Top Banner
MAKING BIG DATA SMALL BARON SCHWARTZ STRATA CONFERENCE NYC 2013
46

Making big data small

Jul 04, 2015

Download

Software

andertech

Baron Schwartz' slides from his Making big data small at Strata Conf 2013
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Making big data small

MAKING BIG DATA SMALLBARON SCHWARTZ

STRATA CONFERENCE NYC 2013

Page 2: Making big data small

ME@XAPRB ON TWITTER

[email protected]

LINKEDIN.COM/IN/XAPRB

Optimization, Backups, Replication, and more

Baron Schwartz, Peter Zaitsev &

Vadim Tkachenko

High PerformanceMySQL

3rd Edition

Covers Version 5.5

Page 3: Making big data small
Page 4: Making big data small

STORAGETRANSFER / BANDWIDTHCOMPUTE (CPU, RAM, I/O)TIME

Page 5: Making big data small

DATA LIFECYCLE

COMPUTERETRIEVESTORETRANSFERGATHER

Page 6: Making big data small

THE MEGABYTE HOUR WHAT DOES 350 MILLISECONDS COST?

Page 7: Making big data small

BIG-O NOTATION

Page 8: Making big data small

GUIDING PRINCIPLES

Page 9: Making big data small

DO MORE WITH LESS

Page 10: Making big data small

ADDRESS THE AUDIENCE

Page 11: Making big data small

GENERATE KNOWLEDGE

DATA

INFORMATION

KNOWLEDGE

Page 12: Making big data small

SUCH KNOWLEDGE IS TOO WONDERFUL FOR ME;IT IS HIGH, I CANNOT ATTAIN IT.

- PSALM 139

Page 13: Making big data small

ANSWER THE QUESTION

Page 14: Making big data small

DOING MORE WITH LESS

Page 15: Making big data small

LOSSLESS COMPRESSION

Page 16: Making big data small

LOSSY REPRESENTATION

Page 17: Making big data small

DATA LIFECYCLE

COMPUTERETRIEVESTORETRANSFERGATHER

Page 18: Making big data small

THE RISK?

Page 19: Making big data small

LESS COMPUTELESS DATA

LOSSINESS FEATURES

TECHNIQUES

Page 20: Making big data small

LESS COMPUTELESS DATA

LOSSINESS FEATURES

TECHNIQUES

Page 21: Making big data small

DECREASE REDUNDANCY

LESS D

ATA

Page 22: Making big data small

DELTA COMPRESSION

LESS D

ATA

Page 23: Making big data small

NULLS AND ZEROES

LESS D

ATA

Page 24: Making big data small

SAMPLING

LESS D

ATA

Page 25: Making big data small

AGGREGATION

LESS D

ATA

Page 26: Making big data small

DISCARDING DUPLICATE SIGNALS

LESS D

ATA

Page 27: Making big data small

COMPRESSION

LESS D

ATA

Page 28: Making big data small

LESS COMPUTELESS DATA

LOSSINESS FEATURES

TECHNIQUES

Page 29: Making big data small

BLOOM FILTERS

LESS C

OM

PUT

E

Page 30: Making big data small

EXPONENTIALLY WEIGHTED MOVING AVERAGES

LESS C

OM

PUT

E

0

100

200

300

400

1 2 3 4 5

0

100

200

300

400

1 2 3 4 5

Page 31: Making big data small

APPROXIMATION

LESS C

OM

PUT

E

Page 32: Making big data small

LESS COMPUTELESS DATA

LOSSINESS FEATURES

TECHNIQUES

Page 33: Making big data small

MODELS

LO

SSINESS

Page 34: Making big data small

LO

SSINESS

Page 35: Making big data small

LO

SSINESS

Page 36: Making big data small

LESS COMPUTELESS DATA

LOSSINESS FEATURES

TECHNIQUES

Page 37: Making big data small

HISTOGRAMS

FEAT

UR

ES

Page 38: Making big data small

HISTOGRAMS

FEAT

UR

ES

Page 39: Making big data small

PERCENTILES

FEAT

UR

ES

Page 40: Making big data small

SPARSE OUTLIER STORAGE

FEAT

UR

ES

Page 41: Making big data small

LOST HAYSTACKS

Page 43: Making big data small

SMALL DATA METHOD

APPLYDISCARDISOLATEEXPLOREGATHER

SAMPLES

VISUALIZE

NEEDLES!

