A New Function-based Framework for Classification and Evaluation of Mutual Exclusion Algorithms in Distributed Systems

Leila Omrani, Zahra Rafinezhad & Mohammadreza keyvanpour
International Journal of Computer Science and Security (IJCSS), Volume (5) : Issue (2) : 2011 193
A New Function-based Framework for Classification and
Evaluation of Mutual Exclusion Algorithms in Distributed
Systems
Leila Omrani L.Omrani@Qiau.ac.ir
Department of Computer Engineering
Islamic Azad University of Qazvin
Qazvin , 34185-1416,Iran
Zahra Rafinezhad Z.Rafinezhad@Qiau.ac.ir
Mohammadreza Keyvanpour Keyvanpour@Qiau.ac.ir
Abstract
This paper presents a new function-based framework for mutual exclusion algorithms in
distributed systems. In the traditional classification mutual exclusion algorithms were divided
in to two groups: Token-based and Permission-based. Recently, some new algorithms are
proposed in order to increase fault tolerance, minimize message complexity and decrease
synchronization delay. Although the studies in this field up to now can compare and evaluate
the algorithms, this paper takes a step further and proposes a new function-based framework
as a brief introduction to the algorithms in the four groups as follows: Token-based,
Permission-based, Hybrid and K-mutual exclusion. In addition, because of being dispersal
and obscure performance criteria, introduces four parameters which can be used to compare
various distributed mutual exclusion algorithms such as message complexity, synchronization
delay, decision theory and nodes configuration. Hope the proposed framework provides a
suitable context for technical and clear evaluation of existing and future methods.
Keywords: Mutual Exclusion, Critical Section, Token
1. INTRODUCTION
The mutual exclusion problem involves the allocation of a single, non shareable resource
among n processes [1], by means that just one process can execute in its critical section at a
given time. Mutual exclusion problem was first introduced in centralized systems. In these
systems mutual exclusion ensure with preserving semaphores and monitors, and one of the
nodes function as a central coordinator which is fully responsible for having all the information
of the system and processes ask only the coordinator for permission to enter their critical
sections. But in distributed systems the decision making is distributed across the entire
system and the solution to the mutual exclusion problem is far more complicated because of
the lake of common shared memory and physical clock. So obtain a complete knowledge of
the total system is difficult.
Lots of algorithms are proposed in distributed systems. They classified into two groups
traditionally. One of them is token-based, in this group there is a unique token in the system
which ensure mutual exclusion. So the requesting node must have it to enter the critical
section. Another one is permission-based group, that requesting node has to ask all other
nodes for their permissions to enter the critical section [5]. According to approach of new
algorithms, in this paper we proposed a new function-based framework which classified these
algorithms in four groups as Token-based, Permission-based, Hybrid and K-mutual exclusion.
Also in new approach spite of synchronization delay and message complexity in light and

heavy loads, we introduced two new measures including decision theory and nodes
configuration.
This paper is organized as follows: in section 2, presents general model of distributed system
and formally describes the mutual exclusion problem. In section 3, introduces the proposed
framework and measures with classification of algorithms. Finally in last section concludes
our work.
2. MODEL AND PROBLEM DEFINITION
2.1. System Model
In general, most of mutual exclusion algorithms use a common model. In this model, a
distributed system is a set of independent and autonomous computers. These computers are
called as node or site and connected via a communication network. Each node has abstract
view of whole system that communicates with message passing [4]. The most important
purposes of distributed system are assigned as providing an appropriate and efficient
environment for sharing resource, having an acceptable speed and high reliability and
availability.
2.2. The Mutual Exclusion Problem
Mutual exclusion problem in distributed systems has received great consideration in recent 3
decades. This problem ensures that concurrent processes access common resource and
data sequentially. In addition each process that executes in its critical section for a finite time,
must do without interfering with other processes. Also, when no process is in a critical section,
any process that request entry to its critical section must be permitted to enter without delay
[3]. Eventually mutual exclusion must be without deadlock and starvation.
