SAWZALL

1
SAWZALL

• Rob Pike, Sean Dorward, Robert Griesemer, Sean
  Quinlan
• Google, Inc

2
GUIDLINES

• BACKGROUD
• STRUCTURE
• EXAMPLES
• CHARACTERISTCS
• PERFORMANCE
• MY OPINIONS

3
BACKGROUD

• Large data set
• Flat but regular structure
• Span multiple disks and machines
• Bottleneck lies in I/O, not CPUs
• Task division and distribution
• Keep computation near to data
• Tolerance of kinds of failures

Paralyze to improve throughput
4
STRUCTURE
5
Structure
High level language
Software libraries
Scheduling software
Application files ystem
6
EXAMPLES
count table sum of int total table sum of
float sum_of_squares table sum of
float x floatinput emit countlt1 emit t
otalltx emit sum_of_squareslt- xx
proto document.proto max_pagerank_uri table
maximun(1)domainstring of urlstring weight
pagerankint doc Document input url string
doc.url emit max_pagerank_urldomain(url)lt-
url weight doc.pagerank
7
CHARACTERISTCS

• Similar to C and Pascal
• Type-safe scripting language
• Code is much shorter than C
• Pure value semantics, no reference types
• Statically typed
• No exception processing

8
PERFORMANCE
9
MY OPINION

• Thousands of machines are powerful in parallel
• Cluster and large-scale distributed system
• Demo
• 6PLANETLAB