P1253814649GaZAp - PowerPoint PPT Presentation

Title: P1253814649GaZAp


1
source Alex Szalay
2
Example Sloan Digital Sky Survey

• The SDSS telescope array is systematically
  mapping ¼ of the entire sky
• Discoveries are made by querying the database,
  not through a zero-sum wrestling match for
  telescope time
• Managed by an RDBMS
• (MS SQL Server), equipped with a hierarchical
  triangular mesh index, among other customizations
• 15 TB in the final release in 2007
• 818 GB in the RDBMS (13.6B tuples)

3
source Alex Szalay
4
Drowning in data starving for information
Empirical X ? Analytical X ? Computational X
? X-informatics

• Acquisition eventually outpaces analysis
• Medicine Online publishing, digital charts
• Astronomy Big telescopes (more in a bit)
• Genetics PCR, Shotgun Sequencing
• Oceanography ??
• Marine Microbiology ??

Increase Data Collection Exponentially in Less
Time, with FlowCAM
5
Cyber-Observatories

• Arctic Observing Network (AON)
• Ocean Observing Initiative (OOI)
• National Ecological Observatory Network (NEON)
• The Waters Network
• The Long-Term Ecological Research (LTER) network
• The Geosciences Network (GEON)
• Earthscope/Incorporated Research Institutions for
  Seismology (IRIS)
• Virtual Solar-Terrestrial Observatory (VSTO)
• Linked Environments for Atmospheric Discovery
  (LEAD)

6
source Alex Szalay
7
source Jim Gray
8
Relational Databases (In Codd we Trust)
At IBM Almaden in 60s and 70s, Ted Codd worked
out a formal basis for tabular data
representation, organization, and access1. The
early systems were buggy and slow (and sometimes
reviled), but programmers only had to write 5 of
the code the previously did. Key Idea Programs
that manipulate tabular data exhibit an algebraic
structure proposed a relational algebra to
manipulate these data sets in their logical form,
indpendently of their physical representation
phsyical data independence
logical data independence
1 E. F. Codd, A Relational Model of Data for
Large Shared Data Banks, Communications of the
ACM 13(6), pp 377-387, 1970
9
source Raghu Ramakrishnan
10
Characteristicsof Cloud Computing

• Virtual Physical location and underlying
  infrastructure details are transparent to users
• Scalable Able to break complex workloads into
  pieces to be served across an incrementally
  expandable infrastructure
• Efficient Services Oriented Architecture for
  dynamic provisioning of shared compute resources
• Flexible Can serve a variety of workload types
  both consumer and commercial

11
Cloud Computing as Hosted Data Management Services

• Yahoo
• Yahoo Distributed Hash Tables Key/value pairs
• Yahoo Distributed Ordered Tables Ordered ranges
• PNUTS Relational-style storage, indexing and
  query
• Amazon
• S3 Simple Storage
• SimpleDB Quasi-Relational features
• Google
• APIs for Storage, Visualization, Document
  processing, Images, Mail
• Microsoft
• CloudDB Relational-style features

12
Workflow at CMOP
Washington University
Cloud
PNNL
OHSU
Cloning/ cDNA/
Sequencing
plates
FASTA files
Inspection
FASTA files
e.g., trim bad reads at the end
Cleaning
BLAST
Post processing
Hit tables
Link
Analyze
Hit tables metadata
synopsis
Shared Knowledge