Declarative, Secure, Convergent, Edge Computation

Erlang Factory 2016
San Francisco, CA

Transcript

1. Declarative, Secure, Convergent Edge Computation Christopher Meiklejohn Erlang Factory 2016, March 10th, 2016 1
2. Example Application Hospital Refrigerators 2
3. Hospital Refrigerators Typical Topology 3
5. Internet 5
6. Internet HDFS 6
7. Internet HDFS Hadoop Client Client Internet 7
8. Internet HDFS Hadoop Client Client Internet Spark Client Client Internet 8
9. Hospital Refrigerators Ideal Execution 9
10. Internet HDFS Spark 10
11. Internet HDFS Spark 35F 11
12. Internet HDFS Spark 35F 12
13. Internet HDFS Spark 35F 40F 13
14. Internet HDFS Spark 35F 40F 14
15. Client Internet HDFS Spark 35F 40F 15
16. Problem Connectivity 16
17. Internet HDFS Spark 17
18. Internet HDFS Spark 35F 18
19. Internet HDFS Spark 35F 19
20. Internet HDFS Spark 35F 40F 20
21. Internet HDFS Spark 35F 40F 21
22. Internet HDFS Spark 35F 40F 40F 22
23. Internet HDFS Spark 35F 40F 40F 23
24. Solution Local Decisions 24
25. Internet HDFS Spark 25
26. Internet HDFS Spark 35F 26
27. Internet HDFS Spark 35F 27
28. Internet HDFS Spark 35F 40F 28
29. Internet HDFS Spark 35F 40F 29
30. Solution Transitive Dissemination 30
31. Internet HDFS Spark 31
32. Internet HDFS Spark 35F 32
33. Internet HDFS Spark 35F 33
34. Internet HDFS Spark 35F 40F 34
35. Internet HDFS Spark 35F 40F 35
36. Internet HDFS Spark 35F 40F Client 36
37. Problem State Transmission 37
38. Internet 38
39. Internet 35F 39
40. Internet 35F 35F 35F 35F 40
41. Internet 35F 35F 35F 35F 41
42. Solution Aggregate Dissemination 42
43. Internet 43
44. Internet OK NO OK 44
45. Internet OK OK OK OK OK NO OK OK OK 45
46. Internet OK OK OK OK OK NO OK OK OK HDFS OK OK OK OK OK OK OK OK NO 46
47. Internet OK OK OK OK OK NO OK OK OK OK OK OK OK OK OK OK OK NO = = HDFS OK OK OK OK OK OK OK NO 47
48. Internet OK OK OK OK OK NO OK OK OK ? OK OK OK OK OK OK OK OK NO = = 48
49. Internet OK OK OK OK OK NO OK OK OK ? OK OK OK OK OK OK OK OK NO = = 49
50. Local Computation • Reduce state transmission
 Perform some local computation to reduce transmitted state on the wire 50
51. Local Computation • Reduce state transmission
 Perform some local computation to reduce transmitted state on the wire • Make local decisions
 Make decisions based on results of local computation 50
52. Databases Consistency Models 51
53. Databases Strong Consistency 52
54. R1 C1 C2 53
55. R1 C1 C2 Value 1 54
56. R1 C1 C2 Read Value 1 Value 1 Value 1 55
57. R1 C1 C2 Read Value 1 Value 1 Value 1 Value 1 56
58. R1 C1 C2 CAS Value 1 with Value 2 Value 2 Value 1 Value 2 57
59. R1 C1 C2 CAS Value 1 with Value 3 Value 2 Value 1 Value 2 58
60. I won’t diagram the Paxos protocol 59
61. R2 C1 C2 Value 2 Value 1 Value 2 R1 R3 Paxos 60
62. Databases Eventual Consistency 61
63. R1 R2 R3 C1 C2 62
64. R1 R2 R3 C1 C2 Value 1 @ [1, 0, 0] Value 1 @ [1, 0, 0] Value 1 @ [1, 0, 0] 63
65. R1 R2 R3 C1 C2 Value 1 @ [1, 0, 0] Value 1 @ [1, 0, 0] Value 1 @ [1, 0, 0] Value 1 @ [1, 0, 0] 64
66. R1 R2 R3 C1 C2 Value 1 @ [1, 0, 0] Value 1 @ [1, 0, 0] Value 1 @ [1, 0, 0] Value 1 @ [1, 0, 0] Value 1 @ [1, 0, 0] 65
67. R1 R2 R3 C1 C2 Value 1 @ [1, 0, 0] Value 1 @ [1, 0, 0] Value 1 @ [1, 0, 0] Value 1 @ [1, 0, 0] Value 1 @ [1, 0, 0] Write 2 66
