Apache Pegasus is a distributed key-value storage system which is designed to be:
- horizontally scalable: distributed using hash-based partitioning
- strongly consistent: ensured by PacificA consensus protocol
- high-performance: using RocksDB as underlying storage engine
- simple: well-defined, easy-to-use APIs
Pegasus targets to fill the gap between Redis and HBase. As the former is in-memory, low latency, but does not provide a strong-consistency guarantee. And unlike the latter, Pegasus is entirely written in C++ and its write-path relies merely on the local filesystem.
Apart from the performance requirements, we also need a storage system to ensure multiple-level data safety and support fast data migration between data centers, automatic load balancing, and online partition split.
Persistence of data: Each write is replicated three-way to different ReplicaServers before responding to the client. Using PacificA protocol, Pegasus has the ability for strong consistent replication and membership changes.
Automatic load balancing over ReplicaServers: Load balancing is a builtin function of MetaServer, which manages the distribution of replicas. When the cluster is in an inbalance state, the administrator can invoke a simple rebalance command that automatically schedules the replica migration.
Cold Backup: Pegasus supports an extensible backup and restore mechanism to ensure data safety. The location of snapshot could be a distributed filesystem like HDFS or local filesystem. The snapshot storing in the filesystem can be further used for analysis based on pegasus-spark.
Eventually-consistent intra-datacenter replication: This is a feature we called duplication. It allows a change made in the local cluster accesible after a short time period by the remote cluster. It help achieving higher availability of your service and gaining better performance by accessing only local cluster.