Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
RICON West 2012: Bringing Consistency to Riak (...
Search
Joseph Blomstedt
October 30, 2013
Programming
3
750
RICON West 2012: Bringing Consistency to Riak (Part 2)
Joseph Blomstedt
October 30, 2013
Tweet
Share
More Decks by Joseph Blomstedt
See All by Joseph Blomstedt
Data Structure Adventures for Fun, Profit, and Performance
jtuple
2
340
Riak -- Seattle Scalability
jtuple
2
470
Hansei: Property-based Development of Concurrent Systems
jtuple
3
2.1k
Riak 1.2 Webinar Preview
jtuple
1
270
Test First Construction of Distributed Systems
jtuple
3
1.1k
Other Decks in Programming
See All in Programming
Go言語の特性を活かした公式MCP SDKの設計
hond0413
1
200
ИИ-Агенты в каждый дом – Алексей Порядин, PythoNN
sobolevn
0
150
なぜGoのジェネリクスはこの形なのか? Featherweight Goが明かす設計の核心
ryotaros
7
1k
Local Peer-to-Peer APIはどのように使われていくのか?
hal_spidernight
2
450
Breaking Up with Big ViewModels — Without Breaking Your Architecture (droidcon Berlin 2025)
steliosf
PRO
1
340
Things You Thought You Didn’t Need To Care About That Have a Big Impact On Your Job
hollycummins
0
160
クラシルを支える技術と組織
rakutek
0
190
Serena MCPのすすめ
wadakatu
4
900
詳しくない分野でのVibe Codingで困ったことと学び/vibe-coding-in-unfamiliar-area
shibayu36
3
4.5k
私はどうやって技術力を上げたのか
yusukebe
43
17k
iOSアプリの信頼性を向上させる取り組み/ios-app-improve-reliability
shino8rayu9
0
150
Domain-centric? Why Hexagonal, Onion, and Clean Architecture Are Answers to the Wrong Question
olivergierke
1
260
Featured
See All Featured
Practical Tips for Bootstrapping Information Extraction Pipelines
honnibal
PRO
23
1.5k
Dealing with People You Can't Stand - Big Design 2015
cassininazir
367
27k
Distributed Sagas: A Protocol for Coordinating Microservices
caitiem20
333
22k
Reflections from 52 weeks, 52 projects
jeffersonlam
352
21k
Fireside Chat
paigeccino
40
3.7k
Making the Leap to Tech Lead
cromwellryan
135
9.5k
Why Our Code Smells
bkeepers
PRO
339
57k
Docker and Python
trallard
46
3.6k
The Art of Delivering Value - GDevCon NA Keynote
reverentgeek
15
1.7k
The MySQL Ecosystem @ GitHub 2015
samlambert
251
13k
Designing for humans not robots
tammielis
254
25k
Sharpening the Axe: The Primacy of Toolmaking
bcantrill
45
2.5k
Transcript
Joseph Blomstedt (@jtuple) Basho Technologies Bringing Consistency To Riak (Part
2) Tuesday, October 29, 13
CAP Theorem 2 Tuesday, October 29, 13
3 Partition-tolerance Consistency Availability Tuesday, October 29, 13
4 Partition-tolerance Consistency Availability Tuesday, October 29, 13
5 Partition-tolerance Consistency Availability CP AP Tuesday, October 29, 13
6 Partition-tolerance Consistency Availability CP AP Tuesday, October 29, 13
7 Partition-tolerance Consistency Availability CP AP Tuesday, October 29, 13
8 C/P Strict Quorum A/P Sloppy Quorum A/P Tuesday, October
29, 13
9 C/P Strict Quorum A/P Sloppy Quorum A/P Tuesday, October
29, 13
10 Node 1 Node 2 Node 3 Node 4 Node
5 client client client Tuesday, October 29, 13
11 Node 1 Node 2 Node 3 Node 4 Node
5 client client client Tuesday, October 29, 13
12 Node 1 Node 2 Node 3 Node 4 Node
5 client client client Tuesday, October 29, 13
13 Node 1 Node 2 Node 3 Node 4 Node
5 client client client Tuesday, October 29, 13
14 C/P Strict Quorum A/P Sloppy Quorum A/P Tuesday, October
29, 13
15 Node 1 Node 2 Node 3 Node 4 Node
5 client client client Tuesday, October 29, 13
16 Node 1 Node 2 Node 3 Node 4 Node
