The third Apache HBaseConAsia summit takes place in Beijing on July 20. As the Apache HBase community’s premier user conference in Asia, HBaseCon has run since 2012. More than 20 experts and community leaders from leading internet and big-data companies will share the latest on HBase and the surrounding ecosystem.
Apache HBase is a highly available, high-performance, multi-version distributed NoSQL database on Apache Hadoop—Google Bigtable’s open-source counterpart. It delivers fast random read/write at scale on commodity servers.
With mobile and IoT data growth, HBase has spread widely. Alibaba, Facebook, Yahoo, Xiaomi, Huawei, Tencent, JD, DiDi, NetEase, 360, Kuaishou, and others are heavy users and active contributors. China is a major driver of the HBase ecosystem: five PMC members and 17 Committers domestically; Xiaomi alone has two PMC members and nine Committers.
Sessions Preview
Opening Keynote Speaker: Baoqiu Cui (Vice President and Chair of the Technology Committee, Xiaomi Group)
HBase Today and Tomorrow
Session: State of HBase Abstract: HBase 2.0.0 shipped recently; 3.0.0 is on the horizon. PMC member Duo Zhang discusses HBase 2.x and 3.x—Procedure V2, Assignment V2, HBCK2, cross-datacenter replication, async client, and more. Speaker: Duo Zhang (HBase PMC, Head of Xiaomi Storage, Secretary of Xiaomi Open Source Committee)
Session: HBase on the Cloud—Advantages and Trends Abstract: How does cloud HBase compare to on-prem datacenters? What are the challenges and future trends?
- Why HBase fits the cloud naturally
- HDFS on cloud disks
- Using diverse cloud storage tiers
- Serverless HBase—value and implementation
- Elastic VMs—what else can cloud HBase do?
- GPUs, FPGAs, and new hardware Speaker: Chunhui Shen (HBase PMC, Senior Expert, Alibaba)
Session: HBase BucketCache with Persistent Memory Abstract: Intel DCPMM (Data Centre Persistent Memory) offers larger capacity with persistence. Intel engineers share building BucketCache on DCPMM with performance numbers. Speakers: Anoop Jam John (HBase PMC), Ramkrishna S Vasudevan (HBase PMC), Xu Kai (Intel Engineer)
HBase 2.x Kernel Improvements
Session: Further GC Optimization: Reading HFileBlock into Off-Heap Directly Abstract: HBase 2.0.0 off-heaped core paths and cut GC impact on latency. With low cache hit rate, read p999 still matched GC STW. Intel and Xiaomi engineers on pushing GC optimization further. Speakers: Anoop Jam John (HBase PMC), Zheng Hu (Xiaomi HBase Engineer, HBase Committer)
Session: HBCK2: Concepts, Trends and Recipes for Fixing Issues within HBase 2 Abstract: HBCK2 concepts, details, and best practices for developers and operators. Speaker: Wellington Chevreuil (Cloudera HBase Engineer, HBase Committer)
Session: WAL Splitting and ACL via Procedure V2 Abstract: Procedure V2 coordinates distributed workflows. Xiaomi’s design for split WAL and ACL on top of it. Speaker: Yi Mei (Xiaomi HBase Engineer, HBase Committer)
HBase in Production
Session: Latest HBase Progress at Pinterest Abstract: Transactions, streaming, secondary indexes, SQL:
- ACID via Apache Omid
- In-house Sparrow vs Omid—2× throughput, 20% lower latency
- Argus + Kafka for WAL notifications
- Lily-based Ixia for real-time secondary indexes
- Ixia + Muse for SQL-like queries Speaker: Lianghong Xu (Pinterest Engineer)
Session: HBase at Tencent Abstract: Social, payments, WeChat, and more. Committer Guangxu Cheng on use cases and improvements. Speaker: Guangxu Cheng (Tencent HBase Engineer, HBase Committer)
Session: HBase for Trillion-Scale Short Video and Billion-Scale User Profiles at Kuaishou Abstract: How Kuaishou stores massive short-video and user-profile data on HBase and HDFS. Speaker: Ming Xu (Kuaishou Big Data Architecture Engineer)
More sessions: HBaseConAsia2019
Registration
Scan the QR code for free registration and a commemorative HBaseCon Asia 2019 T-shirt.
T-shirt sample
Live Stream
Three tracks: Track 1: HBase kernel design and improvements—Xiaomi Live ID: 3004163 Track 2: Ecosystem products and solutions—ID: 2329658282 Track 3: Production use cases—ID: 2312959070
How to watch on Xiaomi Live:
- Install the Xiaomi Live app;
- Search for the track ID (e.g. 3004163);
- Enter the room for that track.