Virtualization Technology News and Information
Q&A: Haoyuan Li of Alluxio Talks About Huawei FusionStorage


Huawei, the China-based networking giant, has announced a new Big Data analytics acceleration solution using its FusionStorage product co-developed with Alluxio Inc., the Silicon Valley-based startup backed by venture legends Andreessen Horowitz. FusionStorage is Huawei's distributed software-defined storage system. The most recent release, FusionStorage 6.0, supports distributed block, file, and object storage, with classification, encryption, and deduplication.

Alluxio's software is a memory-centric, virtual distributed storage system. It sits between Big Data computational applications and a wide range underlying storage systems accessible through an API so the application crunches data in real-time without caring about where the data is stored. To find out more about the new storage solution, I spoke with the founder and CEO of Alluxio, Haoyuan Li.

VMblog:  What industries will most benefit from this solution?

Haoyuan Li:  In verticals such as oil/gas and media/entertainment, the ability of object storage to store large amounts of data within a single namespace, and to integrate with programs such as Hadoop, can help improve analysis of big data. One of the main object storage use cases is active archiving, in which rarely accessed data is stored on low-cost storage while remaining accessible by applications and users. This process is used to store data that is not often needed, but is important enough that the ability to access it if necessary is essential.    

VMblog:  What kinds of new workloads does this joint solution permit that were a challenge before?

Haoyuan Li:  The problem solved here is moving big data at memory speed across a range of different storage systems and technologies. Transferring data to object storage is very hard to do at scale. One of the main problems with implementing object storage is how do you get data from point A to point B?  The most efficient way is to use software that can move data from either NAS or SAN systems into an object storage system, rather than using an object storage gateway.  

VMblog:  Do you have any examples of scale, performance, or other metrics you can share?

Haoyuan Li:  Baidu, the Chinese search giant, can ran Spark queries 30 times faster with Alluxio. Batch queries that used to take 15 minutes to complete, finished in under 30 seconds.  A 1,000-user Alluxio cluster at Baidu provided more than 50TB of RAM space.

The global bank Barclays accelerated Spark jobs from hours to seconds by running Alluxio between the query-running compute nodes and a Teradata data repository.

VMblog:  When will this joint solution be available?

Haoyuan Li:  It's available globally now through Huawei.


Once again, thanks to Haoyuan Li, founder and CEO of Alluxio, for taking time to speak with

Published Friday, September 02, 2016 10:04 AM by David Marshall
There are no comments for this post.
To post a comment, you must be a registered user. Registration is free and easy! Sign up now!
<September 2016>