Levyx Adds a RocksDB Integration That Extends its High-Performance Attributes to Many of the World’s Largest Big Data Application Installations

IRVINE, Calif., Aug. 22, 2017 -- Levyx Inc., whose high-performance, ultra-low latency data processing software dramatically reduces Big Data infrastructure complexity and costs, announced results today highlighting that its RocksDB implementation (HeRocks™) improves the overall efficiency of conventional (open-source) RocksDB by up to 15X.  HeRocks solution is ideal for data center environments that run RocksDB to process extensive key-value sets (billion-plus) in real-time.  A free version of the software can be easily installed with just a few clicks to non-disruptively improve the underlying RocksDB deployment by contacting [email protected].

Reza Sadri, CEO of Levyx, said: “Our optimizations are important because they come at a time when we see accelerating adoption of RocksDB in our key markets by major enterprises. In fact, the larger the conventional deployment of RocksDB, the better our value proposition and market opportunity. We estimate that high-performance data centers like those supporting the major social networking and e-commerce sites such as Facebook, Yahoo! and Google can save on the order of tens of millions of dollars annually by running their peta-scale applications using our software as an embedded database. The best part is that they can incorporate our solution seamlessly within their existing infrastructure.”

Using db_bench, RocksDB’s own benchmarking tool, and running on a system with 88 cores and Intel’s NVMe DCP3700 SSDs, HeRocks test results were more than an order of magnitude better than the latest open-source version of RocksDB. Value sizes of 100 bytes were used, representing workloads such as system logs or network logs for data centers and cybersecurity applications; stock quotes for high-frequency trading and analysis; Facebook status updates; or Twitter messages. For a mixed load of 70% read/ 30% write, RocksDB achieved 330K operations/second (ops/sec) and HeRocks achieved 5.1 million ops/sec. For random reads, RocksDB topped at 1.3 million ops/sec while HeRocks ran at 16 million ops/sec. For “Read while Writing,” RocksDB was at 2.0 million ops/sec, and HeRocks was at 16.4 million ops/sec. 

Levyx software technologies are hardware-independent and serve as enhancements to many of the world’s most popular Big Data platforms including RocksDB, Apache Spark, LevelDB, and Memcached. Levyx’s Helium™ engine is one of the world's fastest key value stores and is optimized for flash, storage class memory, and multi-core systems. Due to its highly embeddable characteristics, Levyx has integrated Helium into these prolific Big Data platforms as a high-performance core data processing engine.

RocksDB was begun at Facebook in 2012 and is a popular high-performance embedded database for key-value data. It is derived from LevelDB and was optimized for input/output (I/O) bound workloads. Based on a log-structured merge-tree (LSM tree) data structure, RocksDB is open-source software written in C++ that provides official application programming interface (API) language bindings for C++, C, and Java.

About Levyx Inc.
Levyx’s software solutions fundamentally disrupt the economics of Big-Data applications, bringing the benefits of high-speed Big-Data processing to all enterprises and commodity hardware. Optimized for all of the latest hardware innovations, Levyx’s products make running large-scale workloads in real-time far more efficient than conventional methods and brings all of the enterprise features required by the most demanding datacenter environments to those interactive data sets. More information is available at www.levyx.com.

Levyx, the Levyx logo, Helium, and HeRocks are trademarks of Levyx Inc.  All other trademarks or brand names referred in this press release are the property of their respective owners.

Levyx Press Contact:
Luis Morales
(949) 466-2738
[email protected]