68. R1 R2 R3 C1 C2 Value 2 @ [2, 0, 0] Value 2 @ [2, 0, 0] Value 2 @ [2, 0, 0] Value 2 @ [2, 0, 0] Value 1 @ [1, 0, 0] Write 2 Write Value 2 @ [2, 0, 0] 67
69. R1 R2 R3 C1 C2 Value 2 @ [2, 0, 0] Value 3 @ [1, 0, 1] Value 2 @ [2, 0, 0] Value 3 @ [1, 0, 1] Value 2 @ [2, 0, 0] Value 3 @ [1, 0, 1] Value 2 @ [2, 0, 0] Value 2 @ [2, 0, 0] Value 3 @ [1, 0, 1] Write 3 68
70. R1 R2 R3 C1 C2 Value 2 @ [2, 0, 0] Value 3 @ [1, 0, 1] Value 2 @ [2, 0, 0] Value 3 @ [1, 0, 1] Value 2 @ [2, 0, 0] Value 3 @ [1, 0, 1] Value 2 @ [2, 0, 0] Value 3 @ [1, 0, 1] Value 2 @ [2, 0, 0] Value 3 @ [1, 0, 1] 69
71. R1 R2 R3 C1 C2 Value 4 @ [2, 0, 1] Value 4 @ [2, 0, 1] Value 4 @ [2, 0, 1] Value 4 @ [2, 0, 1] Value 2 @ [2, 0, 0] Value 3 @ [1, 0, 1] Write 4 70
72. Eventual Consistency As The Model 71
73. Clients Own Their Data 72
74. OK OK OK 73
75. OK OK OK OK OK OK OK OK OK 74
76. OK OK OK OK OK OK OK OK OK 75
77. OK OK OK OK OK OK OK 76
78. OK OK OK OK OK OK OK OK OK 77
79. Computations Mergability & Provenance 78
80. A A B C 79
81. A A B C D’ D’’ F F 80
82. A A B C D’ D’’ D’ D’’ D   F F 81
83. D’’ A A B C D’ D’ D’’ D   D D Merge F F 82
84. Example Application Preliminary Results 83
85. Preliminary Results • Conflict-Free Replicated Data Types
 Distributed data structures designed for convergence 
 [Shapiro et al., 2011] 84
86. Preliminary Results • Conflict-Free Replicated Data Types
 Distributed data structures designed for convergence 
 [Shapiro et al., 2011] • Lattice Processing
 Make decisions based on results of local computation
 [Meiklejohn & Van Roy, 2015] 84
87. Preliminary Results • Conflict-Free Replicated Data Types
 Distributed data structures designed for convergence 
 [Shapiro et al., 2011] • Lattice Processing
 Make decisions based on results of local computation
 [Meiklejohn & Van Roy, 2015] • Selective Hearing
 Epidemic broadcast based runtime system
 [Meiklejohn & Van Roy, 2015/2016] 84
88. Conflict-Free 
 Replicated Data Types • Collection of types
 Sets, counters, registers, flags, maps 85
89. Conflict-Free 
 Replicated Data Types • Collection of types
 Sets, counters, registers, flags, maps • Strong Eventual Consistency
 Objects that receive the same updates, regardless of order, will reach equivalent state 85
90. RA RB RC
91. RA RB RC {1} (1, {a}, {}) add(1)
92. RA RB RC {1} (1, {a}, {}) add(1) {1} (1, {c}, {}) add(1)
93. RA RB RC {1} (1, {a}, {}) add(1) {1} (1, {c}, {}) add(1) {} (1, {c}, {c}) remove(1)
94. RA RB RC {1} (1, {a}, {}) add(1) {1} (1, {c}, {}) add(1) {} (1, {c}, {c}) remove(1) {1} {1} {1} (1, {a, c}, {c}) (1, {a, c}, {c}) (1, {a, c}, {c})
95. Lattice Processing • Distributed, deterministic dataflow
 Distributed, dataflow programming model 91
96. Lattice Processing • Distributed, deterministic dataflow
 Distributed, dataflow programming model • Convergent data structures
 Data abstraction is the CRDT 91
97. Lattice Processing • Distributed, deterministic dataflow
 Distributed, dataflow programming model • Convergent data structures
 Data abstraction is the CRDT • Enables composition
 Composition preserves SEC 91
98. 92 %% Create initial set. S1 = declare(set), %% Add elements to initial set and update. update(S1, {add, [1,2,3]}), %% Create second set. S2 = declare(set), %% Apply map operation between S1 and S2. map(S1, fun(X) -> X * 2 end, S2).
