The following section covers the simplest way to start using Speedb:
- Speedb Example - the fastest way to get Speedb up and running
- How to compile Speedb from source
- Use prebuilt binaries
- Drop-in Replacement for existing Rocksdb users that would like to start using Speedb with minimum steps
Speedb OSS Features
Speedn Open-source can be found in GitHub, rebasing periodically on RocksDB and includes unique features to improve performance and application stability, enhance resource utilization and better operational usability. Including:
- Live Configuration Changes - Live Configuration Changes allow a user to change the mutable options of a Speedb database without restarting the process. Mutable options allow the user to change certain behavior without closing the database.
- Dynamic Delayed Writes - Improved performance stabilization especially in an environment with multiple databases and/or multiple column families.
- Sorted Hash Memtable - New memtable type that improves the performance of reads and writes while maintaining seek performance.
- Paired Bloom Filter - The Paired Bloom Filter reduces the false positive rates of FastLocalBloomFilterPolicy, RockDB’s default cache-local bloom filter, while consuming the same amount of memory and processing power.
- Reduce switch memtable latency - provides better and more stable write performance. create a switch background thread responsible for constructing a new memtable object to be used when switch occurs. A new memtable will be created when we fetch this memtable, so we'll always have one memtable available for use
- Report Index Size per Column Family - enable smarter resource allocation decisions. Provides a report about the size of indexes and filters in the block cache.
To learn more about Speedb OSS features, check our documentation
The Story of Speedb
Finding the root cause of typical data performance problems
Eleven years ago, Google identified a need for a novel architecture to serve its extensive operations, but traditional structured databases were inadequate for the purpose. This led to the creation of LevelDB, a well-known open-source Key-value storage engine, which was then forked by Facebook and rechristened as RocksDB to meet their expansive requirements. Before long, the open-source version was made available to the public, leading to a surge in its adoption, with millions of installations worldwide. However, what was left unnoticed was that developers lacked the necessary expertise to properly use and configure this storage engine for the application, thus compromising the performance, scalability, and stability of the software, resulting in time and effort being invested in the wrong direction.
It was from this realization that Speedb OSS emerged.
We aimed to provide a revolutionary storage engine, built and perfected by the community, that was both highly versatile and easy to configure for meeting the data needs of modern applications. Unlike the preceding generations, Speedb, the third-generation key-value storage engine, boasts greater proficiency in data and metadata management, while the former focused mostly on data alone. As such, Speedb serves as an easy replacement for RocksDB in existing applications, making it incredibly worthwhile in terms of performance gains and minimizing integration time