ExaGrid vs HPE StoreOnce: Backup Architecture Compared

ExaGrid and HPE StoreOnce both reduce the cost of keeping backups on disk through deduplication. They differ in when that deduplication happens and how each system grows, and those two choices shape your backup window, your restore speed, and what you pay as data grows.

StoreOnce is a mature inline deduplication appliance family with deep HPE ecosystem integration through its Catalyst protocol. ExaGrid is tiered backup storage that lands backups undeduplicated first and scales out by adding full appliances. This comparison walks through both designs on mechanism, not marketing ratings, so you can decide which fits your environment.

Key Takeaways

• When deduplication runs and how the system scales directly control backup window, restore time, and the cost of growth.
• HPE StoreOnce deduplicates inline, and its Catalyst protocol can deduplicate at the source to cut network traffic.
• ExaGrid‘s Landing Zone keeps recent backups undeduplicated, delivering up to 3x faster backups and up to 20x faster restores than inline deduplication appliances, according to ExaGrid.
• StoreOnce scales up within each appliance, to 576TB usable on the 5720. ExaGrid scales out to 32 appliances in one system.
• Unplanned downtime averages $14,056 per minute, so restore architecture is a direct cost decision.

Why Deduplication Architecture Decides More Than Storage Cost

Deduplication architecture determines backup window duration, restore speed, and long-term growth cost. Price-per-terabyte comparisons don’t predict any of those outcomes. Two design decisions drive all three: whether deduplication runs before or after data lands on disk, and whether the system grows by adding disk to a fixed controller or by adding full compute-and-capacity appliances.

Restore speed carries a real price tag. Unplanned IT downtime now averages $14,056 per minute, rising to $23,750 per minute for large enterprises, according to Enterprise Management Associates research published by BigPanda. Every minute a restore takes during an outage is a minute at that rate.

ExaGrid vs HPE StoreOnce at a Glance

Specifications below reflect current HPE StoreOnce Gen5 models and ExaGrid Tiered Backup Storage as published by each vendor.

Dimension	ExaGrid Tiered Backup Storage	HPE StoreOnce (Gen5)
Deduplication timing	Adaptive: backups land undeduplicated in the Landing Zone, then deduplicate in parallel	Inline on the appliance; Catalyst can also deduplicate at the backup source
Restore path for recent backups	Read directly from the undeduplicated Landing Zone; no rehydration	Reassembled from deduplicated data; the all-flash 7700 targets faster restores
Published backup speed	Up to 3x faster than inline deduplication appliances (ExaGrid)	Up to 70 TB/hr (5720) and up to 300 TB/hr (all-flash 7700) with Catalyst
Growth model	Scale-out: up to 32 appliances in one system, each adding compute and capacity	Scale-up within each appliance: 5720 grows from 144TB to 576TB usable; Cloud Bank extends to object storage
Maximum single-system size	Up to a 6PB full backup (HDD) or 17.3PB (SSD)	Up to 576TB usable locally (5720); up to 1.7PB with Cloud Bank Storage
Ransomware protections	Non-network-facing tier, delayed deletes, immutable objects, AI-based delete detection, S3 Object Lock	MFA, dual authorization, encryption, configurable immutability, Cloud Bank or tape air gap
Backup application support	More than 25 applications and utilities	Catalyst integration with Commvault, NetBackup, Veeam, and others, plus Oracle, SAP HANA, and SQL Server plug-ins
HPE ecosystem integration	Works with HPE servers and storage like any other backup target	Direct backup of HPE Alletra Storage MP B10000 snapshots and HPE SimpliVity integration

How Does the Timing of Deduplication Affect Backup and Restore Speed?

Deduplication timing has the largest measurable effect on both backup and recovery performance, and it decides whether you can meet your Recovery Time Objective (RTO), a core metric in NIST SP 800-34 Rev. 1, the federal contingency planning guide.