99. 93 %% Create initial set. S1 = declare(set), %% Add elements to initial set and update. update(S1, {add, [1,2,3]}), %% Create second set. S2 = declare(set), %% Apply map operation between S1 and S2. map(S1, fun(X) -> X * 2 end, S2).
100. 94 %% Create initial set. S1 = declare(set), %% Add elements to initial set and update. update(S1, {add, [1,2,3]}), %% Create second set. S2 = declare(set), %% Apply map operation between S1 and S2. map(S1, fun(X) -> X * 2 end, S2).
101. 95 %% Create initial set. S1 = declare(set), %% Add elements to initial set and update. update(S1, {add, [1,2,3]}), %% Create second set. S2 = declare(set), %% Apply map operation between S1 and S2. map(S1, fun(X) -> X * 2 end, S2).
102. 96 %% Create initial set. S1 = declare(set), %% Add elements to initial set and update. update(S1, {add, [1,2,3]}), %% Create second set. S2 = declare(set), %% Apply map operation between S1 and S2. map(S1, fun(X) -> X * 2 end, S2).
103. Selective Hearing • Epidemic broadcast protocol
 Runtime system for application state & scope 97
104. Selective Hearing • Epidemic broadcast protocol
 Runtime system for application state & scope • Peer-to-peer dissemination
 Pairwise synchronization between peers without a central coordinator 97
105. Selective Hearing • Epidemic broadcast protocol
 Runtime system for application state & scope • Peer-to-peer dissemination
 Pairwise synchronization between peers without a central coordinator • No ordering guarantees on messages
 Programming model can tolerate message reordering and duplication 97
106. What can we build? Leaderboard 98
107. Leaderboard • Mobile game platform
 Local leaderboard tracking top-k highest scored games 99
108. Leaderboard • Mobile game platform
 Local leaderboard tracking top-k highest scored games • Clients will go offline
 Clients have limited connectivity and the system still needs to make progress while clients are offline 99
109. Client 1 Leaderboard Client 3 Leaderboard Client 2 Leaderboard Lasp Operation User-Maintained CRDT Lasp-Maintained CRDT 100
110. Leaderboard • Peer-to-peer dissemination
 Nodes periodically “merge” their state with a random peer 101
111. Leaderboard • Peer-to-peer dissemination
 Nodes periodically “merge” their state with a random peer • Complexity in the data type
 Each node tracks a top-k set of its own games in a bounded set 101
112. 102 %% Create a leaderboard datatype. L = declare({top_k, [2]}). %% Update leaderboard. update({set, Name, Score}, L).
113. 103 %% Create a leaderboard datatype. L = declare({top_k, [2]}). %% Update leaderboard. update({set, Name, Score}, L).
114. 104 %% Create a leaderboard datatype. L = declare({top_k, [2]}). %% Update leaderboard. update({set, Name, Score}, L).
