Distributed Garbage Collection Algorithms - PowerPoint PPT Presentation

About This Presentation
Title:

Distributed Garbage Collection Algorithms

Description:

no race conditions (latency) cons: susceptible to unreliable messaging. duplicated messages, ... based on on-the-fly garbage collector as proposed by [Dijkstra ... – PowerPoint PPT presentation

Number of Views:153
Avg rating:3.0/5.0
Slides: 34
Provided by: stefanbr2
Category:

less

Transcript and Presenter's Notes

Title: Distributed Garbage Collection Algorithms


1
Distributed Garbage Collection Algorithms
stefan brunthaler
2
Overview
  • Brief Introduction
  • Reference Counting
  • Weighted Reference Counting
  • Indirect Reference Counting
  • Mark Sweep
  • Marktree
  • Tracing with timestamps
  • Stop Copy

3
Introduction Definitions, 1
What is a distributed system? A system in
which hardware or software components located at
networked computers communicate and coordinate
their actions only by message passing. --Colour
is, 2001
4
Introduction Definitions, 2
What is a distributed system? A distributed
system is a collection of independent computers
that appear to the users of the system as a
single computer. --Tanenbaum, 1994
5
Introduction Nomenclature
A distributed system
A
B
C
6
Introduction Nomenclature
local objects
A
B
C
7
Introduction Nomenclature
global objects
A
B
C
8
Introduction Nomenclature
export record
A
B
C
9
Introduction Nomenclature
import record
A
B
C
10
Introduction Nomenclature
intrasite references
A
B
C
11
Introduction Nomenclature
intersite references
A
B
C
12
Introduction Networking
problems latency
A
B
C
13
Introduction Networking
problems non-causality
A
B
C
14
Reference Counting
  • Reference Counting
  • Weighted Reference Counting
  • Indirect Reference Counting

15
Weighted Reference Counting
  • proposed independently by Watson and Watson,
    1987, and Bevan, 1987

16
Weighted Reference Counting
Where are we?
A
B
C
17
Weighted Reference Counting
Memory cell layout
count
weight
ref
18
Weighted Reference Counting
Algorithm
19
Weighted Reference Counting
Algorithm
20
Weighted Reference Counting
Domino effect Rudalics, 1990
A
B
C
1
21
Weighted Reference Counting
  • pros
  • no race conditions (latency)
  • cons
  • susceptible to unreliable messaging
  • duplicated messages,
  • message failures.

22
Indirect Reference Counting
  • proposed by Ichisugi and Yonezawa, 1990,
    Rudalics, 1990, and Piquer, 1991

23
Indirect Reference Counting
Where are we?
A
B
C
24
Indirect Reference Counting
Memory cell layout
copies
parent
ref
25
Indirect Reference Counting
Algorithm
A
B
C
26
Indirect Reference Counting
inverted diffusion tree
2
1
2
1
0
0
27
Mark Sweep
  • Mark Sweep
  • Marktree
  • Tracing with timestamps

28
Marktree
  • after Hudak and Keller, 1982,
  • based on on-the-fly garbage collector as proposed
    by Dijkstra et al., 1978,
  • instead of recursion uses mark tasks,
  • back propagation using uptree tasks,
  • tricolor scheme.

29
Marktree
30
Tracing with timestamps
  • based on Hughes, 1985

31
Tracing with timestamps
Algorithm
ms
ms
ms
A
B
C
32
Stop Copy
  • based on Rudalics, 1986

rootspace
fromspace
tospace
33
Danke für Eure Aufmerksamkeit
Fragen?
Write a Comment
User Comments (0)
About PowerShow.com