3. PROPOSED FRAMEWORK AND CRITERIA FOR CLASSIFICATION OF
MUTUAL EXCLUSION ALGORITHMS
By reason of the mutual exclusion importance in distributed system for keeping system
consistency and increasing concurrently, various algorithms are proposed. In order to
evaluate performance of these algorithms various criteria are defined as synchronization
delay and message complexity. Synchronization delay is the average time delay in granting
critical section. Message complexity is the number of messages exchanged by a process per
critical section entry. Also evaluates message complexity in different heavy and light load of
system state.
In addition, two new measures are proposed as decision theory and nodes configuration. In
the first one, if each node need not keep information about the concurrent state of the system,
the algorithm will be called static. On the other hand the algorithm is called dynamic. Also, in
nodes configuration if nodes are assumed to be arranged in logical configuration, the
algorithm is called structural, otherwise is called nonstructural. In next sections, the proposed
framework is presented for classifying mutual exclusion algorithms with their evaluation as
follows: 1.Token-based, 2.Permission-based, 3.Hybrid, 4.K-mutual exclusion.
3.1.Description of Token-based Approach
In this approach the right to enter a critical section is produced by a unique message, named
token. The concurrent owner of the token chooses the next token owner and sends it the
token. So granting the privilege to enter the critical section performs by the owner of the
token. In 1985, Suzuki and Kasami [5], presented an algorithm that token by means of
privilege message transmitted to requesting process based on sequence number.
Some of the algorithms use a logical structure like in Raymond [6]. In this algorithm, nodes
(each process performs in a node) arranged in rootless tree structure and every node is
related only to its neighbors and is aware of their information. In Naimi and Trehel algorithm
[7], each node sends its request only to another one that knows as a current root and waits for
its permission. In addition this algorithm uses two data structures, one of them is a queue for
keeping requests and the other is a logical rooted dynamic tree for assigning token. But this
algorithm is so sensitive to node failure and recovery. In consequence in [8] presented a
dynamic algorithm which is able to failure detection, regenerates lost token and robust against
failures.

In 2006 a new algorithm presented as queue migration [9], that nodes arranged in a
distributed and fully connected network. In this network, nodes in logical group can
communicate directly together for the purpose of entering the critical section. One node from
each group is selected to form part of the global group this node is called a link node. Link
node collects logical and global requests and sends them to the owner of token. So token
transmits among local and global groups for ensuring mutual exclusion.
3-1-1. Evaluation of Token-based Approach
The most advantage of Token-based approach is simplicity. In this approach, for example if
the logical structure of algorithm is a ring then the token transmits from a node to another
continuously. Table1 represent the comparison and evaluation of well-known algorithms
which mentioned in previous section according to proposed measures. Under light load, this
method is so expensive because token is broadcasted without using by any nodes. But under
heavy load it's so efficient. According to results that show in table1 Suzuki and Kasami's
algorithm [5] has least message complexity under light load and queue migration algorithm [9]
has least message complexity under heavy load. As you see in table1, the algorithm which
has a structural configuration and dynamic decision theory can has less synchronization
delay.
Token-based approach is so capable to lose token and deadlock occurs if it can't regenerate
token. But if algorithm assumes the existing token is lost and regenerate another one, it will
violate mutual exclusion [10]. Another problem of this approach is low scalability because all
nodes arranged in a logical structure. So, when the number of nodes increment, the average
of waiting time increase. Most of these algorithms have a structural configuration and decision
theory make dynamically.
3.2.Description of Permission-based Approach
In Permission-based approach, requesting node asks to obtain permissions from a set of
nodes in the systems. A priority or an order of events have to be established between
competing requesting nodes, so only one of them receives permission from all other nodes in
the set [5]. After receiving permission from a sufficient number of nodes, it is allowed to enter
the critical section. When a node is completed its execution in critical section must sends
release message to the other nodes. The main problem is finding a minimal number of nodes
from which a node has to obtain permission to enter its critical section. Many algorithms have
been developed to find this minimal, such as Lamport algorithm [11]. This algorithm uses a
mechanism based on logical clocks for the total ordering of requests in the system.