5 client client client Tuesday, October 29, 13
17 Node 1 Node 2 Node 3 Node 4 Node
5 client client client Tuesday, October 29, 13
18 Node 1 Node 2 Node 3 Node 4 Node
5 client client client Tuesday, October 29, 13
19 C/P Strict Quorum A/P Sloppy Quorum A/P Tuesday, October
29, 13
20 Node 1 Node 2 Node 3 Node 4 Node
5 client client client Tuesday, October 29, 13
21 Node 1 Node 2 Node 3 Node 4 Node
5 client client client Tuesday, October 29, 13
22 Node 1 Node 2 Node 3 Node 4 Node
5 client client client Tuesday, October 29, 13
23 Node 1 Node 2 Node 3 Node 4 Node
5 client client client Tuesday, October 29, 13
24 Node 1 Node 2 Node 3 Node 4 Node
5 client client client client client Tuesday, October 29, 13
Eventual Consistency 25 Tuesday, October 29, 13
26 A A A Tuesday, October 29, 13
27 A A A Tuesday, October 29, 13
28 A A A B Tuesday, October 29, 13
29 A A A B Tuesday, October 29, 13
30 A A A B B B B Tuesday, October
29, 13
31 A A A Tuesday, October 29, 13
32 A A A B C Tuesday, October 29, 13
33 A A A B C Tuesday, October 29, 13
34 A A A B {B,C} {B,C} {B,C} C Tuesday,
October 29, 13
35 Write Once Immutable Last Write Wins Business Rules CRDTs/Monotonicity
Tuesday, October 29, 13
36 Write Once Immutable Last Write Wins Business Rules CRDTs/Monotonicity
Tuesday, October 29, 13
37 Write Once Immutable Last Write Wins Business Rules CRDTs/Monotonicity
Tuesday, October 29, 13
38 Write Once Immutable Last Write Wins Business Rules CRDTs/Monotonicity
Tuesday, October 29, 13
39 Write Once Immutable Last Write Wins Business Rules CRDTs/Monotonicity
Tuesday, October 29, 13
40 Write Once Immutable Last Write Wins Business Rules CRDTs/Monotonicity
Tuesday, October 29, 13
Strong Consistency 41 Tuesday, October 29, 13
Strong Consistency 42 Why? Tuesday, October 29, 13
Strong Consistency 43 Recency Tuesday, October 29, 13
Strong Consistency 44 Recency Partial Writes Tuesday, October 29, 13
Strong Consistency 45 Recency Partial Writes Atomicity Tuesday, October 29,
13
46 Recency Partial Writes Atomicity Tuesday, October 29, 13
47 Recency Partial Writes Atomicity Tuesday, October 29, 13
48 Eventual consistency is great Tuesday, October 29, 13
49 But, when is eventual? Tuesday, October 29, 13
50 Do I have the most recent value? Tuesday, October
29, 13
51 CRDTs don’t help Tuesday, October 29, 13
52 (a,1) (a,1) (a,1) =1 Tuesday, October 29, 13
53 (a,1) (a,1) (a,1) Tuesday, October 29, 13
54 (a,1) +1 +3 (a,1) (a,1) (a,2) (a,1),(b,3) =2 =4
Tuesday, October 29, 13
55 (a,1) +1 +3 (a,1) (a,1) (a,2) (a,1),(b,3) Tuesday, October
29, 13
56 (a,1) +1 +3 (a,1) (a,1) (a,2) (a,1),(b,3) (a,2),(b,3) (a,2),(b,3)
(a,2),(b,3) Tuesday, October 29, 13
57 (a,1) +1 +3 (a,1) (a,1) (a,2) (a,1),(b,3) (a,2),(b,3) (a,2),(b,3)
(a,2),(b,3) =5 Tuesday, October 29, 13
58 (a,1) +1 +3 (a,1) (a,1) (a,2) (a,1),(b,3) =2 =4
Tuesday, October 29, 13
59 Recency Partial Writes Atomicity Tuesday, October 29, 13
60 A write B (fail) A A B A A
Tuesday, October 29, 13
61 B A A Tuesday, October 29, 13
62 B A A read A read A read A
Tuesday, October 29, 13
63 B A A read A read A read A
Tuesday, October 29, 13
64 B A A read A read A read A
read B Tuesday, October 29, 13
65 Recency Partial Writes Atomicity Tuesday, October 29, 13
Strong Consistency 66 Tuesday, October 29, 13
Strong Consistency 67 What does mean for Riak 2.0? Tuesday,