HPE StoreOnce deduplicates inline, processing data before it’s written to disk. Its Catalyst protocol lets compatible backup applications deduplicate at the source, so less data crosses the network. HPE rates the StoreOnce 5720 at up to 70 TB/hr and the all-flash 7700 at up to 300 TB/hr with Catalyst. Because data is stored only in deduplicated form, every restore must reassemble the original data from deduplicated blocks, a process called rehydration. Faster hardware, such as the all-flash 7700, shortens rehydration but doesn’t remove the step.

ExaGrid uses adaptive deduplication. Backups write first to a disk-cache Landing Zone at full disk speed, and deduplication into the Repository Tier runs in parallel without slowing ingest. The most recent backups stay in full, undeduplicated form, so restores, instant VM recoveries, and tape copies read directly from the Landing Zone with no rehydration. ExaGrid states this design delivers up to 3x faster backups and up to 20x faster restores than inline deduplication appliances. Older retention points in the Repository Tier are reassembled on restore, as with any deduplicated store.

How Do I Read Raw, Usable, and Effective Capacity Without Being Misled?

Raw capacity is the total physical disk installed. Usable capacity subtracts RAID protection and reserved space. Effective capacity multiplies usable capacity by an assumed deduplication ratio, and that ratio is rarely yours.

Both vendors publish effective-capacity figures. HPE’s StoreOnce documentation, for example, calculates effective capacity using a 60:1 ratio between effective and usable capacity. VM images, databases, and already-compressed or encrypted files each deduplicate at very different rates, so a generic ratio can overstate what you’ll actually store.

Before comparing price per terabyte, ask each vendor to size the system at your specific data type mix, daily change rate, and retention. With ExaGrid, plan two tiers: the Landing Zone sized for recent undeduplicated backups and the Repository Tier sized for deduplicated retention.

How Do Scale-Up and Scale-Out Growth Models Differ?

Scale-up systems add disk capacity behind a fixed controller. Capacity grows, but processor, memory, and bandwidth don’t, so the backup window can lengthen as data grows. When the controller reaches its ceiling, the next step is a larger appliance or an additional system.

HPE StoreOnce scales up within each appliance. The 5720 grows from 144TB to 576TB usable by adding drives in the chassis, and the all-flash 7700 grows from 92TB to 552TB. Cloud Bank Storage extends capacity to object storage, up to 1.7PB for the 5720 and 2.7PB for the 7700, for longer-term retention. Organizations that outgrow a single appliance add more StoreOnce systems.

ExaGrid scales out by adding full appliances, each bringing its own processor, memory, bandwidth, and disk. Because compute grows with data, the backup window stays fixed. A single system supports up to 32 appliances under one interface, handling up to a 6PB full backup on HDD appliances or 17.3PB on SSD appliances. Appliances of different sizes and generations can be mixed, which avoids forklift upgrades and planned product obsolescence.

How Do ExaGrid and HPE StoreOnce Protect Backups from Ransomware?

Backup storage is a prime ransomware target. Ransomware 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. The CISA #StopRansomware Guide advises keeping offline, encrypted backups and testing them regularly, because many ransomware variants hunt for accessible backups to delete or encrypt.

HPE StoreOnce offers multifactor authentication, dual authorization for sensitive changes, encryption in flight and at rest, and configurable immutability to block unauthorized changes or deletion. For an air gap, HPE pairs StoreOnce with Cloud Bank Storage or HPE tape.

ExaGrid builds the air gap into the appliance. Its Repository Tier is non-network-facing, so attackers can’t see or access it. AI-Powered Retention Time-Lock delays every delete request in that tier for a configurable 1 to 30 days, and Auto Detect & Guard extends the delay indefinitely when deletes fall outside normal patterns. Clearing that hold requires both an administrator and a security officer to authenticate with two-factor authentication. ExaGrid also supports S3 Object Lock for Veeam.

Both approaches are credible. The difference is that ExaGrid’s tiered air gap and delayed deletes are part of the primary backup appliance, while StoreOnce’s air gap depends on adding cloud or tape copies.

