This Meetup is past

125 people went

Location visible to members

Details

Alluxio (formerly Tachyon) features a chance to interact with other Alluxio (http://www.alluxio.com/) users and developers, along with two presentations. Registration is required.

Bin Fan from Alluxio and Chen Tian from Huawei will co-present the need and architecture for a memory-speed distributed storage solution, using Alluxio and Huawei’s Fusion Storage as an example.

Lin Ma from Huawei will be sharing their experience leveraging Alluxio to gain faster insights from mobile data.

Agenda:

6:30 - 7:00: Food & Networking

7:00 - 8:00: Presentations

8:00 - 8:15: Q&A

8:15 - 8:30: Wind down

Food will be available starting at 6:30pm, presentations will begin at 7:00pm.

Special thanks to Huawei for sponsoring and hosting this meetup!

Using Alluxio for a Memory-Speed, Scalable, Object Storage Solution for Enterprise Big Data Analytics

Abstract:

Enterprises typically store large amounts of data in existing storage systems, which are often separate from big data analytics systems. Therefore, importing petabytes of data into a big data analytics system takes a long time with large overheads and high costs. Even worse, transferring large amounts of data results in data silos and unnecessary duplication, which creates serious data management problems.

Alluxio solves this problem by transparently connecting applications to existing storage, and by accelerating access by keeping the data in memory. We present how an architecture looks like with Alluxio and existing storage solutions. Also, Alluxio and Huawei work together to provide a memory-speed distributed storage solution with dynamic data tiering, achieving transparent classification for hot and cold data and bring more efficient data analysis while reducing more than 30% of the total storage cost.

Bio:

Bin Fan is a software engineer at Alluxio Inc. He is one of the top contributors and the PMC maintainer of the open source Alluxio project. Prior to Alluxio, he worked for Google to build the next-generation storage infrastructure and won Google's Technical Infrastructure award. Bin got his Ph.D. in Computer Science from Carnegie Mellon University in 2013.

Chen Tian is a director in Software Lab at Huawei U.S. R&D Center. He is leading a team responsible for delivering cutting-edge solutions that improve software performance in various domains ranging from android phones, 4G/5G wireless network, SDN, NFV, storage server, to cloud computing and data center. Mr. Tian is also a researcher with a proven track record on publishing papers on top-tier computer science conferences. His research interests span the areas of distributed and parallel systems, operating systems, compilers, programming languages, software engineering, and computer architecture.

Crash-Proofing Smartphones with Alluxio

Abstract:

Huawei is one of the worldwide market leaders for mobile smartphones. In order to improve overall quality assurance for handsets, Huawei uses its own internal, distributed program analysis framework in order to process all the data from mobile devices. These processes are critical to identifying root causes for crashes, failures, and performance issues. Improved analysis on the mobile data can prevent future crashes and increase performance for mobile devices.

High throughput and low latency program analysis on the data is important for faster discovery of issues, more complete analysis, and better efficiency of resources. Alluxio improves the throughput and latency by storing data in memory, thus resulting in faster insights from the data. We describe how we detect and prevent mobile phone issues, discuss how we use Alluxio in our architecture, and present results from initial experiments.

Bio:

Lin Ma is Senior Staff Research Scientist in Huawei America Research Center, affiliated in Parallel and Distributed Computing Lab. He received his PhD from Washington University in St. Louis. His research interest includes parallel architecture and algorithms, performance evaluation and tuning model for multi-threading system, accelerating application-specific architectures, and high-performance distributed computing over architecturally diverse system of CPUs and GPUs.