lfs-bg-2

LizardFS Software Defined Storage is a distributed, parallel, scalable,
fault-tolerant, Geo-Redundant and highly available file system.

What LizardFS gives you

Distribution

Distribution means that all the data is distributed among multiple chunk servers. You can easily add and remove drives or the whole storage nodes. It is as simple as “plug and scale”.

Replication

Despite single replica (goal replication) we added Erasure Coding which enables you to save disc space and allows parallel writes and reads to multiple chunkservers for increased performance.

Scalability

You can easily scale your LizardFS cluster up and down both vertically and horizontally by a single drive or a single node. You can have thousands of chunkservers and up to 1 Exabyte of data.

Ease of Installation and Administration

If you have good knowledge of LINUX/UNIX systems you shouldn’t need more than 2 hours to get your cluster up and running (our current record is 28 minutes – let us know if you beat it).

Hardware Flexibility

Digging deeper there are standard LINUX/UNIX operating systems (you can even mix them up in one installation) – so you can use any hardware that supports any of this kind of the systems.

Containers

Lizardfs can be containerized, which means you can run chunkservers within LXC and Docker. That allows for servers that have low I/O requirements to be utilized as chunkservers at the same time.

Fine Grained Policies

Policies and goals can be applied per namespace, per directory as well as per file giving you endless options to configure where your data will be located and how it will be replicated within your LizardFS installation.

Metadata

Metadata is secured in multiple layers via HA setups of the metadataservers and additionally via multiple metadata loggers who can be utilized to rebuild a metadataserver in minimal time decreasing RTO and RPO significantly.

Performance

You can juggle with architecture configuration and hardware components to adjust desirable performance or cost per TB. You can even set few tiers of your storage by defining a group of faster and slower chunkservers.

Below LizardFS are lying classic LINUX or UNIX file systems so you can use all the performance options offered by modern FS’s (e.g. ZFS):

  • Speed up from SSD’s
  • Speed up from NvME
  • Automatic handling of tape drives
  • Tiering between groups of differently configured chunkservers

Multi Data Center support

Geo replication allows you to create geo stretched clusters stretching virtually any distance.

© Copyright - Skytechnology Sp. z o.o. - LizardFS