Description
Evaluation measures
Algorithm
name decision
theory
configuration
synchronization
delay
Message complexity
Light
load
Heavy
load
Token as a
privilege
message
dynamicNon structural_0N
Suzuki-
Kasami
Nodes in
rootless tree
staticstructural((log N)/2)TO(log N)O(log N)
Raymond
with tree
structure
Use two
structures
:queue&
logical tree
dynamicstructuralTO(log N)O(log N)
Naimi-
Trehel
With failure
detection &
recovery
dynamicstructuralTO(log N)O(log N)
Dynamic
tree
With global &
local groups
dynamicstructural_)( NO2
Queue
migration
TABLE 1: comparison and evaluation of Token-based approach
In Ricart and Agrawala algorithm [12], when a node receives a request compares its
sequence number with previous request and allows the request with smallest sequence

number to enter the critical section. After that in [13], both of above algorithms exchanged and
introduced a new algorithm which is the best known algorithm that guarantees fairness in the
same sense. It means, when there is a high priority request to do, the low priority request is
delayed.
3.2.1. The Group Mutual Exclusion Problem
In this problem, every request for a critical section is associated with a type or group. It means
at any time none of two processes which have requested critical sections belonging two
different groups in their critical sections are simultaneously. In addition it's free of starvation it
means a process wishing to enter critical section succeeds eventually.
Also, concurrent entry property is the most important issue in group mutual exclusion, it
means if all requests are for critical sections belonging to the same group, then no requesting
process should wait for entry in to its critical section until some other processes have left it
[14]. The concept of quorum is used to solve this problem. In fact quorum is a subset of
processes. Each process enters its critical section only after it has successfully locked all
nodes in its quorum. There are two properties in the concept of quorum which can ensure
mutual exclusion requirements: the first one is intersection and the second one is minimally
[14].
At first Maekawa in 1985[15], used the concept of quorum in his algorithm. In this algorithm
requests serviced based on their sequence numbers. After that another algorithm [16]
represented by using the concept of dynamic quorum. The purpose of this algorithm was high
scalability and low message complexity. This algorithm acts independent on its quorum
system. In Maekawa algorithm a node never changes its quorum or requests from the other
quorums but here a node can link to member of other quorums and respond all of its requests
by making some changes during execution. Hence the quorums change dynamically. Then,
the delay optimal algorithm was presented in [17]. This algorithm has least delay with fix
message complexity among of above algorithms.
3.2.2.Evaluation of Permission-based Approach
Fairness is a very important measure for solutions to the most contention problems. In the
concept of mutual exclusion is that requests for access to the critical section are satisfied in
the order of their timestamps. This concept is obvious in the algorithms which mentioned
above, clearly [13].
Table2 presents the comparison and evaluation of well-known algorithms which mentioned in
previous section according to proposed measures. Beside the number of messages are
exchanged for accessing the resources capture the overhead imposed on the system. In the
various algorithms are tried to get an optimal value. This value in group mutual exclusion
algorithms is limited to the number of quorum members [16].
As mentioned in section 3.1.1, most of the token-based algorithms configuration was
structural, which decreased scalability. Also detection of lost token and regenerate were the
other problems of this approach. But in permission-based approach, are tried to solve these
problems. Here configuration of nodes in most of the primary algorithms was nonstructural.
However after introducing the group mutual exclusion, is imposed the logical structure to the
system and the configuration is became structural. But as show in table2 synchronization
delay increased when the configuration of algorithms became structural. So delay optimal
algorithm solved this problem.
