Fast, efficient backup
Deduplication engine
Backups of large applications without pushing tens of gigabytes for every small change. Our engine sends only the data blocks that actually changed.
Where standard backup can't keep up
Databases and business systems store data in large files. A small change is enough for a classic backup to transfer the whole file again.
Whole files for every change
Changing a few database records means a new copy of the entire file.
A flooded network
Tens of gigabytes a day slow down the network and people's work.
Long backup windows
Backups don't finish overnight and spill into the working day.
Fast capacity growth
Every version takes full space, so storage fills up quickly.
Our own block-level deduplication engine
Instead of whole files we work with small data blocks. Each block is stored only once, and the next backup transfers only the new ones.
Enterprise storage such as NetApp ONTAP and Dell PowerStore is built on the same block-level deduplication principle - without the enterprise price tag.
- 01
Splitting
Data is split into small blocks.
- 02
Fingerprint
Each block gets a unique fingerprint (hash).
- 03
Comparison
Fingerprints are compared with blocks we already store.
- 04
Transfer changes
Only new or changed blocks travel over the network.
- 05
Backup version
A complete, restorable version is assembled from blocks.
A 10 GB file, a few blocks changed
The file is split into thousands of small blocks. When a few database records change, only the blocks containing those records change. Only those are backed up - the rest of the file is already in the backup.
What you get
Illustrative example
A 10 GB database file in which about 1 % of the data changes during the day.
Actual savings depend on the data type and the extent of changes.
- Only changed blocks are transferred, not whole files
- Shorter backup windows and less network load
- Storage savings even with many versions
- More frequent backups without affecting operations
Technical details
How the engine ensures only what actually changed is transferred - and that data can always be fully restored.
- Variable-length blocks
- Block boundaries are set by the data content (content-defined chunking), not by fixed positions. If a few bytes are inserted at the start of a file, only the affected block changes - the rest of the file doesn't shift and doesn't need to be sent again.
- Cryptographic fingerprint
- Each block gets a fingerprint (hash, e.g. SHA-256). Identical content always has the same fingerprint, so a block is stored only once - even if it repeats in other files or versions.
- Block index
- The engine keeps an index of fingerprints for all stored blocks. On the next backup it compares fingerprints first and sends only the blocks missing from the index.
- Compression and encryption
- New blocks are compressed and encrypted before they are sent. Transfers are even smaller and data is protected in transit and at rest.
- Versions without full backups
- Every backup is a complete version built from block references (incremental forever). Periodic full backups aren't needed, yet you can restore a file from any day.
- Integrity checks
- Each block's fingerprint is verified during restores and regular checks. A damaged block is detected long before you'd need the backup.
Where we deploy the engine
It delivers the most where large files change only in small parts.
Databases
SQL and other database files that grow and change constantly.
ERP and accounting
Business systems with large data files and frequent changes.
Virtual disks
Virtual server images where only a fraction of the content changes.
Branches on slow links
Backup to headquarters or off-site without flooding the network.
Which package includes it
The deduplication engine is part of the Ransomware Shield Pro and Smart Processor packages.
Data Shield
Everything Synology offers
For businesses that want their data safely backed up and the NIS2 basics covered.
View packageRansomware Shield Pro
Active attack protection
For businesses that cannot afford downtime or data loss.
View packageSmart Processor
Protection + custom automation
For businesses that want to automate routine work on top of protecting their data.
View packageFAQ
Why a custom engine when Synology already deduplicates?
For everyday data we use the built-in tools. We deploy our own engine for specific applications where large files change only in small parts and standard backup would needlessly transfer whole files.
Can the engine be added to our existing backup?
In most cases, yes. We first review which applications and data you back up and recommend where the engine brings the most benefit.
Is a restore from a deduplicated backup complete?
Yes. Every backup version consists of a complete set of blocks, so you restore the whole file as it was at that point in time. We verify recovery regularly with tests.
Do your backups take too long?
We'll look at your applications and data and show how much time and capacity you can save.
Contact us