Scribd
Upload
52 views15 pages

Distributed Progrms

This document outlines distributed programs and their semantics. It discusses how distributed programs can handle concurrency without interference and scale to large systems. The syntax and semantics of distributed programs are defined, including how communication between processes is modeled. Methods for proving properties like partial correctness, weak total correctness, and total correctness are presented. Activities are included to apply the concepts to examples.

Uploaded by

Sanapala Amit
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
52 views15 pages

Distributed Progrms

This document outlines distributed programs and their semantics. It discusses how distributed programs can handle concurrency without interference and scale to large systems. The syntax and semantics of distributed programs are defined, including how communication between processes is modeled. Methods for proving properties like partial correctness, weak total correctness, and total correctness are presented. Activities are included to apply the concepts to examples.

Uploaded by

Sanapala Amit
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 15

.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Distributed Programs
Dr. Mattox Beckman
Illinois Institute of Technology
Department of Computer Science
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Outline
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Motivation

Distributed programs are an amazing way of handling concurrency.

No interference!

Scales well to large systems.


.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Syntax

A sequential process has the form


S S
0
; do
m
j1
g
j
S
j
od
where

m 0 and S
0
, . . . , S
n
are nondeterminstic programs.

S
0
is the initialization code.

There may be loops inside of each S


i
.

g B;

B is a guard

is of form c?u an input command, or

c!t an output command.


.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Example
SR [SENDERRECIVER] (1)
SENDER i := 0; do i = |a|; link!a[i] i := i + 1 od (2)
RECEIVER j := 0; do j = |b|; link?b[j] j := j + 1 od (3)
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Activity
Do the Distributed Program Activity
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Semantics

First we talk about termination.


< do
m
j=1
g
j
S
j
od , >< E, >
where for j {1, . . . , n} g
j
B
j
;
j
and |=
m
j=1
B
j

We also want to talk about the meaning of a communication.

1
= c?u

2
= c!t
Eff(
1
,
2
) = Eff(
2
,
1
) = u := t
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Communication
< [S
1
S
n
], >< [S

1
S

n
], >
S
k
= do
m
1
j=1
g
j
R
j
od
S
l
= do
m
2
j=1
h
j
T
j
od
g
j
1
= B
1
;
1
h
j
2
= B
2
;
2
|= B
1
B
2
M[[Eff(
1
,
2
)]]() = {}
S

i
S
i
for i = k, l
S

k
R
j
1
; S
k
S

l
T
j
2
; S
l
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Transforming to a nondeterministic program

Suppose each process has form


S
i
S
i,0
; do
m
i
j=1
B
i,j
;
i,j
S
i,j
od

Let be a list of matching communications:


= {(i, j, k, l)|
i,j
and
k,l
match and i < k}

Then S [S
1
S
n
] can be transformed as...
T(S) S
1,0
; S
2,0
; ; S
n,0
;
do
(i,j,k,l)
B
i,j
B
k,l
Eff(
i,j
,
k,l
); S
i,j
; S
k,l
od
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Activity
Do the Distributed Conversion Activity
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Proving Deadlock

The partial correctness will be the same, but what about deadlock?

Given
T(S) S
1,0
; S
2,0
; ; S
n,0
;
do
(i,j,k,l)
B
i,j
B
k,l
Eff(
i,j
,
k,l
); S
i,j
; S
k,l
od

We dene TERM and BLOCK. How are they different?


TERM

n
i=1

m
i
j=1
B
i,j
BLOCK

(i,j,k,l)
(B
i,j
B
k,l
)

Which is the stronger of TERM and BLOCK?

What does BLOCK TERM tell us?


.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Partial Correctness: Rule 19
{p}S
1,0
; ; S
n,0
{I},
{I B
i,j
B
k,l
}Eff(
i,j
,
k,l
); S
i,j
; S
k,l
{I}
for all (i, j, k, l)
{p}S{I TERM}
What do we have to do to add weak total correctness?
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Weak Total Correctness: Rule 20
{p}S
1,0
; ; S
n,0
{I},
{I B
i,j
B
k,l
}Eff(
i,j
,
k,l
); S
i,j
; S
k,l
{I}
for all (i, j, k, l)
{I B
i,j
B
k,l
t = z}Eff(
i,j
,
k,l
); S
i,j
; S
k,l
{t < z}
for all (i, j, k, l)
I t 0
{p}S{I TERM}
What do we have to do to add total correctness?
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Total Correctness: Rule 21
{p}S
1,0
; ; S
n,0
{I},
{I B
i,j
B
k,l
}Eff(
i,j
,
k,l
); S
i,j
; S
k,l
{I}
for all (i, j, k, l)
{I B
i,j
B
k,l
t = z}Eff(
i,j
,
k,l
); S
i,j
; S
k,l
{t < z}
for all (i, j, k, l)
I t 0
I BLOCK TERM
{p}S{I TERM}
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Activity
Do the Distributed Deadlock Activity

You might also like