Features
»Replication »POSIX »Security »Cloud File System »Elasticity »Extensibility »More...


Replication

All XtreemFS services can be replicated with hot stand-bys which automatically take over when the primary server fails. With XtreemFS you can create a highly available file system installation without a single point of failure (SPOF-free).



File Replication

XtreemFS replicates your file data across multiple storage servers, which can be distributed worldwide. Our replication algorithm is designed to cope with all the problems and failure scenarios that can occur in truly distributed systems: message loss, network partitions, server crashes etc. For each replicated file, XtreemFS keeps a list of replicas which gives you complete control where to place replicas. You can even use different placement schemes for each file. When a replicated file is opened, XtreemFS automatically elects a primary replica for that file which processes all file operations. If the primary fails, one of the backup replicas will automatically take over after a short failover period. Your applications won't see a difference between regular and replicated file because the XtreemFS file replication is completeley transparent. In the video to the right you can see XtreemFS file replication and automatic failover in action. We used an unmodified mplayer showing a video stored on XtreemFS.
Find out how the file replication in XtreemFS works »

Read-Only Replication

In addition to regular file replication, XtreemFS provides read-only replication. This replication mode works on immutable files and supports a large number of replicas. The read-only replication helps you quickly build a caching infrastructure on top of XtreemFS in order to reduce latency and bandwidth consumption between datacenters. In addition to full replicas that contain a complete copy, XtreemFS also supports partial replicas. These replicas are filled on demand when a client accesses data.

Metadata Replication

XtreemFS can also replicate the Metadata (MRC) and Directory (DIR) services. Like file replication, these services are replicated with a primary/backup scheme. One of the replicas is elected as the primary and answers all client requests. Should the primary crash or get disconnected, one of the backups automatically takes over. In release 1.3.0, automatic failover is not supported by the client and other services and metadata replication should be considered highly experimental.