3.3. Description of Hybrid Approach
Providing deadlock-free distributed mutual exclusion algorithms is often difficult and it involves
passing many messages, so the hybrid algorithms are introduced. Since call such algorithm
hybrid that uses both Token-based and Permission-based approach for mutual exclusion
assurance, simultaneously. One of the algorithms of this group is proposed by Paydar et
al[10]. In which setsn nodes in 2-dimension array. This array is composed of rows and

Description
Evaluation measures
Algorithm
name decision
theory
configuration
synchronization
delay
Message complexity
Light
load
Heavy
load
Give priority
with timestamp
staticNon structuralT3(N-1)3(N-1)Lamport
Get n-1
permissions
dynamicNon structuralT2(N-1)2(N-1)
Ricart-
Agrawala
Give priority
with FIFO
dynamicNon structuralT(N-1)2(N-1)fair
Use quorumstaticstructural2TN3N5Maekawa
Generate
dynamic
quorum
dynamicstructural3TO(q)O(q)
quorum
Dynamic
Least
synchronization
delay
dynamicstructuralT)1(6 −N)1(6 −NDelay
optimal
TABLE 2: comparison and evaluation of Permission-based approach
columns. Every node i (i=1,2,3,…,n) with the other nodes in the same row and the
other nodes in the same column form the quorum Qi. Every node can enter to critical
section when it obtains the permissions of all nodes in the same quorum and gets the token
too.
Kakugawa et al [18] presented an algorithm that used the coterie concept. Coterie is a set of
quorums, each of which is a subset of the process set and any two quorums share at least
one process. In coterie concept both intersection and minimally properties are satisfied. This
algorithm uses two classes of token, main-token and sub-token. If each process of a coterie
requests to enter the critical section, the owner of the main-token by generating necessary
sub-tokens can respond their requests.
Then an algorithm based on Suzuki and Kasami's algorithm [27] is proposed in [19,20]. This
algorithm uses non-uniform groups, in which, the same groups set in one session. Also, uses
two kind of tokens to enhance concurrency of Suzuki and Kasami algorithm: primary and
secondary. The owner of the primary token can grant the secondary token to others.
3.3.1. Evaluation of Hybrid Approach
In algorithms of this approach, every node begin its function in a permission-based manner
but continues in token-based. This approach can overcome token-based problems, Because
of using both token-based and permission-based techniques.
Table3 represents the comparison and evaluation of well-known algorithms which mentioned
in previous section according to proposed measures. Most of the algorithms in this group
improve message complexity and increase the degree of concurrency. This issue avoids
unnecessary blocking [19,20], also they can increase fault tolerance by using failure recovery
and failure detection mechanisms.
According to table3 these algorithms because of using quorums have structural configuration
and their decision theory is dynamically, so these features help them to overcome the only
token-based or permission-based algorithm's problems.
3.4. Description of K-mutual Exclusion Approach
The K-mutual exclusion problem is a fundamental distributed problem that completes the
mutual exclusion issue. It guarantees the integrity of the k units of a shared resource by
restricting the number of processes that can access them simultaneously [21]. Likewise these
algorithms divided in to token-based and permission-based groups. In token-based group, k
tokens are generated to let requested processes to enter their critical sections. But in
permission-based group a node gets in to the critical section only after sending requests to

description
Evaluation measures
Algorithm
name decision
theory
configuration
synchronization
delay
Message
complexity
Light
load
Heavy
load
Quorum contains of
same column and
row nodes
Dynamicstructural_N4N4
Paydar
algorithm
With coterie and
main token & sub
token
Dynamicstructural3T05|Q|+1Kakugawa
With non-uniform
groups
DynamicstructuralT2N-12N-1
Non-uniform
groups
TABLE 3: Comparison and evaluation of Hybrid approach
the other nodes and receiving permission from nodes [22].
One of the algorithms of this group that uses queue migration is extended on [9]. In this
algorithm for ensuring mutual exclusion, the link node is the owner of parent token which has
the capacity of generating tokens. In [23] the Raymond's algorithm [6] extended.
