docs/iterator-pipeline.puml


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45

@startuml iterator-pipeline
!theme plain
title Volcano Iterator Model - Process Pipeline

component "Scan Iterator\n(Process)" as P1 {
  [Read from Relation] as S1
}

component "Select Iterator\n(Process)" as P2 {
  [Apply Predicate] as S2
}

component "Project Iterator\n(Process)" as P3 {
  [Filter Attributes] as S3
}

component "Sort Iterator\n(Process)" as P4 {
  [Sort All Tuples] as S4
}

component "Take Iterator\n(Process)" as P5 {
  [Limit N Tuples] as S5
}

P1 -right-> P2: {next, Pid}\n{tuple, T} | done
P2 -right-> P3: {next, Pid}\n{tuple, T} | done
P3 -right-> P4: {next, Pid}\n{tuple, T} | done
P4 -right-> P5: {next, Pid}\n{tuple, T} | done

note bottom of P1
  Each iterator is an
  independent Erlang process

  Message passing protocol:
  - Request: {next, CallerPid}
  - Response: {tuple, Data} | done

  Benefits:
  - Lazy evaluation
  - Memory-bounded
  - Composable
  - Concurrent
end note

@enduml