Data-driven Innovation

Slides from my session at the #AWS Public Sector Summit, 2012.

Transcript

1. Data-driven innovation matthew@amazon.com Dr. Matt Wood @mza
2. Hello
3. Hello
4. Data
5. DNA
6. Chromosome 11 : ACTN3 : rs1815739
7. Chromosome X : rs6625163
8. Chromosome 19 : FUT2 : rs601338
9. Chromosome 2 : rs10427255
10. TYPE II Chromosome 10 : rs7903146
11. +0.25 Chromosome 15 : rs2472297
12. I know this, because...
14. A T C G G T C C A G G
15. A T C G G T C C A G G A G C C A G G U C C Transcription
16. A T C G G T C C A G G A G C C A G G U C C Translation Ser Glu Val Transcription
19. Chromosome 11 : ACTN3 : rs1815739
20. Chromosome X : rs6625163
21. Chromosome 19 : FUT2 : rs601338
22. Chromosome 2 : rs10427255
23. TYPE II Chromosome 10 : rs7903146
24. +0.25 Chromosome 15 : rs2472297
25. I know all that, because...
26. Human Genome Project
27. 40 species ensembl.org
28. Compare
29. Change
30. Less
33. Compare
34. Transformative
36. Data generation costs are falling everywhere
37. Customer segmentation, financial modeling, system analysis, line of sight, business intelligence.
38. Opportunity
39. Transformation
40. Innovation
41. Generation Collection & storage Analytics & computation Collaboration & sharing
42. Generation Collection & storage Analytics & computation Collaboration & sharing lower cost, increased throughput
43. Generation Collection & storage Analytics & computation Collaboration & sharing lower cost, increased throughput highly constrained
44. Barrier
45. Data generation challenge X
46. Analytics challenge
47. Accessibility challenge
48. Enter the AWS Cloud
49. Utility
50. Remove constraints
51. Data-driven innovation
52. Distributed
54. 2 Software for distributed storage & analysis
55. 2 Software for distributed storage & analysis Infrastructure for distributed storage & analysis
56. Software Frameworks for data-intensive work loads. Distributed by design.
57. Infrastructure Platform for data-intensive work loads. Distributed by design.
58. Support the data timeline
59. Generation Collection & storage Analytics & computation Collaboration & sharing highly constrained
60. Generation Collection & storage Analytics & computation Collaboration & sharing
61. Lower the barrier to entry
62. Agility
63. Responsive
64. Generation Collection & storage Analytics & computation Collaboration & sharing
65. Generation DynamoDB Analytics & computation Collaboration & sharing
66. Generation DynamoDB EC2, Elastic MapReduce Collaboration & sharing
67. Generation DynamoDB EC2, Elastic MapReduce S3, Public Datasets
68. Tools and techniques for working productively with data
69. Scale
70. Secure
71. 2 Software for distributed storage & analysis Infrastructure for distributed storage & analysis
72. Amazon EC2
73. Scale out systems Embarrassingly parallel Queue based distribution Small, medium and high scale
74. High performance
75. High performance Compute performance
76. Cluster Compute Intel Xeon E5-2670 10 gigabit, non-blocking network 60.5 Gb Placement groupings
77. Cluster Compute Intel Xeon E5-2670 10 gigabit, non-blocking network 60.5 Gb Placement groupings +GPU
78. 240 TFLOPS
79. High performance Compute performance IO performance
80. Unstructured
81. Variable
82. Amazon DynamoDB Predictable, consistent performance Unlimited storage Single digit millisecond latencies No schema. Zero admin.
83. ...and SSDs for all
84. hi1.4xlarge 2 x 1Tb SSD storage 10 gigabit networking HVM: 90k IOPS read, 9k to 75k write PV: 120k IOPS read, 10k to 85k write
85. 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
86. Provisioned IOPS Provision required IO performance EBS optimized instances
87. Cost optimization
88. Reserved capacity
89. Reserved capacity On-demand
90. Reserved capacity On-demand
91. Spot instances
93. $0.2530 vs $2.40
94. 2 Software for distributed storage & analysis Infrastructure for distributed storage & analysis
95. map/reduce
96. Map. Reduce.
97. Write functions. Scale up.
98. Hadoop
99. Undi erentiated heavy lifting
100. Amazon Elastic MapReduce Managed Hadoop Clusters Easy to provision and monitor Write two functions. Scale up. Choice of Hadoop flavors
101. Amazon Elastic MapReduce Integrates with S3 Analytics for DynamoDB Perfect for Spot pricing
102. Input data S3
103. Elastic MapReduce Code Input data S3
104. Elastic MapReduce Code Name node Input data S3
105. Elastic MapReduce Code Name node Input data S3 Elastic cluster
106. Elastic MapReduce Code Name node Input data S3 Elastic cluster HDFS
107. Elastic MapReduce Code Name node Input data S3 Elastic cluster HDFS Queries + BI Via JDBC, Pig, Hive
108. Elastic MapReduce Code Name node Output S3 + SimpleDB Input data S3 Elastic cluster HDFS Queries + BI Via JDBC, Pig, Hive
109. Output S3 + SimpleDB Input data S3
110. CDC Centers for Disease Control and Prevention
111. “BioSense 2.0 protects the health of the American people by providing timely insight into the health of communities, regions, and the nation by o ering a variety of features to improve data collection, standardization, storage, analysis, and collaboration”
112. Health data Collection & storage Analytics & computation Collaboration & sharing
113. Health data Collection & storage Analytics & computation Collaboration & sharing highly constrained
114. HIPAA, HITECH, FISMA Moderate
115. GovCloud
116. Beyond a definition of Big Data
117. Chromosome 11 : ACTN3 : rs1815739
118. Chromosome X : rs6625163
119. Chromosome 19 : FUT2 : rs601338
120. Chromosome 2 : rs10427255
121. TYPE II Chromosome 10 : rs7903146
122. +0.25 Chromosome 15 : rs2472297
123. Thank you aws.amazon.com @mza matthew@amazon.com