Huawei, Startup Collaborate on Big Data Object Storage
Chinese networking giant Huawei has collaborated with a University of California at Berkeley spinoff on a big data storage acceleration platform with the goals of boosting the performance of analytics and other applications while extending analytics to a broader customer base.
Huawei said this week it worked with Alluxio Inc., a spinoff from UC-Berkeley’s Algorithms, Machines and People Lab (AMPLab), on the release of a data storage acceleration platform that incorporates the Chinese company’s FusionStorage distributed management system with Alluxio’s memory-boosted virtual distributed storage system.
Alluxio is backed by the venture capital firm Andreessen Horowitz. The startup, formerly know as Tachyon Nexus, was founded in 2015 by Haoyuan Li. Spun out as an open-source project from the AMPLab, Li’s team has focused on new ways to store and access data via a “memory-centric architecture.”
The startup claims to have tackled the problem of moving big data at “memory speed” across different storage systems such as network attached and storage-area networks. The startup focused on the problem of scaling the transfer of object storage by coming up with what it claims is a more efficient way to use its software rather than an object storage gateway to move big data from traditional storage systems to object storage.
The approach reflects the growing trend toward object storage to collect large amounts of data in a single namespace, a step that is seen as a way of integrating applications like Hadoop to improve data analysis. The startup cites active archiving as a prime object storage use case in which little used data remains accessible to applications and users via cheaper storage options
Now, the startup has the deep pockets of the Chinese networking giant behind it as it deploys its “memory-centric” storage approach for big data processing “across multiple industries.”
In a statement, Li noted that the goals of the partnership with Huawei was “popularizing big data” as well as accelerating analytics applications.
From Huawei’s perspective, more of its storage customers are “looking towards big data solutions to become more efficient and cheaper,” noted Zhao Chunhui, vice president of Huawei’s storage unit. The company said the integration of Alluxio’s memory-focused approach to object storage with its exabyte-level data storage pool aimed at production big data systems and on-demand data analytics helped double storage utilization on its FusionStorage platform.
The partners further claimed the big data storage accelerator boosted analytics efficiency by 50 percent while reducing total cost of ownership by a similar amount.
Along with delivering memory-based data access and storage for uses cases such as enterprise analytics, Alluxio said it also is targeting the growing adoption of hybrid cloud infrastructure. The startup has so far raised $7.5 million in one venture-funding round. Whether Huawei emerges as an investor is unclear given recent disputes between the Chinese networking giant and U.S. lawmakers over network security and eavesdropping, allegations Huawei has denied.