Start from the recovery time you need, then work backward. Where the newest backup sits, and in what format, is the controlling variable for how fast you can restore. Most IT teams design backup storage around capacity and cost, then discover recovery time problems only during an incident. Putting RTO first and storage tier second fixes the gap before it becomes a crisis.
Key Takeaways
- Storage design, not backup software, sets how fast you can restore after an outage or attack.
- Rehydration from deduplicated storage adds latency that can break RTO targets for time-sensitive systems.
- ExaGrid’s Landing Zone keeps the newest backups undeduplicated, enabling restores and VM boots up to 20x faster than inline deduplication appliances, according to ExaGrid.
- 53% of ransomware victims fully recovered within a week in 2025, up from 35% in 2024 (Sophos State of Ransomware 2025).
- 71% of organizations skip failover testing, so many recovery targets are untested assumptions (Cockroach Labs, State of Resilience 2025).
What Is Storage Backup and Recovery, and Why Does the Storage Layer Control Restore Speed?
Storage data backup and recovery is the combination of where backup data is written, in what format it’s stored, and how quickly it can be retrieved. The backup application handles scheduling, policy, and deduplication logic. The storage target controls what lands on disk and whether it can be read back at full speed without an intermediate step.
Recovery Time Objective (RTO) is the maximum acceptable time to restore a system after failure. Recovery Point Objective (RPO) is the maximum acceptable data loss, measured in time. Both are core inputs to contingency planning in NIST SP 800-34 Rev. 1, the federal contingency planning guide. Both targets are meaningless if the storage layer can’t deliver data at the speed the RTO requires.
Rehydration is the process of reassembling backup data from deduplicated blocks before it can be restored. When the newest backup is stored in deduplicated form, every restore starts with rehydration, and that step adds latency a fast RTO can’t absorb. The storage format on disk is where recovery time is won or lost.
How Do You Classify Systems by Recovery Time and Assign Each to the Right Storage Tier?
Group every system into one of three RTO bands, then assign each band to a storage tier that can physically meet the target. Classify first, then design the storage to match.
Three bands work for most organizations. Systems needing instant VM boot, such as databases, ERP, and core infrastructure, require the newest backup to be undeduplicated and immediately accessible. Same-day systems can tolerate a few hours of restore work. Low-priority or archival systems can wait for next-day or longer recovery.
A Landing Zone is a disk-cache tier that holds the most recent backups in undeduplicated form, ready for instant VM recovery or fast file restores. A Repository Tier is the deduplicated long-term tier, where weekly, monthly, and yearly restore points live at lower cost. Adaptive deduplication runs after backups land, so the Landing Zone stays fast while the Repository Tier stays economical.
The 3-2-1 rule calls for three copies of data, on two different media, with one offsite. That maps to this model: Landing Zone on fast disk, Repository Tier on deduplicated disk, and an offsite or isolated copy for disaster and ransomware scenarios. Federal guidance reinforces the isolated copy: the CISA #StopRansomware Guide recommends offline, encrypted backups that are regularly tested in a disaster recovery scenario.
RTO-to-Storage Tier Planning Framework
| Criterion | Instant Recovery (minutes) | Same-Day Restore (hours) | Archival Restore (days) |
| Newest backup format | Undeduplicated (Landing Zone) | Undeduplicated preferred | Deduplicated acceptable |
| Rehydration required | No | No, if restored from the Landing Zone | Yes, from the Repository Tier |
| Instant VM boot | Required | Not required | Not required |
| Storage tier | Fast disk-cache Landing Zone | Fast disk or Landing Zone | Deduplicated Repository Tier |
| Ransomware exposure | High if only network-facing storage is used | High if only network-facing storage is used | Reduced by a non-network-facing tier |
Protecting Backup Data From Ransomware Without Slowing Recovery
Ransomware remains a constant pressure on backup infrastructure. It was again the most pervasive threat to US critical infrastructure in 2024, with complaints up 9% from 2023, according to the FBI’s 2024 Internet Crime Report. Attackers increasingly take control of the backup application and issue delete commands against backup data. A network-facing storage target with no delayed-delete policy is fully exposed, and the attacker’s delete command can destroy the last recovery path before the IT team knows an attack is underway.
A tiered air gap separates the network-facing Landing Zone from a non-network-facing Repository Tier. Attackers who compromise the backup application can’t see or access the Repository Tier from the network, so prior backups remain intact even when the Landing Zone is encrypted.
AI-Powered Retention Time-Lock (RTL) delays every delete request in the Repository Tier for a configurable 1 to 30 days, so a threat actor’s delete command doesn’t destroy retention data during that window. At the 10-day default, the policy uses roughly 5% to 10% additional storage.
Auto Detect & Guard learns the pattern of normal operational deletes. When a request falls outside that pattern, it alerts the IT team and extends the delayed-delete policy indefinitely. Clearing the hold requires an administrator and a security officer, each authenticating with two-factor authentication. Immutable data objects in the Repository Tier add another layer: retained objects can’t be changed or overwritten.
