Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Big Data Analytics
Search
Matt Wood
August 01, 2012
Technology
7
1.3k
Big Data Analytics
An introduction to Big Data Analytics in the cloud.
Matt Wood
August 01, 2012
Tweet
Share
More Decks by Matt Wood
See All by Matt Wood
Field Notes from Expeditions in the Cloud
mza
2
420
A Platform for Big Data
mza
6
770
The Data Lifecycle
mza
5
520
Provision Throughput Like a Boss
mza
0
460
Impact of Cloud Computing: Life Sciences
mza
2
880
Latency's Worst Nightmare: Performance Tuning Tips and Tricks
mza
4
1.1k
Under the Covers of DynamoDB
mza
4
1.1k
From Analytics to Intelligence: Amazon Redshift
mza
9
1k
Scaling Science
mza
3
530
Other Decks in Technology
See All in Technology
[re:Inent2025事前勉強会(有志で開催)] re:Inventで見つけた人生をちょっと変えるコツ
sh_fk2
1
430
serverless team topology
_kensh
3
230
Okta Identity Governanceで実現する最小権限の原則 / Implementing the Principle of Least Privilege with Okta Identity Governance
tatsumin39
0
180
QA業務を変える(!?)AIを併用した不具合分析の実践
ma2ri
0
150
プロファイルとAIエージェントによる効率的なデバッグ / Effective debugging with profiler and AI assistant
ymotongpoo
1
250
OCIjp_Oracle AI World_Recap
shinpy
1
180
SOTA競争から人間を超える画像認識へ
shinya7y
0
570
ブラウザのAPIで Nintendo Switch用の特殊なゲーム用コントローラーを体験型コンテンツに / IoTLT @ストラタシス・ジャパン
you
PRO
0
140
SQLAlchemy の select(User).where(User.id =="123") を理解してみる/sqlalchemy deep dive
3l4l5
3
520
SRE × マネジメントレイヤーが挑戦した組織・会社のオブザーバビリティ改革 ― ビジネス価値と信頼性を両立するリアルな挑戦
coconala_engineer
0
270
だいたい分かった気になる 『SREの知識地図』 / introduction-to-sre-knowledge-map-book
katsuhisa91
PRO
3
1.4k
Azure Well-Architected Framework入門
tomokusaba
1
130
Featured
See All Featured
How to train your dragon (web standard)
notwaldorf
97
6.3k
Producing Creativity
orderedlist
PRO
347
40k
Thoughts on Productivity
jonyablonski
70
4.9k
Save Time (by Creating Custom Rails Generators)
garrettdimon
PRO
32
1.7k
Statistics for Hackers
jakevdp
799
220k
Gamification - CAS2011
davidbonilla
81
5.5k
How to Ace a Technical Interview
jacobian
280
24k
Docker and Python
trallard
46
3.6k
GraphQLとの向き合い方2022年版
quramy
49
14k
Optimising Largest Contentful Paint
csswizardry
37
3.5k
The Success of Rails: Ensuring Growth for the Next 100 Years
eileencodes
46
7.7k
BBQ
matthewcrist
89
9.9k
Transcript
Big Data Analytics w i t h A m a
z o n W e b S e r v i c e s Dr. Matt Wood An Online Seminar for Partners. Wednesday 1st August.
Hello, and thank you.
Big Data Analytics An introduction
Big Data Analytics An introduction The story of analytics on
AWS
Big Data Analytics An introduction The story of analytics on
AWS Integrating partners
Big Data Analytics An introduction The story of analytics on
AWS Integrating partners Partner success stories
INTRODUCING BIG DATA 1
Data for competitive advantage.
Customer segmentation, financial modeling, system analysis, line-of-sight, business intelligence. Using
data
Generation Collection & storage Analytics & computation Collaboration & sharing
Cost of data generation is falling.
Generation Collection & storage Analytics & computation Collaboration & sharing
lower cost, increased throughput
Generation Collection & storage Analytics & computation Collaboration & sharing
HIGHLY CONSTRAINED
Very high barrier to turning data into information.
Move from a data generation challenge to analytics challenge.
Enter the Cloud.
Remove the constraints.
Enable data-driven innovation.
Move to a distributed data approach.
Maturation of two things.
Maturation of two things. Software for distributed storage and analysis
Maturation of two things. Software for distributed storage and analysis
Infrastructure for distributed storage and analysis
Frameworks for data-intensive workloads. Software Distributed by design.
Platform for data-intensive workloads. Infrastructure Distributed by design.
Support the data timeline.
Generation Collection & storage Analytics & computation Collaboration & sharing
HIGHLY CONSTRAINED
Generation Collection & storage Analytics & computation Collaboration & sharing
Lower the barrier to entry.
