Tweet
Tweet

Slide 1

Slide 1 text

RIAK

Slide 2

Slide 2 text

A BRIEF DIVERSION...

Slide 3

Slide 3 text

A BRIEF DIVERSION...

Slide 4

Slide 4 text

DISTRIBUTED SYSTEMS

Slide 5

Slide 5 text

CAP THEOREM

Slide 6

Slide 6 text

CONSISTENCY AVAILABILITY PARTITION TOLERANCE

Slide 7

Slide 7 text

No content

Slide 8

Slide 8 text

No content

Slide 9

Slide 9 text

CONSISTENCY

Slide 10

Slide 10 text

A B chunky: ‘bacon’ chunky: ‘bacon’

Slide 11

Slide 11 text

A B chunky: ‘bacon’ chunky: ‘bacon’ chunky: ‘chicken’

Slide 12

Slide 12 text

A B chunky: ‘chicken’ chunky: ‘bacon’ chunky: ‘chicken’

Slide 13

Slide 13 text

A B chunky: ‘chicken’ chunky: ‘bacon’ chunky: ‘chicken’

Slide 14

Slide 14 text

A B chunky: ‘chicken’ chunky: ‘chicken’ chunky: ‘chicken’

Slide 15

Slide 15 text

A B chunky: ‘chicken’ chunky: ‘chicken’ chunky: ‘chicken’

Slide 16

Slide 16 text

A B chunky: ‘chicken’ chunky: ‘bacon’ chunky: ‘chicken’

Slide 17

Slide 17 text

A B chunky: ‘chicken’ chunky: ‘bacon’ chunky: ‘chicken’ ?

Slide 18

Slide 18 text

A B chunky: ‘chicken’ chunky: ‘bacon’ chunky: ‘chicken’ chunky: ‘bacon’

Slide 19

Slide 19 text

A B chunky: ‘chicken’ chunky: ‘bacon’ chunky: ‘chicken’ chunky: ‘chicken’

Slide 20

Slide 20 text

“ATOMIC, OR LINEARIZABLE, CONSISTENCY IS THE CONDITION EXPECTED BY MOST WEB SERVICES TODAY. UNDER THIS CONSISTENCY GUARANTEE, THERE MUST EXIST A TOTAL ORDER ON ALL OPERATIONS SUCH THAT EACH OPERATION LOOKS AS IF IT WERE COMPLETED AT A SINGLE INSTANT. THIS IS EQUIVALENT TO REQUIRING REQUESTS OF THE DISTRIBUTED SHARED MEMORY TO ACT AS IF THEY WERE EXECUTING ON A SINGLE NODE, RESPONDING TO OPERATIONS ONE AT A TIME.”

Slide 21

Slide 21 text

“ATOMIC, OR LINEARIZABLE, CONSISTENCY IS THE CONDITION EXPECTED BY MOST WEB SERVICES TODAY. UNDER THIS CONSISTENCY GUARANTEE, THERE MUST EXIST A TOTAL ORDER ON ALL OPERATIONS SUCH THAT EACH OPERATION LOOKS AS IF IT WERE COMPLETED AT A SINGLE INSTANT. THIS IS EQUIVALENT TO REQUIRING REQUESTS OF THE DISTRIBUTED SHARED MEMORY TO ACT AS IF THEY WERE EXECUTING ON A SINGLE NODE, RESPONDING TO OPERATIONS ONE AT A TIME.”

Slide 22

Slide 22 text

AVAILABILITY

Slide 23

Slide 23 text

A B C D E

Slide 24

Slide 24 text

A B C D E

Slide 25

Slide 25 text

“FOR A DISTRIBUTED SYSTEM TO BE CONTINUOUSLY AVAILABLE, EVERY REQUEST RECEIVED BY A NON-FAILING NODE IN THE SYSTEM MUST RESULT IN A RESPONSE. THAT IS, ANY ALGORITHM USED BY THE SERVICE MUST EVENTUALLY TERMINATE … [WHEN] QUALIFIED BY THE NEED FOR PARTITION TOLERANCE, THIS CAN BE SEEN AS A STRONG DEFINITION OF AVAILABILITY: EVEN WHEN SEVERE NETWORK FAILURES OCCUR, EVERY REQUEST MUST TERMINATE.”