115. What if we want to enhance the behavior? 105
116. What if we want to enhance the behavior? 105 Without the creation of a new datatype
117. What can we build? Per-User Leaderboard 106 EdgeCom 2016
118. Per-User Leaderboard • Enhance existing design
 Only the top score for each user at each device 107
119. Per-User Leaderboard • Enhance existing design
 Only the top score for each user at each device • Minimize transmitted state
 Prevent transmission of state that is not necessary to perform the computation 107
120. Per-User Leaderboard • Enhance existing design
 Only the top score for each user at each device • Minimize transmitted state
 Prevent transmission of state that is not necessary to perform the computation • Compose data types
 Build a per-user leaderboard through the composition of existing types 107
121. Client1 Scores Local Top-K Fold Global Top-K Lasp Operation Input User-Maintained CRDT Output Lasp-Maintained CRDT Fold Client3 Scores Local Top-K Fold Global Top-K Fold Client2 Scores Local Top-K Fold Global Top-K Fold 108
122. Client1 Scores Local Top-K Fold Global Top-K Fold 109 Client1 Scores Local Top-K Fold Global Top-K Fold Client3 Scores Local Top-K Fold Global Top-K Fold Client2 Scores Local Top-K Fold Global Top-K Fold
123. Client1 Scores Local Top-K Fold Global Top-K Fold 110 Client1 Scores Local Top-K Fold Global Top-K Fold Client3 Scores Local Top-K Fold Global Top-K Fold Client2 Scores Local Top-K Fold Global Top-K Fold
124. Client1 Scores Local Top-K Fold Global Top-K Fold 111 Client1 Scores Local Top-K Fold Global Top-K Fold Client3 Scores Local Top-K Fold Global Top-K Fold Client2 Scores Local Top-K Fold Global Top-K Fold
125. Client1 Scores Local Top-K Fold Global Top-K Fold 112 Client1 Scores Local Top-K Fold Global Top-K Fold Client3 Scores Local Top-K Fold Global Top-K Fold Client2 Scores Local Top-K Fold Global Top-K Fold
126. Client1 Scores Local Top-K Fold Global Top-K Fold 113 Client1 Scores Local Top-K Fold Global Top-K Fold Client3 Scores Local Top-K Fold Global Top-K Fold Client2 Scores Local Top-K Fold Global Top-K Fold
127. Client1 Scores Local Top-K Fold Global Top-K Fold 114 Client1 Scores Local Top-K Fold Global Top-K Fold Client3 Scores Local Top-K Fold Global Top-K Fold Client2 Scores Local Top-K Fold Global Top-K Fold
128. Client1 Scores Local Top-K Fold Global Top-K Lasp Operation Input User-Maintained CRDT Output Lasp-Maintained CRDT Fold Client3 Scores Local Top-K Fold Global Top-K Fold Client2 Scores Local Top-K Fold Global Top-K Fold 115 Client1 Scores Local Top-K Fold Global Top-K Fold Client3 Scores Local Top-K Fold Global Top-K Fold Client2 Scores Local Top-K Fold Global Top-K Fold
129. Per-User Leaderboard • Dynamically scoped variables
 Variable which take different values depending on where it is executing 116
130. Per-User Leaderboard • Dynamically scoped variables
 Variable which take different values depending on where it is executing • Dynamically scoped fold operation
 Perform a distributed “reduce” operation that combines the state of a dynamically scoped variables across 116
131. 117 %% Create a global leaderboard. G = declare({top_k, [10]}). %% Create a local leaderboard. L = declare_dynamic({top_k, [10]}). %% Create a set of scores. S = declare_dynamic(set). %% Compute local top-k list. fold(S, fun max_by_name/2, L). %% Compute global top-k list. fold_dynamic(L, fun max_by_name/2, G).
132. 118 %% Create a global leaderboard. G = declare({top_k, [10]}). %% Create a local leaderboard. L = declare_dynamic({top_k, [10]}). %% Create a set of scores. S = declare_dynamic(set). %% Compute local top-k list. fold(S, fun max_by_name/2, L). %% Compute global top-k list. fold_dynamic(L, fun max_by_name/2, G).