October 29, 13
68 Conditional single key atomic operations Tuesday, October 29, 13
69 No siblings Tuesday, October 29, 13
70 get sees most recent put Tuesday, October 29, 13
71 get/modify/put fails if object changed Tuesday, October 29, 13
72 get/modify/put fails if object changed (eg. concurrent put) Tuesday,
October 29, 13
73 puts w/o vclock fails if object exists Tuesday, October
29, 13
74 partial writes resolved on read Tuesday, October 29, 13
75 Consensus Tuesday, October 29, 13
76 Paxos Tuesday, October 29, 13
77 1RGH 1RGH
1RGH 1 SUHSDUH 1 SURPLVH 1 9 % SURPLVH 1 9 & 9 1 I 9 $ 9 % 9 & FRPPLW 1 9 1 DFFHSW 1 Tuesday, October 29, 13
78 Rinse/repeat for each request Tuesday, October 29, 13
79 2 round trips/request Tuesday, October 29, 13
80 Multi-Paxos Tuesday, October 29, 13
81 First Request Tuesday, October 29, 13
82 1RGH 1RGH
1RGH 1 , SUHSDUH 1 , SURPLVH 1 , 9 % SURPLVH 1 , 9 & 9 1 I 9 $ 9 % 9 & FRPPLW 1 , 9 1 DFFHSW 1 , Tuesday, October 29, 13
83 Each Additional Request Tuesday, October 29, 13
84 1RGH 1RGH
1RGH , FRPPLW 1 , 9 DFFHSW 1 , Tuesday, October 29, 13
85 1 round trip/request (common case) Tuesday, October 29, 13
86 Problem Shipping entire state each request is expensive Tuesday,
October 29, 13
87 Solution Paxos + Replicated Log Tuesday, October 29, 13
88 Problem Now I have N problems Tuesday, October 29,
13
89 Log recovery Log trimming Rollup Snapshots Fault Recovery Tuesday,
October 29, 13
90 Choose your own adventure... Tuesday, October 29, 13
91 Better Solution Build log replication into protocol Tuesday, October
29, 13
92 Better Solution ZK Atomic Broadcast Raft Tuesday, October 29,
13
Zab 93 Tuesday, October 29, 13
94 Tuesday, October 29, 13
95 Tuesday, October 29, 13
96 Tuesday, October 29, 13
97 Tuesday, October 29, 13
Raft 98 Tuesday, October 29, 13
99 Tuesday, October 29, 13
100 raftconsensus.github.io Tuesday, October 29, 13
101 Text Tuesday, October 29, 13
Back to Riak 102 Tuesday, October 29, 13
103 Key/Value Keys are independent Active Anti-Entropy Tunable backends Tuesday,
October 29, 13
104 Each key is independent state Tuesday, October 29, 13
105 Simple multi-paxos per key Tuesday, October 29, 13
106 1B keys = 1B consensus groups? Tuesday, October 29,
13
107 No Tuesday, October 29, 13
108 Consensus group per preflist (replica set) Tuesday, October 29,
13
109 Emulate paxos per key Tuesday, October 29, 13
Node 0 Node 1 Node 2 Tuesday, October 29, 13
111 1 2 3 4 5 6 7 123 Tuesday,
October 29, 13
112 1 2 3 4 5 6 7 123 234
Tuesday, October 29, 13
113 1 2 3 4 5 6 7 123 234
345 Tuesday, October 29, 13
114 1 2 3 4 5 6 7 123 234
345 456 Tuesday, October 29, 13
115 1 2 3 4 5 6 7 123 234
345 456 567 ... Tuesday, October 29, 13
116 1 2 3 4 5 6 7 123 234
345 456 567 Ensembles ... Tuesday, October 29, 13
117 64 partition ring = 64 ensembles Tuesday, October 29,
13
118 Each Ensemble Elects leader Establishes epoch Supports get/put ops
Tuesday, October 29, 13
119 Establish a new epoch Tuesday, October 29, 13
120 1RGH 1RGH
1RGH 1 , SUHSDUH 1 , SURPLVH 1 , 9 % SURPLVH 1 , 9 & 9 1 I 9 $ 9 % 9 & FRPPLW 1 , 9 1 DFFHSW 1 , Tuesday, October 29, 13
121 consensus state epoch sequence membership leader Tuesday, October 29,
13
122 K/V objects epoch sequence key value Tuesday, October 29,
13
123 GET leader reads local object if obj.epoch old: refresh
