Infinit alpha

A Resilient and Scalable Platform


Node A
Node B
Node C
Node D
Node E
Node F
Node G

In this scenario, an Infinit cluster composed of 5 nodes maintains 3 redundant copies of a given piece of data.

PlayPlay Demo ReplayReplay

1
Node C crashes, taking down with it one of the 3 copies of the data it hosted.

2
Infinit ejects the unresponsive node from the cluster as soon as the failure is detected.

3
In order to replace the missing copy, a new replica is created and placed on another node.

4
At some point, two new nodes join the cluster, extending it to 6 nodes.

5
The cluster is rebalanced to take into account the new nodes, moving a replica from Node D to G.





Scalable Model

Unlike most distributed systems relying on a master/slave model (also known as leader/follower), Infinit benefits from a truly decentralized architecture.

This model protects Infinit from single points of failure and bottlenecks while allowing for better scalability and performance.





Resilient Architecture

Every data item can be stored redundantly in order to allow for Infinit to tolerate failures being software bugs, hardware crashes and more.

Should a failure occur, Infinit will automatically self-heal by recreating the missing redundant copies in order to quickly get back to a healthy state.

Customizable Through Policies

Redundancy

Choose the redundancy policy of your choice (replication, erasure coding or none at all) and configure it: replication factor etc.

Encryption

While data in transit is always encrypted, you can decide to activate the encryption policy for every data item to be encrypted at rest (RSA/AES-256).

Compression

Activate compression to reduce storage consumption should the nature of the data being stored be adapted.

Data Placement

Control the placement of the redundant pieces of data through several policies to prevent critical scenarios: rack-aware, region-aware (different data centers) etc.

Deduplication

Provides a way to activate deduplication in order to store identical pieces of data only once, hence further reducing storage consumption.

Versioning

Allow one to browse through the history of a file/directory/object and to rollback to a specific version if necessary.

For Each Use Case its Logic

Block Device

Deploy a low-level virtual block device that can be formatted in the file system of your choice (ext4, NTFS etc.) and used for a single node to access it.

File System

A POSIX-compliant hierarchical organization of files and folders that can be concurrently accessed from multiple client nodes.

Object Storage

Use an object storage interface (Amazon S3, OpenStack Swift etc.) to access and store large amounts of unstructured data a.k.a blobs.

Check out

Bringing Persistent Storage to Docker

Deploy your stateful and/or legacy applications as easily as your modern stateless ones thanks to Infinit's elastic storage platform.

Get Started

Deploy your first Infinit architecture

Follow a simple step-by-step guide to set up your first infrastructure for a single user with two devices.