133. 119 %% Create a global leaderboard. G = declare({top_k, [10]}). %% Create a local leaderboard. L = declare_dynamic({top_k, [10]}). %% Create a set of scores. S = declare_dynamic(set). %% Compute local top-k list. fold(S, fun max_by_name/2, L). %% Compute global top-k list. fold_dynamic(L, fun max_by_name/2, G).
134. 120 %% Create a global leaderboard. G = declare({top_k, [10]}). %% Create a local leaderboard. L = declare_dynamic({top_k, [10]}). %% Create a set of scores. S = declare_dynamic(set). %% Compute local top-k list. fold(S, fun max_by_name/2, L). %% Compute global top-k list. fold_dynamic(L, fun max_by_name/2, G).
135. 121 %% Create a global leaderboard. G = declare({top_k, [10]}). %% Create a local leaderboard. L = declare_dynamic({top_k, [10]}). %% Create a set of scores. S = declare_dynamic(set). %% Compute local top-k list. fold(S, fun max_by_name/2, L). %% Compute global top-k list. fold_dynamic(L, fun max_by_name/2, G).
136. 122 %% Create a global leaderboard. G = declare({top_k, [10]}). %% Create a local leaderboard. L = declare_dynamic({top_k, [10]}). %% Create a set of scores. S = declare_dynamic(set). %% Compute local top-k list. fold(S, fun max_by_name/2, L). %% Compute global top-k list. fold_dynamic(L, fun max_by_name/2, G).
137. Let’s look at a larger application 123
138. Let’s look at a larger application 123 With visible 
 non-monotonicity
139. What can we build? Advertisement Counter 124
140. Advertisement Counter • Mobile game platform selling advertisement space
 Advertisements are paid according to a minimum number of impressions 125
141. Advertisement Counter • Mobile game platform selling advertisement space
 Advertisements are paid according to a minimum number of impressions • Clients will go offline
 Clients have limited connectivity and the system still needs to make progress while clients are offline 125
142. Ads Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Riot Ad Counter 2 Contracts Ads Contracts Ads With Contracts Riot Ads Rovio Ads Filter Product Read 50,000 Remove Increment Read Union Lasp Operation User-Maintained CRDT Lasp-Maintained CRDT Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Client Side, Single Copy at Client 126
143. Ads Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Riot Ad Counter 2 Contracts Ads Contracts Riot Ads Rovio Ads Product Read 50,000 Remove Increment Union 127 Ads Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Riot Ad Counter 2 Contracts Ads Contracts Ads With Contracts Riot Ads Rovio Ads Filter Product Read 50,000 Remove Increment Read Union Lasp Operation User-Maintained CRDT Lasp-Maintained CRDT Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Client Side, Single Copy at Client
144. Ads Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Riot Ad Counter 2 Contracts Ads Contracts Ads With Contracts Riot Ads Rovio Ads Filter Product Read 50,000 Remove Increment Read Union Rovio Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 1 Client 128 Ads Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Riot Ad Counter 2 Contracts Ads Contracts Ads With Contracts Riot Ads Rovio Ads Filter Product Read 50,000 Remove Increment Read Union Lasp Operation User-Maintained CRDT Lasp-Maintained CRDT Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Client Side, Single Copy at Client
145. Ads ovio Ad ounter 1 ovio Ad ounter 2 Riot Ad ounter 1 Riot Ad ounter 2 Contracts Ads Contracts Ads With Contracts Riot Ads Rovio Ads Filter Product Read 50,000 Remove Increment Read Union Rovio Ad Counter 1 Ro C Rovio Ad Counter 1 Ro C Rovio Ad Counter 1 Ro C Rovio Ad Counter 1 Ro C Client Side, Sing 129 Ads Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Riot Ad Counter 2 Contracts Ads Contracts Ads With Contracts Riot Ads Rovio Ads Filter Product Read 50,000 Remove Increment Read Union Lasp Operation User-Maintained CRDT Lasp-Maintained CRDT Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Client Side, Single Copy at Client
