ALLUXIO (FORMERLY TACHYON): UNIFY DATA AT MEMORY SPEED

- EFFECTIVE USING SPARK WITH ALLUXIO

Spark Summit at BostonFeb. 2017

Haoyuan Li, Alluxio Inc.

HISTORY

• Started at UC Berkeley AMPLab In Summer 2012 • Originally named as Tachyon • Rebranded to Alluxio in early 2016

• Open Sourced in 2013 • Apache License 2.0 • Latest Stable Release: Alluxio 1.4.0 • Alluxio 1.5.0 Planned For Q2, 2017

2

3

• Fastest growing open-source project in the big data ecosystem

• 400+ contributors from 100+ organizations

• Running in large production clusters

• Community members are welcome!

FASTEST GROWING BIG DATA PROJECTS

Popular Open Source Projects’ Growth

Months

Num

ber o

f Con

trib

utor

s

INDUSTRY ADOPTION

4

…

…

5

BIG DATA ECOSYSTEM YESTERDAY

BIG DATA ECOSYSTEM TODAY

…

…

5

…

…

BIG DATA ECOSYSTEM ISSUES

5

BIG DATA ECOSYSTEM WITH ALLUXIO

…

…

FUSE Compatible File SystemHadoop Compatible File System Native Key-Value InterfaceNative File System

GlusterFS InterfaceAmazon S3 Interface Swift InterfaceHDFS Interface

5

BIG DATA ECOSYSTEM WITH ALLUXIO

…

…

FUSE Compatible File SystemHadoop Compatible File System Native Key-Value InterfaceNative File System

Unifying Data at Memory Speed

GlusterFS InterfaceAmazon S3 Interface Swift InterfaceHDFS Interface

5

6

7

WHY ALLUXIO

8

Co-located compute and data with memory-speed access to data

Virtualized across different storage systems under a unified namespace

Scale-out architecture

File system API, software only

9

Unification

New workflows across any data in any storage system

Orders of magnitude improvement in run time

Choice in compute and storage – grow each independently, buy only what is needed

Performance Flexibility

BENEFITS

#1 – ACCELERATING REMOTE STORAGE I/O

10

• Scenario: Compute and Storage Separation • Meet different compute and storage hardware requirements

• Scale compute and storage independently

• Store data in traditional filers/SANs and object stores

• Analyze existing data with Big Data compute frameworks

• Limitation

• Accessing data requires remote I/O

I/O WITHOUT ALLUXIO

Spark

Storage

11

I/O WITHOUT ALLUXIO

Spark

Storage

Low latency, memory throughput

High latency, network throughput

11

I/O WITH ALLUXIO

Spark

Storage

Alluxio

12

I/O WITH ALLUXIO

Spark

Storage

AlluxioKeeping data in Alluxio accelerates data access

12

CASE STUDY: BAIDU

13

The performance was amazing. With Spark SQL alone, it took 100-150 seconds to finish a query; using Alluxio, where data may hit local or remote Alluxio nodes, it took 10-15 seconds.

- Baidu

RESULTS

• Data queries are now 30x faster with Alluxio

• Alluxio cluster runs stably, providing over 50TB of RAM space

• By using Alluxio, batch queries usually lasting over 15 minutes were transformed into an interactive query taking less than 30 seconds

Baidu’s PMs and analysts run

interactive queries to gain insights

into their products and business

• 200+ nodes deployment

• 2+ petabytes of storage

• Mix of memory + HDD

ALLUXIO

Baidu File System

#2 – SHARING DATA AT MEMORY-SPEED AMONG APPLICATIONS

• Scenario: Data Sharing Architecture

• Pipelines: output of one job is input of the next job

• Applications, jobs, and contexts reading the same data

• Limitation

• Sharing data requires I/O

14

SHARING WITHOUT ALLUXIO

Spark

Storage

MapReduce Spark

15

SHARING WITHOUT ALLUXIO

Spark

Storage

MapReduce Spark

Network I/O

Disk I/O

I/O slows down

sharing

15

SHARING WITH ALLUXIO

Spark

Storage

MapReduce Spark

Alluxio

16

SHARING WITH ALLUXIO

Spark

Storage

MapReduce SparkMemory-speed

sharingAlluxio

16

CASE STUDY: BARCLAYS

Thanks to Alluxio, we now have the raw data immediately available at every iteration and we can skip the costs of loading in terms of time waiting, network traffic, and RDBMS activity.

- Barclays

RESULTS

• Barclays workflow iteration time decreased from hours to seconds

• Alluxio enabled workflows that were impossible before

• By keeping data only in memory, the I/O cost of loading and storing in Alluxio is now on the order of seconds

Barclays uses query and machine

learning to train models for risk

management

• 6 node deployment

• 1TB of storage

• Memory only

ALLUXIO

Relational Database

17

#3 – UNIFYING DATA ACCESS FROM DIFFERENT STORAGE

• Scenario: Multiple Storage Systems

• Most enterprises have multiple storage systems

• New (better, faster, cheaper) storage systems arise

• Limitation

• Accessing data from different systems requires different APIs

18

ACCESSING DATA THROUGH ALLUXIO

Storage B

Alluxio

Spark MapReduce Spark

19

ACCESSING DATA THROUGH ALLUXIO

Storage B

Alluxio

Spark MapReduce Spark

19

ACCESSING DATA THROUGH ALLUXIO

Storage B

Alluxio

Spark MapReduce Spark

Storage A Storage C

Flexible,

simple

no application changes,

new mount point

19

CASE STUDY: QUNAR

We’ve been running Alluxio in production for over 9 months, Alluxio’s unified namespace enable different applications and frameworks to easily interact with data from different storage systems.

- Qunar

RESULTS

• Data sharing among Spark Streaming, Spark batch and Flink jobs provide efficient data sharing

• Improved the performance of their system with 15x – 300x speedups

• Tiered storage feature manages storage resources including memory and HDD

• 200+ nodes deployment

• 6 billion logs (4.5 TB) daily

• Mix of Memory + HDD

ALLUXIO

Qunar uses real-time machine

learning for their website ads.

20

SUMMARY

21

• Adopted by industry leaders

• Unified, memory-speed data access across compute frameworks and storage systems

• Rapidly growing OS community

JOIN THE COMMUNITY

22

Contribute @ www.alluxio.org/contribute Get started @ goo.gl/55ApFx

Contact: haoyuan@alluxio.com

Twitter: @haoyuan

Websites: www.alluxio.com and www.alluxio.org

Thank you! We are hiring!Demo: Spark + Alluxio + S3Alluxio Unified Namespace

23