“WANdisco Fusion provides functionality and availability not possible with Cloudera BDR, and avoids BDR’s risk of data loss when outages occur.”
|FEATURE||WANDISCO FUSION||CLOUDERA BDR||COMMENTS|
|TRANSFER DATA BETWEEN CLUSTERS||
Fusion uses patented active transactional replication to continuously transfer data between clusters running on any Hadoop distribution. BDR is built on MapReduce and data is transferred in batch at pre-scheduled intervals to other clusters running on the same major release of CDH.
Fusion’s active-transactional replication means data is continuously available, with automated forward recovery for clusters that go offline. BDR is batch, requires manual recovery, and the entire BDR job has to be rerun after an outage.
Both products allow selected HDFS files and folders to be copied between clusters but BDR is limited to clusters running the same major release of CDH.
|FULL READ/WRITE ON ALL CLUSTERS||
Fusion replicates every change to the same files at multiple locations. BDR is a one-way, time-based (batch) solution that requires target clusters to be read-only to avoid divergence.
|GUARANTEED DATA CONSISTENCY||
Fusion delivers guaranteed consistency and has built in recovery to automatically resynchronize clusters after an outage.
|INGEST & REPLICATE SIMULTANEOUSLY||
Fusion replicates data as it’s ingested-ideal for Spark streaming applications that require replication. BDR requires files to be written and closed before moving data.
|MINIMIZES ATTACK SURFACE EXPOSED TO HACKERS||
Only Fusion servers communicate through a firewall. BDR requires every cluster node to have a firewall port configured increasing the network security administration burden as well as increasing vulnerability to hacking.
BDR only copies data between specific versions of CDH, and does not support migration to non-CDH clusters. Fusion replicates data running on any distribution, version and storage, including cloud storage, and supports migration between platforms with no downtime.
“WANdisco Fusion replicated the same data volumes up to 90% faster than BDR, without impacting the performance of the other applications running on the clusters.”
“Unlike BDR, WANdisco’s replication technology in Fusion enables clusters to be available even when full cluster backups are engaged.”