Learning to Build Distributed Systems the Hard Way

Slide 1

Slide 1 text

LEARNING TO BUILD DISTRIBUTED SYSTEMS THE HARD WAY @iconara NEW and improved!

Slide 2

Slide 2 text

LEARNING TO BUILD DISTRIBUTED SYSTEMS THE HARD WAY @iconara NEW and improved!

Slide 3

Slide 3 text

speakerdeck.com/iconara

Slide 4

Slide 4 text

Theo / @iconara

Slide 5

Slide 5 text

chief architect at BURT

Slide 6

Slide 6 text

FAILURE embrace it

Slide 7

Slide 7 text

SCALE how hard can it be? let’s worry about that later.

Slide 8

Slide 8 text

KNOW YOUR LIMITS who’s the largest customer you could sign? what would happen if you did?

Slide 9

Slide 9 text

BALANCE “customer” is a really bad shard key, ﬁnd something that distributes evenly & uniformly

Slide 10

Slide 10 text

SCALE OUT, NOT UP bigger boxes aren’t going to save you <

Slide 11

Slide 11 text

START WITH TWO OF EVERYTHING going from one to two is the hardest, force yourself to solve the scaling problem up front

Slide 12

Slide 12 text

START WITH TWO OF EVERYTHING you’ll solve the scaling problem, and need less overcapacity THREE

Slide 13

Slide 13 text

LIMITS we’ll probably never run out of memory

Slide 14

Slide 14 text

BACK PRESSURE what happens when the system is working at full capacity? what happens next?

Slide 15

Slide 15 text

PRODUCTION = QA production is where the weird shit happens, can you test production trafﬁc without deploying to production? =

Slide 16

Slide 16 text

MONOLITHS running all the things on the same box is really fast. what could ever go wrong? 1:4:9

Slide 17

Slide 17 text

DECOUPLE UNTIL IT BREAKS moving things to separate services means that you will be able to scale them independently

Slide 18

Slide 18 text

PROCESSING & STORAGE separate processing from storage, they almost never scale together.

Slide 19

Slide 19 text

SCALE exponential scaling is also scaling, but you want it as cheaply as possible