Compared with the standard edition, ApsaraDB for HBase Performance-enhanced Edition has the following improvements:
ApsaraDB for HBase Performance-enhanced Edition is deeply optimized for RPCs, memory management, caching, and logging. It uses core technologies such as high-performance data schema, coroutines, merging and submission, and traceable coding to greatly improve the performance of reads and writes. The throughput of ApsaraDB for HBase Performance-enhanced Edition is over seven times higher when compared with HBase Community Edition using the same amount of resources. It can help you reduce 90% of the noise in HBase Community Edition. For more information, see Performance white paper.
ApsaraDB for HBase Performance-enhanced Edition has improved the mean time to repair (MTTR). The fault recovery is over 10 times faster than HBase Community Edition. Based on the "Logging is Storing" and "if Partition then Availability or Consistency, Else Latency or Consistency (PACELC)" theories, a multi-replica architecture is adopted to guarantee the consistency of data at different levels. This allows applications to choose between consistency, availability, latency, and programmability. It provides more disaster recovery solutions for your applications to ensure high availability, such as active zone-redundancy, active geo-redundancy, and three data center consistency. Currently, only single-zone deployment is available. More deployment solutions will be available soon.
Developed with a deeply optimized compression algorithm: Zstandard. It supports automatic compression level tuning and compression dictionaries provided based on sampling data. The highest compression ratio can reach 13, which is over 50% higher when compared with SNAPPY.
Transparent hot and cold data separation
In most scenarios, hot data is closer to users than cold data. ApsaraDB for HBase Performance-enhanced Edition supports automatic hot and cold data separation to help you accelerate hot data consumption without the need to modify your applications. You can store cold data in cost-effective medium such as Object Storage Service (OSS).
Built-in data security and resource isolation features for multi-tenancy. It supports username-password authentication, ACLs, quotas, and resource groups, allowing you to build an all-in-one enterprise-class HBase platform, and improve development efficiency and optimize resource utilization. For more information, see Cluster management.