HAYSTACK

REVIEW

Page 44: Making big data small

SMART-FRIENDS METHOD

Page 45: Making big data small

CONTACT / QUESTIONS@XAPRB ON TWITTER

[email protected]

LINKEDIN.COM/IN/XAPRBPLEASE

RATE THIS TALK :-)

Page 46: Making big data small

IMAGE CREDITSHTTP://WWW.FLICKR.COM/PHOTOS/PIPER/458847242/

HTTP://WWW.FLICKR.COM/PHOTOS/MIONG/4222013704/

HTTP://WWW.FLICKR.COM/PHOTOS/BENSTEVINSON/4908627827/

HTTP://EN.WIKIPEDIA.ORG/WIKI/FILE:BLOOM_FILTER.SVG

HTTP://WWW.FLICKR.COM/PHOTOS/JINGLESLENOBEL/4915172257/

HTTP://WWW.FLICKR.COM/PHOTOS/MAUROPPP/429231880/

HTTP://WWW.FLICKR.COM/PHOTOS/ECSTATICIST/3979554514/

HTTP://WWW.FLICKR.COM/PHOTOS/TAMBAKO/9158311021/

HTTP://WWW.FLICKR.COM/PHOTOS/FUZZYLITTLEMANPEACH/4689679368/

HTTP://WWW.FLICKR.COM/PHOTOS/FWP/49514206/

HTTP://WWW.FLICKR.COM/PHOTOS/QTHOMASBOWER/3511960429/

HTTP://WWW.FLICKR.COM/PHOTOS/JIMNISTA/7239478006/

HTTP://WWW.FLICKR.COM/PHOTOS/FLOWERS-OF-THE-SEA/6971586987/

HTTP://WWW.FLICKR.COM/PHOTOS/OLDPATTERNS/5817264314/

HTTP://WWW.FLICKR.COM/PHOTOS/BREWBOOKS/184343329/

HTTP://WWW.FLICKR.COM/PHOTOS/PAREEERICA/4235666733/

HTTP://WWW.FLICKR.COM/PHOTOS/JDHANCOCK/5177358991/

HTTP://WWW.FLICKR.COM/PHOTOS/ROSEMARY/47396922/

HTTP://WWW.FLICKR.COM/PHOTOS/ECSTATICIST/1337749333/

HTTP://WWW.FLICKR.COM/PHOTOS/PINKSHERBET/233228813/

HTTP://BOST.OCKS.ORG/MIKE/SIMPLIFY/

HTTP://WWW.FLICKR.COM/PHOTOS/PSOUP216/7540279198/

HTTPS://GITHUB.COM/BRENDANGREGG/HEATMAP/WIKI

HTTP://WWW.FLICKR.COM/PHOTOS/WALKINGSF/6221430560/

HTTP://WWW.FLICKR.COM/PHOTOS/WALKINGSF/7835295762/

HTTP://WWW.FLICKR.COM/PHOTOS/NASAMARSHALL/8615696355/

HTTP://WWW.FLICKR.COM/PHOTOS/BOOLEANSPLIT/8482641188/

HTTP://WWW.FLICKR.COM/PHOTOS/HTAKASHI/6861028238/

HTTP://WWW.FLICKR.COM/PHOTOS/ZDETO/4383899069/

JPL VIA HTTP://WWW.PLANETARY.ORG/EXPLORE/PROJECTS/PIONEER-ANOMALY-1/PIONEER_ANOMALY_STAIRS_DATA_TAPES_3.HTML

HTTP://EN.WIKIPEDIA.ORG/WIKI/FILE:PIONEER_10-11_SPACECRAFT.JPG