its own tier, held offsite, indexed, and scanned on request, so it falls under the same retention schedule as your digital files.
Egress fees grow with how much data you pull out of a cloud and how often, so a record class you retrieve constantly can cost more than its storage fees suggest. Records that rarely leave storage, and paper archives that never enter the cloud at all, keep egress low.
How We Match Your Records With Storage Providers
Record Nations connects you with providers for each tier of a hybrid build, from document scanning for active and day-forward files to cloud hosting and offsite storage with scan on demand for the archive. Providers in our network serve organizations nationally, from Spokane, WA to Knoxville, TN, and beyond.
Record Nations is an i-SIGMA member, and we work closely with i-SIGMA to ensure our network providers meet its PRISM Privacy+ standards for storage and information management. Answers to common questions about pickup, pricing, and access are in the Record Nations FAQ.
Fill out our form or call (866) 385-3706, and we’ll match each record class with providers who can quote it. You’ll often receive multiple competitive quotes within minutes or the hour, and generally by day’s end, depending on how integrated each provider is.