Accelerate time to market and increase agility.
Enable new business opportunities.
Washington Post Pinterest NASA
“AWS enables Pfizer to explore difficult or deep scientific questions
in a timely, scalable manner and helps us make better decisions more quickly” Michael Miller, Pfizer
THE STORY OF ANALYTICS 2
EC2 Utility computing. 6 years young.
Embarrassingly parallel problems. Scale out systems Queue based distribution. Small,
medium and high scale.
None
None
None
EC2 Utility computing. 6 years young. Cost optimization.
Achieving economies of scale 100% Time
Reserved capacity Achieving economies of scale 100% Time
Reserved capacity Achieving economies of scale 100% Time On-demand
Reserved capacity Achieving economies of scale 100% Time On-demand UNUSED
CAPACITY
Bid on unused EC2 capacity. Spot Instances Very large discount.
Perfect for batch runs. Balance cost and scale.
$650 per hour
Pattern for distributed computing. Map/reduce Software frameworks such as Hadoop.
Write two functions. Scale up.
Pattern for distributed computing. Map/reduce Software frameworks such as Hadoop.
Write two functions. Scale up. Complex cluster configuration and management.
Managed Hadoop clusters. Amazon Elastic MapReduce Easy to provision and
monitor. Write two functions. Scale up. Optimized for S3 access.
Input data S3 UNDER THE HOOD i i
Elastic MapReduce Code Input data S3 UNDER THE HOOD i
i
Elastic MapReduce Code Name node Input data S3 UNDER THE
HOOD i i
Elastic MapReduce Code Name node Input data S3 Elastic cluster
UNDER THE HOOD i i
Elastic MapReduce Code Name node Input data S3 Elastic cluster
HDFS UNDER THE HOOD i i
Elastic MapReduce Code Name node Input data S3 Elastic cluster
HDFS Queries + BI Via JDBC, Pig, Hive UNDER THE HOOD i i
Elastic MapReduce Code Name node Output S3 + SimpleDB Input
data S3 Elastic cluster HDFS Queries + BI Via JDBC, Pig, Hive UNDER THE HOOD i i
Output S3 + SimpleDB Input data S3 UNDER THE HOOD
i i
None
None
None
None
None
None
None
None
None
None
None
None
None
None
Performance
Performance Compute performance
Intel Xeon E5-2670 Cluster Compute 10 gig E non-blocking network
Placement groupings 60.5 Gb UNDER THE HOOD i i
Intel Xeon E5-2670 Cluster Compute 10 gig E non-blocking network
Placement groupings 60.5 Gb UNDER THE HOOD i i + GPU enabled instances
Performance Compute performance
Performance Compute performance IO performance
NoSQL Unstructured data storage.
Predictable, consistent performance DynamoDB Unlimited storage No schema for unstructured
data Single digit millisecond latencies Backed on solid state drives
...and SSDs for all. New Hi1 storage instances.
2 x 1Tb SSDs hi1.4xlarge 10 GigE network HVM: 90k
IOPS read, 9k to 75k write PV: 120k IOPS read, 10k to 85k write UNDER THE HOOD i i
Netflix “The hi1.4xlarge configuration is about half the system cost
for the same throughput.” http://techblog.netflix.com/2012/07/benchmarking-high-performance-io-with.html
EBS Elastic Block Store
Provisioned IOPS Provision required IO performance
Provisioned IOPS Provision required IO performance + EBS-optimized instances with
dedicated throughput
Generation Collection & storage Analytics & computation Collaboration & sharing
Performance + ease of use
PARTNER INTEGRATION 3
Extend platform with partners
Innovate on behalf of customers
Remove undifferentiated heavy lifting
Rolled the Amazon Hadoop optimizations into MapR MapR distribution for
EMR Choice for EMR customers Easy deployment for MapR customers
Hadoop distribution MapR distribution for EMR Integrated into EMR NFS
and ODBC drivers High availability and cluster mirroring
Enterprise data toolchain Informatica on EMR “Swiss army knife” for
data formats Data integration Available to all on EMR
AWS Marketplace Karmasphere, Marketshare, Acunu Cassandra, Metamarkets, Aspera and more.
aws.amazon.com/marketplace
PARTNER SUCCESS STORIES 4
Razorfish
3.5 billion records 71MM unique cookies 1.7MM targeted ads per
day
3.5 billion records 71MM unique cookies 1.7MM targeted ads per
day 500% improvement in return on ad spend.
Cycle Computing + Schrodinger
30k cores, $4200 an hour (compared to $10+ million)
Marketshare + Ticketmaster Optimize live event pricing
Reduced developer infrastructure management time by 3 hours a day
Thank you!
Q & A
[email protected]
@mza on Twitter