PETASCALE DATA STORAGE INSTITUTE

1
PETASCALE DATA STORAGE INSTITUTE

• The Drive to Petascale Computing
• Faster computers need more data, faster.

• Checkpoint at Terabytes/sec
• Petabyte files
• Billions of files
• Revisit programming for Input/Output
• Data center automation
• Acceleration for search

Everything Must Scale with Compute
2015 100 PF --
Computing Speed
TFLOP/s
Memory
2,500
TeraBytes
Year
2012
250
08
25
04
Disk
-- 2011 10 PF
50
2.5
5
PetaBytes
.5
00
.05
1
5
Application Performance
50
200
Parallel I/O
-- 2008 1 PF
500
.5
.5
5,000
200
GigaBytes/sec
5
2,000

• PDSI Thrusts
• Data Capture
• Education Dissemination
• Innovation

5
20,000
50
50
-- 2005 100 TF
Metadata
500
Inserts/sec
Network Speed
500
ArchivalStorage
Gigabits/sec
GigaBytes/sec
-- 2001 10 TF
2
Steeped in Terascale Experience
PETASCALE DATA STORAGE INSTITUTE
Seaborg GPFS
3
PETASCALE DATA STORAGE INSTITUTE
Strategic Plan

• IT Automation
• Instrumentation
• Visualization
• Machine Learning
• Diagnosis
• Adaptation

• Scaling Further
• Global/WAN access
• Federated security
• Metadata at scale
• Para-virtualization

Peta-Bytes Tera-B/sec Giga-files Mega-CPUs

• Outreach
• Storage-research-list
• Collaboration w/ other Scidacs

Innovation

• Education
• Workshops
• Tutorials
• Course materials

Education Dissemination

• HPC NFS
• Parallel NFS
• Secure NFS
• IETF Standard

• App Workloads
• INCITE resources
• Trace replay tools (e.g. BLAST, CCSM,
  Calore, EVH1, MCNP, GYRO, Sierra, QCDand other
  Scidacs)

Data Capture

• API Standards
• POSIX API
• Rich metadata
• Compute-in-disk
• Archive API
• Quality of Storage

• Failure Data
• Capture publish
• Computer Failure Data Repository
• (e.g. LANLs outages by root cause)

NFSv4extendedw/ layouts
Tera-Bytes Giga-B/sec Mega-files Kilo-CPUs
4
Participating Organizations
PETASCALE DATA STORAGE INSTITUTE

• Carnegie Mellon UniversityGarth Gibson (PI)
• University of California, Santa Cruz Darrell
  Long (co-PI)
• University of Michigan, Ann ArborPeter Honeyman
  (co-PI)
• Los Alamos National Laboratory Gary Grider
  (co-PI)
• Lawrence Berkeley National LaboratoryBill Kramer
  (co-PI)
• Oak Ridge National LaboratoryPhilip Roth (co-PI)
• Pacific Northwest National LaboratoryEvan Felix
  (co-PI)
• Sandia National LaboratoryLee Ward (co-PI)

5
Programming for Storage
PETASCALE DATA STORAGE INSTITUTE

• The Need for Training Programmers for Storage
• HPC IT managers work for users who program apps
• Often performance of apps/workflows dependent on
  storage
• Many times best solutions would be to change the
  program
• Reality is app specialists intolerant of requests
  to reprogram for better storage performance
• That is, reprogramming for storage performance
  often doesnt get done
• Approach Create tools, training to help a priori
• Give programmers libraries, performance debugging
  tools that avoid or detect poor storage patterns
• Give tutorials, case studies, help pages showing
  weak programming approaches and how to improve
  them

6
Example from BioInformatics
PETASCALE DATA STORAGE INSTITUTE

• Pseudo code example from IT manager -- single
  thread
• for( I0, Ilt1000, I)
• for( J0, Jlt1000, J)
• buf compute (I,J)
• f open( file_foo)
• lseek(f, offset(I,J))
• write(f, buf, lengthof(buff))
• close(f)
• Buf turns out to be small, unaligned, fixed
  length
• Obvious fixes
• Open/close outside both loops
• Malloc sizeof 1000000lengthof(buff), copy into
  it in memory, one write at end