RoadMap

v0.1.0

  • Standalone version, local storage
  • Analytical storage format
  • Support SQL

v0.2.0

  • Distributed version supports static topology defined in config file.
  • The underlying storage supports Aliyun OSS.
  • WAL implementation based on OBKV.

v0.3.0

  • Release multi-language clients, including Java, Rust and Python.
  • Static cluster mode with HoraeMeta.
  • Basic implementation of hybrid storage format.

v0.4.0

  • Implement more sophisticated cluster solution that enhances reliability and scalability of HoraeDB.
  • Set up nightly benchmark with TSBS.

v1.0.0-alpha (Released)

  • Implement Distributed WAL based on Apache Kafka.
  • Release Golang client.
  • Improve the query performance for classic time series workloads.
  • Support dynamic migration of tables in cluster mode.

v1.0.0

  • Formally release HoraeDB and its SDKs with all breaking changes finished.
  • Finish the majority of work related to Table Partitioning.
  • Various efforts to improve query performance, especially for cloud-native cluster mode. These works include:
    • Multi-tier cache.
    • Introduce various methods to reduce the data fetched from remote storage (improve the accuracy of SST data filtering).
    • Increase the parallelism while fetching data from remote object-store.
  • Improve data ingestion performance by introducing resource control over compaction.

Afterwards

With an in-depth understanding of the time-series database and its various use cases, the majority of our work will focus on performance, reliability, scalability, ease of use, and collaborations with open-source communities.

  • Add utilities that support PromQL, InfluxQL, OpenTSDB protocol, and so on.
  • Provide basic utilities for operation and maintenance. Specifically, the following are included:
    • Deployment tools that fit well for cloud infrastructures like Kubernetes.
    • Enhance self-observability, especially critical logs and metrics should be supplemented.
  • Develop various tools that ease the use of HoraeDB. For example, data import and export tools.
  • Explore new storage formats that will improve performance on hybrid workloads (analytical and time-series workloads).