The Unigraph Engine:

Graph storage and traversal at lightning speed

Highly distributed, fully concurrent and scalable solution for storing interconnected data. File system and storage medium agnostic. Optimized for Flash with extremely low latencies, working with LSM and Btree data structures

engine
Built to store terabytes of data on locally attached Flash drives or in RAM with linear CPU scaling
Embedded key-value store, optimized for fast storage and high throughput. Fully exploiting the high read/write rates offered by Flash and RAM subsystems

Flexible configuration settings for workload optimization on a variety of production environments including pure memory, Flash and hard disks. Keys and values are stored as pure byte strings, without size limitation. Various compression algorithms are supported.

Storage system agnostic

RAM, SSD or HDD – The Unigraph Engine works on any medium and file system

Immediate availability

After import, your data is immediately available. No warm up time required, no indexes needed.

Highly distributed

From 1 to 100 nodes across datacenters. No problem. Reshaping data has never been so easy

Concurrent reads and writes

Concurrent reads and writes and fault tolerance provides support for real-time applications

Storage concurrency

In LSM, background processes like flush and compaction execute concurrently via threads

Optimized encoding

Very fast and space optimized data encoding and decoding: (strings, floats, ints, etc.)
0
x
Faster access
0
x
Optimized storage
0
x
Less RAM needed
0
ms
Warmup time

Build your own, dedicated Knowledge Graph by combining private and public data

Can the Unigraph Engine run on my own infrastructure?

Yes. If you have lots of data that you want to be transferred into a graph structure, the Unigraph Engine is the ideal solution. All you have to do is provide us with root access to a dedicated server on which we’ll install and configure the engine.

Please note: Your data will be completely isolated from the data in Unigraph.

I have private data. Can I merge it with Unigraph's?

Yes. The multigraph architecture allows us to create a copy of Unigraph on a dedicated system and merge (reconcile) it with your existing data. We’ll take care of everything: Infrastructure set up, API, data reconcilliation, etc.

Please note: Your private data and the Unigraph instance have to physically reside on your own infrastructure.

I don't have a dedicated infrastructure. Can you help me?

Yes. We’ll setup a cluster of dedicated servers in the OVH global network. The required servers depend on the amount of data that will be stored and wether private data will be reconciled with Unigraph’s. There are 5 datacenters located in France and Canada to choose from. We cover the entire process: Cluster management and monitoring, API availability, system updates and upgrades.

How do I import my data into the Unigraph Engine?

The import feature currently supports the following formats: .XLSX, .XLS, .CSV & .TSV. You have to annotate your datasets and describe the relations between the nodes. Don’t worry – even the most complex graphs take less than 15 min to annotate.

The importer also provides several very powerful data cleaning features, to help you clean your data during import, if need be. For example: Text processing (Split, Merge, Regex, Lowercase, Capitalize…), on the fly language detection and powerful conditional statements.

Of course, we can assist you and guide you through the process.

Will my dedicated graph be updated with new Unigraph data?

Yes. Updates, upgrades and reconciliation will be taken care automatically for you.

Starting from:

999 €

Per month, VAT excl. excl. Includes the management and monitoring of your dedicated infrastructure.

Ready to build your own graph storage?

On-premise in or in the cloud – we will get you started in no time!