146. Ads Contracts Ads Contracts Ads With Contracts Riot Ads Rovio Ads Filter Product move Read Union Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Client Side, Single Copy at Client 130 Ads Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Riot Ad Counter 2 Contracts Ads Contracts Ads With Contracts Riot Ads Rovio Ads Filter Product Read 50,000 Remove Increment Read Union Lasp Operation User-Maintained CRDT Lasp-Maintained CRDT Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Client Side, Single Copy at Client
147. Ads Contracts Ads Contracts Ads With Contracts Filter Product Read Union Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Client Side, Single Copy at Client 131 Ads Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Riot Ad Counter 2 Contracts Ads Contracts Ads With Contracts Riot Ads Rovio Ads Filter Product Read 50,000 Remove Increment Read Union Lasp Operation User-Maintained CRDT Lasp-Maintained CRDT Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Client Side, Single Copy at Client
148. Ads Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Riot Ad Counter 2 Contracts Ads Contracts Ads With Contracts Riot Ads Rovio Ads Filter Product Read 50,000 Remove Increment Read Union Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Client Side, Single Copy at Client 132 Ads Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Riot Ad Counter 2 Contracts Ads Contracts Ads With Contracts Riot Ads Rovio Ads Filter Product Read 50,000 Remove Increment Read Union Lasp Operation User-Maintained CRDT Lasp-Maintained CRDT Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Client Side, Single Copy at Client
149. Ads Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Riot Ad Counter 2 Contracts Ads Contracts Riot Ads Rovio Ads Fil Product Read 50,000 Remove Increment Union 133 Ads Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Riot Ad Counter 2 Contracts Ads Contracts Ads With Contracts Riot Ads Rovio Ads Filter Product Read 50,000 Remove Increment Read Union Lasp Operation User-Maintained CRDT Lasp-Maintained CRDT Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Client Side, Single Copy at Client
150. Ads Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Riot Ad Counter 2 Contracts Ads Contracts Ads With Contracts Riot Ads Rovio Ads Filter Product Read 50,000 Remove Increment Read Union Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Client Side, Single Copy at Client 134 Ads Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Riot Ad Counter 2 Contracts Ads Contracts Ads With Contracts Riot Ads Rovio Ads Filter Product Read 50,000 Remove Increment Read Union Lasp Operation User-Maintained CRDT Lasp-Maintained CRDT Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Rovio Ad Counter 1 Rovio Ad Counter 2 Riot Ad Counter 1 Client Side, Single Copy at Client
151. Advertisement Counter • Completely monotonic
 Disabling advertisements and contracts are all modeled through monotonic state growth 135
152. Advertisement Counter • Completely monotonic
 Disabling advertisements and contracts are all modeled through monotonic state growth • Arbitrary distribution
 Use of convergent data structures allows computational graph to be arbitrarily distributed 135
153. Advertisement Counter • Completely monotonic
 Disabling advertisements and contracts are all modeled through monotonic state growth • Arbitrary distribution
 Use of convergent data structures allows computational graph to be arbitrarily distributed • Divergence
 Divergence is a factor of synchronization period 135
154. Client3 Lasp Operation User-Maintained CRDT Lasp-Maintained CRDT Riot Ad Counter 1 Client1 Client2 Rovio Ad Counter 1 Rovio Ad Counter 2 Rovio Ad Counter 1 Rovio Ad Counter 2 Rovio Ad Counter 1 Riot Ad Counter 2 Riot Ad Counter 1 Riot Ad Counter 1 Ads With Contracts Ads With Contracts Ads With Contracts Server Ads With Contracts Server Computation! 136