Validating the Design: Why an Untested Backup Is an Assumption
A backup that has never been restored is an assumption, not a verified capability, and the gap shows up at the worst possible time. Testing is often skipped: in the Cockroach Labs State of Resilience 2025 survey of 1,000 technology executives, 71% of organizations said they skip failover tests.
Recovery outcomes are improving where preparation is in place. In the Sophos State of Ransomware 2025 report, 53% of organizations fully recovered from a ransomware attack within a week, up from 35% the year before. Yet only 54% used backups to restore their data, the lowest share in six years. Backups only help if they’re intact and can be restored fast enough to matter.
Test each RTO band separately. Run an instant VM boot from the Landing Zone, measure elapsed time, and compare it against the RTO target. Run a full restore from the Repository Tier and record the rehydration time. Document both results so the RTO claim is defensible to auditors.
Schedule tests by system criticality, for example monthly for tier-one systems and quarterly for tier-two. A simple worksheet captures what you need: system name, RTO, RPO, storage tier assignment, last restore test date, and measured restore time. That document is your architecture’s proof of performance.
How to Choose Backup Storage That Matches Your RTOs
- Restore format: Does the storage target hold the most recent backup in undeduplicated form? Restores and instant VM boots that skip rehydration are the most reliable way to meet a tight RTO.
- Growth model: Does the architecture scale out by adding full appliances with compute and disk, rather than disk shelves behind a fixed controller? Scale-out keeps the backup window fixed as data grows.
- Ransomware protection: Is protection architectural? A non-network-facing tier with delayed deletes and immutable objects holds up even if the backup application is compromised.
- Application support: Does the system support the backup applications already in use, such as Veeam, Commvault, Cohesity NetBackup, Rubrik, HYCU, and Oracle RMAN?
- Support model: Does the vendor assign a named engineer who knows your backup applications, with proactive health monitoring that flags issues before they affect a recovery?
- Cost predictability: Do maintenance, support, and long-term retention carry no hidden fees, and can you grow without forklift upgrades?
ExaGrid Tiered Backup Storage meets these criteria directly. The Landing Zone holds recent backups undeduplicated for fast restores and VM boots. The Repository Tier holds long-term retention as deduplicated, immutable objects in a non-network-facing tiered air gap. The scale-out architecture keeps the backup window fixed up to a 6PB full backup on HDD appliances or 17.3PB on SSD appliances in a single system. Every ExaGrid customer is assigned a senior Level 2 support engineer matched to the backup applications they run, with no Level 1 queue in between.
Frequently Asked Questions
What is the difference between ExaGrid Tiered Backup Storage and a traditional deduplication appliance?
A traditional deduplication appliance deduplicates data as it arrives, so every restore requires rehydration, reassembling data from deduplicated blocks before it can be read. ExaGrid writes the newest backups to an undeduplicated Landing Zone first, then deduplicates into the Repository Tier. Restores and VM boots run from the Landing Zone without a rehydration step, which ExaGrid states makes them up to 20x faster than inline deduplication appliances.
What happens if ExaGrid hardware fails?
Each ExaGrid appliance includes RAID 6 with a spare drive and dual power supplies, so a single drive or power supply failure doesn’t stop the system. Replacement components are hot-swappable while the system stays in production, and ExaGrid ships replacements from spares depots worldwide. Because each appliance in a scale-out system has its own processor, memory, and disk, there’s no single shared controller for the whole system.
How do I calculate the storage I need for a 30-day retention policy with daily incrementals and weekly fulls?
Start with the size of one full backup. Add the daily incremental size multiplied by the number of daily backups in the retention window, then add the weekly fulls you keep. Apply the deduplication ratio your data actually achieves, which varies widely by data type and retention length. In a tiered design, size the Landing Zone for recent undeduplicated backups and the Repository Tier for deduplicated retention.
How long does it take to install ExaGrid in an existing Veeam or Commvault environment?
Customers rack the appliances and complete setup with their assigned ExaGrid support engineer by phone or web session, at no charge. A typical installation takes 30 minutes to 3 hours, depending on the environment. You then add ExaGrid as a backup target in Veeam, Commvault, or any of the more than 25 supported applications and point your jobs at it.
How do I know if my current backup storage can meet my RTO?
Run a restore test and measure elapsed time. If your newest backup is stored in deduplicated form, measure the rehydration step separately, since that’s time your RTO is absorbing that most teams don’t account for. If measured restore time exceeds the documented RTO for any system class, the storage tier assignment needs to change. An untested RTO isn’t a target. It’s a guess.
- How B2B SaaS Platforms Are Using AI Model Routing to Stop Paying for Intelligence They Don’t Need - October 23, 2026
- Storage Backup and Recovery: How to Design Backup Storage Around Your RTO - October 6, 2026
- The Intelligent Building: How AI-Driven HVAC Systems Are Cutting Energy Waste - September 5, 2026