reply w/ val Tuesday, October 29, 13
124 1RGH 1RGH
1RGH REMHSRFK HSRFK JHW .H\ UHSO\ (SRFK % 6HT % 9DO % UHSO\ (SRFK & 6HT & 9DO & 9DO ODWHVW 9DO $ 9DO % 9DO & 9DOHSRFK HSRFK ZULWH (SRFK 6HT 9DO DFN (SRFK 6HT Tuesday, October 29, 13
125 1RGH 1RGH
1RGH REMHSRFK HSRFK 5HSO\ ORFDOBJHW .H\ Tuesday, October 29, 13
126 2 roundtrips/get (worst) 0 roundtrip/get (best) Tuesday, October 29,
13
127 PUT leader reads local object if obj.epoch old: refresh
if modify(obj) false: fail commit modified obj reply ok Tuesday, October 29, 13
128 1RGH 1RGH
1RGH REMHSRFK HSRFK JHW .H\ UHSO\ (SRFK % 6HT % 9DO % UHSO\ (SRFK & 6HT & 9DO & /DWHVW ODWHVW 9DO $ 9DO % 9DO & 9DO PRGLI\ /DWHVW ZULWH (SRFK 6HT 9DO DFN (SRFK 6HT Tuesday, October 29, 13
129 1RGH 1RGH
1RGH REMHSRFK HSRFK /DWHVW ORFDOBJHW .H\ 9DO PRGLI\ /DWHVW ZULWH (SRFK 6HT 9DO DFN (SRFK 6HT Tuesday, October 29, 13
130 2 roundtrips/put (worst) 1 roundtrip/put (best) Tuesday, October 29,
13
131 Leader abandons leadership if any quorum operation ever fails
Tuesday, October 29, 13
132 Which forces new epoch to be established Tuesday, October
29, 13
133 Partial Writes Tuesday, October 29, 13
failed partial write X (2) X (2) X (2) X
(2) X (2) Y (2) epoch 2 epoch 3 Tuesday, October 29, 13
read / rewrite / reply X X (2) X (2)
Y (2) X (3) X (3) Y (2) epoch 3 epoch 3 Tuesday, October 29, 13
X (3) X (3) Y (2) X (3) X (3)
X (3) read / repair / reply X epoch 3 epoch 3 Tuesday, October 29, 13
Usage 137 Tuesday, October 29, 13
138 AP or CP per bucket type Tuesday, October 29,
13
139 consistent = true Tuesday, October 29, 13
140 $ riak-admin bucket-type create strong \ '{"props": {"consistent": true}}'
strong created Tuesday, October 29, 13
141 $ riak-admin bucket-type activate strong strong has been activated
Tuesday, October 29, 13
142 > riakc_pb_socket:get(Socket, {<<"strong">>, <<"bucket">>}, <<"key">>). {error,notfound} Tuesday, October 29,
13
143 > Obj = riakc_obj:new({<<"strong">>, <<"bucket">>}, <<"key">>, <<"1">>)). > riakc_pb_socket:put(Socket,
Obj). ok Tuesday, October 29, 13
144 > Obj2 = riakc_obj:new({<<"strong">>, <<"bucket">>}, <<"key">>, <<"2">>)). > riakc_pb_socket:put(Socket,
Obj2). {error, failed} Tuesday, October 29, 13
145 {ok, Obj3} = riakc_pb_socket:get(Socket, {<<"strong">>, <<"bucket">>}, <<"key">>). Tuesday, October
29, 13
146 Obj4 = riakc_obj:update_value(Obj3, <<"2">>). Tuesday, October 29, 13
147 Obj5 = riakc_obj:update_value(Obj3, <<"22">>). Tuesday, October 29, 13
148 > riakc_pb_socket:put(Socket, Obj4). ok Tuesday, October 29, 13
149 > riakc_pb_socket:put(Socket, Obj5). {error,<<"failed">>} Tuesday, October 29, 13
150 Your client may vary Tuesday, October 29, 13
151 Your client may vary We’re working on it Tuesday,
October 29, 13
Tech Preview 152 Tuesday, October 29, 13
153 No AAE syncing No 2i No stats Tuesday, October
29, 13
154 Will be in 2.0 final Tuesday, October 29, 13
Coming Soon 155 Tuesday, October 29, 13
156 Datatypes Multi-DC Lightweight Tx? Perf benchmarks Tuesday, October 29,
13
157 Datatypes Multi-DC Lightweight Tx? Perf benchmarks Tuesday, October 29,
13
158 Datatypes Multi-DC Lightweight Tx? Perf benchmarks Tuesday, October 29,
13
159 Datatypes Multi-DC Lightweight Tx? Perf benchmarks Tuesday, October 29,
13
160 Datatypes Multi-DC Lightweight Tx? Perf benchmarks Tuesday, October 29,
13
Questions? 161 Tuesday, October 29, 13