Backup Application Support and Support Model

HPE StoreOnce works with leading backup applications such as Commvault, NetBackup, and Veeam through Catalyst integration, and HPE offers free Catalyst plug-ins for Oracle, SAP HANA, and Microsoft SQL Server. Its strongest advantage is HPE ecosystem integration, including direct backup of HPE Alletra Storage MP B10000 snapshots and HPE SimpliVity protection.

ExaGrid supports more than 25 backup applications and utilities, including Veeam, Commvault, Cohesity NetBackup, Rubrik, HYCU, and Oracle RMAN. Several applications can write to the same system at once, so mixed environments don’t need separate storage targets.

On support, ExaGrid assigns each customer a senior Level 2 engineer who specializes in two to three backup applications, with no Level 1 tier to get through first. Annual maintenance covers installation assistance, failed-component replacement, and all point and full software releases, and ExaGrid offers five-year price protection. For StoreOnce, support terms depend on the HPE service contract and partner you buy through, so compare the specific coverage you’d receive.

Five Questions to Ask Either Vendor Before You Commit

1. Does deduplication run inline or after the backup lands, and what is the measured backup window at my data volume?
2. Does recovery require rehydration, and what is the measured restore time for a full VM at my data set size?
3. What is effective capacity at my specific data type mix, not at a generic assumed ratio?
4. When my data doubles, do I add capacity to an existing controller, add a new system, or add appliances that join the same system with their own compute?
5. What does annual maintenance include, how is installation handled, and does price protection apply to future purchases?

Which Should You Choose?

Choose HPE StoreOnce if you’re standardized on HPE infrastructure, want Catalyst source-side deduplication and native integration with HPE Alletra and SimpliVity, and your capacity fits within a single appliance’s scale-up range.

Choose ExaGrid if restore speed and a fixed backup window matter most, you expect data to keep growing, you run several backup applications, or you want the ransomware air gap built into the primary backup system rather than added through cloud or tape.

For most enterprises focused on fast recovery and predictable growth, ExaGrid is the stronger fit: its Landing Zone removes rehydration from recent restores, scale-out keeps the backup window fixed through multi-petabyte growth, and its tiered air gap protects retention data without extra infrastructure.

Frequently Asked Questions

Is HPE StoreOnce inline or post-process deduplication?

HPE StoreOnce deduplicates inline, before data is written to disk. With Catalyst, compatible backup applications can also deduplicate at the source so less data crosses the network. Because data is stored only in deduplicated form, restores reassemble data from deduplicated blocks.

How is ExaGrid priced for a system sized to 100TB of full backup data with 30-day retention?

ExaGrid pricing depends on full backup size, daily change rate, retention period, and data type mix, all of which affect Landing Zone and Repository Tier sizing, so there’s no single list price for a 100TB configuration. Annual maintenance is all-inclusive and ExaGrid offers five-year price protection, which makes multi-year costs easier to forecast.

Can ExaGrid appliances from different generations be mixed in the same system?

Yes. ExaGrid’s scale-out architecture lets appliances of different sizes and ages operate together in a single system of up to 32 appliances under one management interface. Adding a newer model doesn’t require replacing existing hardware, which avoids forklift upgrades and planned product obsolescence.

How does ExaGrid’s Retention Time-Lock protect backup data from ransomware deletion?

AI-Powered Retention Time-Lock delays every delete request in the non-network-facing Repository Tier for a configurable 1 to 30 days, so a delete issued by ransomware or an attacker doesn’t execute immediately. Auto Detect & Guard learns normal deletion patterns. Requests outside that pattern trigger an alert and extend the delay until both an administrator and a security officer authenticate with two-factor authentication to clear it.

Does ExaGrid support Veeam and Commvault at the same time?

Yes. ExaGrid supports more than 25 backup applications, including Veeam, Commvault, Cohesity NetBackup, Rubrik, HYCU, and Oracle RMAN, and multiple applications can write to the same ExaGrid system simultaneously.

What does an ExaGrid installation involve?

Customers rack the appliances and complete setup with an ExaGrid support engineer by phone or web session, at no charge. A typical installation takes 30 minutes to 3 hours, depending on the environment. Testing with your own backup and restore jobs is the most reliable way to compare performance against your current storage.