Leading tech firm DapuStor has joined forces with Pliops to provide a new alternative to the industry standard RocksDB storage engine, more than doubling the performance and significantly improving the quality of service.
KVrocks serves as the cornerstone for the critical operations of renowned tech companies such as Baidu, Inc. , Trip.com Group Limited, Meitu Inc., RGYUN, U-NEXT, and Baishan Cloud. This open-source, distributed key-value NoSQL database is adept at managing and processing large-scale datasets. Designed as a durable alternative to Redis, KVrocks employs RocksDB for storage, maintaining compatibility with the Redis protocol. Its prowess as a persistent key-value store translates into substantial savings on memory costs for handling extensive datasets.
Nevertheless, when data volumes escalate to petabyte levels, adhering to customer Service Level Agreements (SLAs) poses a challenge, primarily due to the performance and latency constraints of the RocksDB storage engine. To surmount these hurdles, strategic optimisation of the infrastructure and boosting operational efficiency are essential. These measures not only curtail the costs associated with data storage but also elevate overall customer experience by ensuring responsiveness and reliability.
The DapuStor & Pliops solution
Technology leaders, DapuStor and Pliops have joined forces to advance KVRocks with superior key-value storage, aiming to exceed customer expectations in performance, scalability, and QoS. The integration of DapuStor’s SSDs with Pliops’ AccelKV offers unmatched benefits for KVRocks clientele.
DapuStor stands at the forefront of enterprise and data center SSDs storage solutions, boasting a robust portfolio in ASIC controller chip technology. Its products, renowned for their reliability and performance, are pivotal in the telecommunications, service provider, and data centre industries.
"The DapuStor R5101 is a high-performance PCIe4.0 NVMe SSD that offer superior reliability, low latency, and excellent power efficiency, optimising TCO for enterprise IT and cloud facilities. Ideal for core data storage across various sectors like logistics, finance, intelligent manufacturing, and AI, it achieves 1,750 randread and 240 randwrite IOPS. Combined with Pliops’ XDP and library, the solution delivers a high-performance key-value storage interface,” says Grant Li, DapuStor’s solution architect.
Pliops XDP is a PCIe-based accelerator that seamlessly integrates with servers and SSDs, featuring a sleek design for rapid deployment. It’s universally compatible, ensuring swift operability. The XDP-Rocks library extends RocksDB functionalities, optimizing KVrocks integration with features like efficient data manipulation and advanced dataset scanning, essential for robust data retrieval.
“At Pliops, we lead the industry in providing efficient and intelligent data transport solutions. Leveraging our innovative Key-Value technology, our products are designed to accelerate critical AI applications on both GPU and CPU platforms,” says Tony Afshary, Global VP of Products and Marketing at Pliops. “We’re thrilled to partner with DapuStore and accelerate their best-in-class NVMe storage solutions. Our customers will be delighted when deploying these integrated products.”
DapuStor solution offers twice the performance of industry standard
To show the power of the new solution, the Dbbench database benchmark tool was used to compare the Pliops Accel KV performance for KVRocks against RocksDB. This assessment involved a substantial dataset comprising 1.2 billion objects, which provided valuable insights into the systems’ capabilities. The findings revealed that Pliops KV outperforms traditional RocksDB by a factor of two, demonstrating double the performance. Additionally, it offers enhanced scalability for users and improved system efficiency for KV Rocks Applications.
In this rigorous testing environment, the DapuStor R5101 SSDs were used in conjunction with the Pliops XDP Accel KV. This combination was instrumental in pushing the systems to their maximum operational capacities. The results of this evaluation indicate that Pliops KV stands out as a viable and superior alternative to conventional RocksDB implementations across various applications, particularly KV-Rocks.
One of the most significant advantages of Pliops KV is its impact on Quality of Service (QoS). Tail latency drops between 66% and 87% in the mixed workload with 50% read and 50% write, and 10% and 28% in the workload with 70% read and 30% write. It is especially beneficial in scenarios where QoS is paramount for delivering an exceptional customer experience. This is attributed to the considerable reduction in latency that Pliops KV facilitates, ensuring swift and reliable access to data.
“By utilising existing infrastructure and reducing the need for higher memory capacity, this solution can offer a cost-effective solution for large-scale data management, which brings a 53% reduction in price performance due to performance acceleration,“ says Grant Li of DapuStor.
Please contact Harry O'Neill at Insight Media at harry@insightm.co.uk to arrange interviews with the DapuStor technical team or for any further information.