155. Client3 Lasp Operation User-Maintained CRDT Lasp-Maintained CRDT Riot Ad Counter 1 Client1 Client2 Rovio Ad Counter 1 Rovio Ad Counter 2 Rovio Ad Counter 1 Rovio Ad Counter 2 Rovio Ad Counter 1 Riot Ad Counter 2 Riot Ad Counter 1 Riot Ad Counter 1 Ads With Contracts Ads With Contracts Ads With Contracts Server Ads With Contracts Server Computation! 137
156. Client3 Lasp Operation User-Maintained CRDT Lasp-Maintained CRDT Riot Ad Counter 1 Client1 Client2 Rovio Ad Counter 1 Rovio Ad Counter 2 Rovio Ad Counter 1 Rovio Ad Counter 2 Rovio Ad Counter 1 Riot Ad Counter 2 Riot Ad Counter 1 Riot Ad Counter 1 Ads With Contracts Ads With Contracts Ads With Contracts Server Ads With Contracts Server Computation! 138
157. Advertisement Counter • “Servers” as peers to “clients”
 Servers are peers to clients that perform additional computation 139
158. Advertisement Counter • “Servers” as peers to “clients”
 Servers are peers to clients that perform additional computation • Any node can disable an advertisement under this model given enough information 139
159. Advertisement Counter • “Servers” as peers to “clients”
 Servers are peers to clients that perform additional computation • Any node can disable an advertisement under this model given enough information • “Servers” as trusted nodes
 Serve as a location for performing “exactly once” side- effects 139
160. Advertisement Counter • “Servers” as peers to “clients”
 Servers are peers to clients that perform additional computation • Any node can disable an advertisement under this model given enough information • “Servers” as trusted nodes
 Serve as a location for performing “exactly once” side- effects • Billing customers must be done at a central point by a trusted node in the system 139
161. We’ve build up from zero synchronization 140
162. We’ve build up from zero synchronization 140 Instead of working to remove synchronization
163. Challenges Looking Ahead 141
164. Causality State Explosion 142
165. Set Counter
166. Set Counter {1} (1, {c}, {}) 1 {(c, 1)}, {}
167. Set Counter {1} (1, {c}, {}) 1 {(c, 1)}, {} {} (1, {c}, {c}) {(c, 1)}, {(c, 1)} 0
168. Set Counter {1} (1, {c}, {}) 1 {(c, 1)}, {} {} (1, {c}, {c}) {(c, 1)}, {(c, 1)} 0 {1} (1, {a, c}, {c}) {1} 1 1 {(c, 1), (a, 1}}, {(c, 1)}
169. Security Computing at the Edge 147
170. OK OK OK 148
171. Computations Expressiveness 149
172. How restrictive is a programming model where operations must be associative, commutative, and idempotent? 150
173. What’s new? 151
174. Since Erlang Factory 2015 • Erlang 18, rebar3, Common Test
 Faster test suite, on Erlang 18 with rebar3 152
175. Since Erlang Factory 2015 • Erlang 18, rebar3, Common Test
 Faster test suite, on Erlang 18 with rebar3 • No Riak Core
 Plumtree distribution replaces the Riak Core distribution system 152
176. Since Erlang Factory 2015 • Erlang 18, rebar3, Common Test
 Faster test suite, on Erlang 18 with rebar3 • No Riak Core
 Plumtree distribution replaces the Riak Core distribution system • No NIFs
 Allows us to cross-compile to other platforms and not worry about NIF scheduling 152
177. Semantics Improvements • Delta-State Based Conflict-Free Replicated Data Types
 Optimized state dissemination by shipping deltas 
 [Almeida et al., 2016] 153
178. Semantics Improvements • Delta-State Based Conflict-Free Replicated Data Types
 Optimized state dissemination by shipping deltas 
 [Almeida et al., 2016] • Causal CRDTs
 Optimized, garbage-free data structure support
 [Almeida et al., 2016; Meiklejohn 2016, in review] 153
179. Semantics Improvements • Delta-State Based Conflict-Free Replicated Data Types
 Optimized state dissemination by shipping deltas 
 [Almeida et al., 2016] • Causal CRDTs
 Optimized, garbage-free data structure support
 [Almeida et al., 2016; Meiklejohn 2016, in review] • Fold
 New semantics for a more expressive fold operation for arbitrary computation over sets
 [Meiklejohn 2016, in review] 153