Slide 26

Slide 26 text

“FOR A DISTRIBUTED SYSTEM TO BE CONTINUOUSLY AVAILABLE, EVERY REQUEST RECEIVED BY A NON-FAILING NODE IN THE SYSTEM MUST RESULT IN A RESPONSE. THAT IS, ANY ALGORITHM USED BY THE SERVICE MUST EVENTUALLY TERMINATE … [WHEN] QUALIFIED BY THE NEED FOR PARTITION TOLERANCE, THIS CAN BE SEEN AS A STRONG DEFINITION OF AVAILABILITY: EVEN WHEN SEVERE NETWORK FAILURES OCCUR, EVERY REQUEST MUST TERMINATE.”

Slide 27

Slide 27 text

PARTITION TOLERANCE

Slide 28

Slide 28 text

A B C D E

Slide 29

Slide 29 text

A B C D E

Slide 30

Slide 30 text

“IN ORDER TO MODEL PARTITION TOLERANCE, THE NETWORK WILL BE ALLOWED TO LOSE ARBITRARILY MANY MESSAGES SENT FROM ONE NODE TO ANOTHER. WHEN A NETWORK IS PARTITIONED, ALL MESSAGES SENT FROM NODES IN ONE COMPONENT OF THE PARTITION TO NODES IN ANOTHER COMPONENT ARE LOST. (AND ANY PATTERN OF MESSAGE LOSS CAN BE MODELED AS A TEMPORARY PARTITION SEPARATING THE COMMUNICATING NODES AT THE EXACT INSTANT THE MESSAGE IS LOST).”

Slide 31

Slide 31 text

“IN ORDER TO MODEL PARTITION TOLERANCE, THE NETWORK WILL BE ALLOWED TO LOSE ARBITRARILY MANY MESSAGES SENT FROM ONE NODE TO ANOTHER. WHEN A NETWORK IS PARTITIONED, ALL MESSAGES SENT FROM NODES IN ONE COMPONENT OF THE PARTITION TO NODES IN ANOTHER COMPONENT ARE LOST. (AND ANY PATTERN OF MESSAGE LOSS CAN BE MODELED AS A TEMPORARY PARTITION SEPARATING THE COMMUNICATING NODES AT THE EXACT INSTANT THE MESSAGE IS LOST).”

Slide 32

Slide 32 text

No content

Slide 33

Slide 33 text

?

Slide 34

Slide 34 text

No content

Slide 35

Slide 35 text

?

Slide 36

Slide 36 text

A B C D E

Slide 37

Slide 37 text

A B C D E

Slide 38

Slide 38 text

A B C D E

Slide 39

Slide 39 text

A B C D E

Slide 40

Slide 40 text

A B C D E

Slide 41

Slide 41 text

A B C D E

Slide 42

Slide 42 text

A B C D E

Slide 43

Slide 43 text

A B C D E

Slide 44

Slide 44 text

WRITES

Slide 45

Slide 45 text

A B C D E

Slide 46

Slide 46 text

A B C D E

Slide 47

Slide 47 text

B A C D E

Slide 48

Slide 48 text

B A C D E

Slide 49

Slide 49 text

B A C D E

Slide 50

Slide 50 text

B A C D E

Slide 51

Slide 51 text

READS

Slide 52

Slide 52 text

A B C D E

Slide 53

Slide 53 text

B A C D E

Slide 54

Slide 54 text

B A C D E

Slide 55

Slide 55 text

B A C D E

Slide 56

Slide 56 text

...

Slide 57

Slide 57 text

A B C D E

Slide 58

Slide 58 text

MAJORITY

Slide 59

Slide 59 text

A B C D E

Slide 60

Slide 60 text

A B C D E

Slide 61

Slide 61 text

CONSISTENT

Slide 62

Slide 62 text

N W R REPLICAS WRITE SUCCESS READ SUCCESS

Slide 63

Slide 63 text

(N/2) + 1

Slide 64

Slide 64 text

R+W > N

Slide 65

Slide 65 text

FAIL

Slide 66

Slide 66 text

A B C D E

Slide 67

Slide 67 text

A B C D E

Slide 68

Slide 68 text

YOU GET THE CHOICE

Slide 69

Slide 69 text

R+W <= N

Slide 70

Slide 70 text

A B C D E

Slide 71

Slide 71 text

W=N, R=1

Slide 72

Slide 72 text

R=N, W=1

Slide 73

Slide 73 text

TUNE FOR C OR A

Slide 74

Slide 74 text

CAN’T HAVE BOTH

Slide 75

Slide 75 text

CAN’T REALLY DO MUCH MORE ABOUT P

Slide 76

Slide 76 text

IT’S A BUSINESS DECISION

Slide 77

Slide 77 text

No content

Slide 78

Slide 78 text

No content

Slide 79

Slide 79 text

No content