Szymon Sobczak

BigData @ getbase.com

concurrency
IS not
parallelism

PARALLELISM

executing multiple things at the same time

(in parallel)

CONCURRENCY

property of a solution,

decomposition of a problem into pieces

why to care

problems getting bigger

data getting bigger

computers getting bigger, but not faster

massively parallel execution

only in datacenters?

single rack can house > 2500 CPUs

example

(gophers, yay!)

images by Rob Pike - r@golang.org

http://concur.rspace.googlecode.com/hg/talk/concur.html

parallelism

execution

(# of gophers on the screen)

Concurrency

solution structure

(ability to run more / less gophers)

is the burning of old books

a stupid example?

the future

big problems

big data

massively parallel computers

need to think and design solutions
using the tools of concurrency

because we will need
the parallel execution soon

thank you