180. Runtime Improvements • Delta-State Based Anti-Entropy
 Optimized AAE mechanism based on deltas
 [Almeida et al., 2016] 154
181. Runtime Improvements • Delta-State Based Anti-Entropy
 Optimized AAE mechanism based on deltas
 [Almeida et al., 2016] • Mesos and Docker Enabled
 Run large-scale Lasp clusters on Mesos with Marathon
 [Meiklejohn and Yoo 2016, in review] 154
182. Runtime Improvements • Delta-State Based Anti-Entropy
 Optimized AAE mechanism based on deltas
 [Almeida et al., 2016] • Mesos and Docker Enabled
 Run large-scale Lasp clusters on Mesos with Marathon
 [Meiklejohn and Yoo 2016, in review] • Loquat
 Epidemic broadcast, partially replicated with Decentralized Information Flow Control
 [Meiklejohn 2016, in review] 154
183. Lasp Simulator • Docker Containers
 Docker containers for EPMD and Lasp runtime system 155
184. Lasp Simulator • Docker Containers
 Docker containers for EPMD and Lasp runtime system • Service Discovery
 Mechanisms for clustering Erlang nodes based on either Marathon application definitions of Mesos-DNS 155
185. Lasp Simulator • Docker Containers
 Docker containers for EPMD and Lasp runtime system • Service Discovery
 Mechanisms for clustering Erlang nodes based on either Marathon application definitions of Mesos-DNS • Instrumentation
 Transmission instrumentation and divergence measurement 155
186. Lasp Simulator • Plumtree VM-to-VM
 VM-to-VM communication performed using the Plumtree epidemic broadcast protocol 156
187. Lasp Simulator • Plumtree VM-to-VM
 VM-to-VM communication performed using the Plumtree epidemic broadcast protocol • Clients-as-processes
 Multiple clients per virtual machine, acting as mobile/IoT devices that periodically simulate and can be partitioned 156
188. Lasp Simulator • Plumtree VM-to-VM
 VM-to-VM communication performed using the Plumtree epidemic broadcast protocol • Clients-as-processes
 Multiple clients per virtual machine, acting as mobile/IoT devices that periodically simulate and can be partitioned • “Design in the small, run in the large”
 Runtime configuration change for client-to-VM ratio, allows for single laptop design of multi-machine evaluations of the programming model 156
190. �� ��� ��� ��� ���� � �� ��� ��� ��� ��� ��� ���������� �� � �� ������ ���� �� ������� ������� ���������� ��������� ����� ������� ���������� ��������� �� 158
191. ������� ������ ����� ���� ��� � � � � �� ��� ��� ��� ��� ��� ��������� ����������� ���� �� ������� ������� ��������� ������� ���������� ������ 159
192. What’s next? 160
193. Google Summer of Code (and my Ph.D.!) • Partial Evaluation
 Optimize execution based on analysis and annotations where we can determine local-vs-remote usage 161
194. Google Summer of Code (and my Ph.D.!) • Partial Evaluation
 Optimize execution based on analysis and annotations where we can determine local-vs-remote usage • Optimizations
 General optimizations of the Erlang implementation of Lasp to improve performance of the runtime system 161
195. Google Summer of Code (and my Ph.D.!) • Partial Evaluation
 Optimize execution based on analysis and annotations where we can determine local-vs-remote usage • Optimizations
 General optimizations of the Erlang implementation of Lasp to improve performance of the runtime system • Elixir / Macros
 Provide a nicer way for working with Lasp, outside of the current syntax. 161
196. How do I learn more? 162
197. Publications • “Lasp: A Language for Distributed, Coordination-Free Programming” 
 ACM SIGPLAN PPDP 2015 • “Selective Hearing: An Approach to Distributed, Eventually Consistent Edge Computation”
 IEEE W-PSDS 2015 • “The Implementation and Use of a Generic Dataflow Behaviour in Erlang”
 ACM SIGPLAN Erlang Workshop ’15 • “Lasp: A Language for Distributed, Eventually Consistent Computations with CRDTs"
 PaPoC 2015 • “Declarative, Sliding Window Aggregations for Computations at the Edge"
 IEEE EdgeCom 2016 163
198. 164 Christopher Meiklejohn @cmeik http://www.lasp-lang.org Thanks!