Toshiba has optimized the RocksDB for SSD so as to improve the storage endurance, utilization and reduce the application write amplification. And in the following part, we will show more detailed information.
Toshiba Optimizes the RocksDB for SSD
Toshiba is a very famous manufacturer around the world deserving itself to researching and developing all kinds of storage devices, such as external hard drive, flash memory products, solid-state drives, etc and other electronic products. For example, it has announced the BG4 NVMe Client SSD.
And recently, Toshiba has optimized the RocksDB for SSD. Then you may wonder what the RocksDB is? At here, we will show you a brief introduction.
What Is RocksDB for SSD?
RocksDB is a high performance embedded database for key-value data. It is a branch of the LevelIDB by Facebook optimized to exploit many central processing unit cores so as to make efficient use of fast storage, such as solid-state drive, for input or output workloads.
The RocksDB for SSD is based on a log-structured merge-tree data structure. It is written in C++. And this RocksDB for SSD is widely used in production systems at various web-scale enterprises. And for more information about RocksDB for SSD, you can click here to know more.
Toshiba Optimizes RocksDB for SSD
Toshiba optimizes the RocksDB for SSD can effectively improve the write application and endurance. TRocksDB is an enhancement to the RocksDB database design using key values more efficiently with SSDs to enable improvements in storage and DRAM usage. In addition, the TRocksDB has also reduced the repeated data rewriting caused by application-generated WA.
The TRocksDB solution will be shown publicly for the first time at Flash Memory Summit.
The RocksDB takes a long-structure merge-tree approach to storing data and its platform rewrites data at least one time for every level of database, and in many cases, multiple times per level. Hence, the total WA for RocksDB will often be greater than 21x, result in application-level performance delays and early SSD wear-out.
However, Toshiba has addressed this problem. With the key architectural advancements of TRocksDB, it enables values and keys to be stored in separately managed files. TRocksDB has the ability to separate keys from values produces faster, more efficient database lookups, minimizes WA and optimizes SSD utilization.
Peter Zaitsev, CEO, Percona has said “TRocksDB solves important customer application problems encountered in real-world key value database implementations that will improve SSD use and enable a wider variety of capabilities, extending the overall life of flash-based database storage.”
The server software included in the TRocksDB platform will soon be available under the terms of Apache 22 open-sourced licensing. TRocksDB runs on any Linux hardware supported by RocksDB. If you are a member of the OSS community, the TRocksDB can be a part of storage innovation and industry participation that can evolve the platform to new frontiers.
With the technology of RocksDB, the performance of Toshiba SSD may be improved and the endurance will be increased.
To sum up, Toshiba has optimized the RocksDB for SSD to improve the storage endurance and utilization, reduce applications write amplification by up to 21x. And Toshiba also takes an important new step in expanding our role as a cloud storage technology leader by contributing to the OSS community and evolving SSD capabilities in database applications.