Although in this algorithm each node has a sequence number, it has to obtain
permissions to enter the critical section. In [21] the Raymond's algorithm is extended again
and it increases fault tolerance to nodes, in which in Raymond's algorithm it was
nodes. So it ensures that even occurs failure, k processes can execute simultaneously.
3-4-1.Evaluation of K-mutual Exclusion Approach
Most of the k-mutual exclusion algorithms focus on fault tolerance. They provide using more
resource at the same time. For example, some of these algorithms can prevent the system
security until occurs failures. It means when a node fails, It is impossible to any others
can collect permissions and enter to the critical section. But after failures, the
most number of processes that can execute simultaneously reach to one. As a result the
algorithm's performance decreases. Also, if the number of the received permissions
decreased by occurring failures dynamically, the value of fault tolerance will reach to
nodes. So it ensures that even occur failures, k processes can execute in their critical
sections simultaneously [23,24].
Table 4 represents the comparison and evaluation of well-known algorithms which mentioned
in previous section according to proposed measures. As you see decision theory in these
algorithms are dynamically.
description
Evaluation measures
Algorithm
name decision
theory
configuration
synchronization
delay
Message
complexity
Light
load
Heavy
load
Use parent token
& generate k-1
tokens
dynamicstructural4T0)( NO
k-queue
migration
K resources & n-
k replies
dynamicNon structural_2N-k-12N-1
Raymond with
multi entries
Extended of
Raymond
algorithm
dynamicNon structural_2N-k-12N-1
Obtain ni-k
replies from ni
correct nodes
TABLE 4:comparison and evaluation of K-mutual approach
4. CONCLUSION
According to importance of mutual exclusion in achieving goals of distributed systems,
various approaches are proposed. In this research try to make a brief analyzing on most

common distributed mutual exclusion algorithms. Also, present a new classification based on
their functions in four groups: Token-based, Permission-based, Hybrid and K-mutual
exclusion. This framework helps novice researchers to set each new algorithm in specific
category. To achieving this purpose focus on four measures such as message complexity,
synchronization delay, decision theory and nodes configuration for comparison and evaluation
of them. Hope the proposed framework makes a convenient way for future researches.
5.REFERENCES
[1] N.A.Lynch . "Distributed Algorithms" , Morgan Kaufmann Publishers, pp.255-327,(1996)
[2] P.C.Saxena, and J.Rai. "A Survey Of Permission-based Distributed Mutual Exclusion
Algorithms" . Computer Standards & Interfaces, 25: 159-181, 2003
[3] M.G.Velaquez. "A Survey Of Distributed Mutual Exclusion Algorithms". Technical Report
CS. Colarido state university, September 1993
[4] W.Stallings. "Operating Systems Internals and Design Principls", Prentice Hall, pp.205-
261 (2009)
[5] I.Suzuki,and T.Kasami."A Distributed Mutual Exclusion Algorithm", ACM Transactions
On Computer Systems, Vol.3(No.4): 344-349, November 1985
[6] K.Paymond. "A TreeBased Algorithm For Distributed MutualExclusion",ACM Transactions
On Computer System, Vol.7(No.1): 61-77, February 1989
[7] M.Naimi, M.Trehel, and A.Arnold. "A Log(n) Distributed Mutual Exclusion Algorithm
Based On The Path Reversal", Journal Of Parallel And Distributed Computing,34(1): 1-
13 April 1996
[8] J.Sopena , L.Arantes, M.Bertier, and Pierre Sens. "A Fault-tolerant Token-based Mutual
Exclusion Algorithm Using A Dynamic Tree". Euro Par.LNCS 3648, 2005
[9] P.chaudhuri, and Tomas Edward. "An Distributed Mutual Exclusion Algorithm
Using Queue Migration". Journal Of Universal Computer Science, Vol.21(No.2):140-
159, 2006
[10] S.Paydar, M.Naghibzadeh , and A.Yavari. "A Hybrid Distributed Mutual Exclusion
Algorithm", IEEE International Conference On Emerging Technologies In Pakistan,
November 2006
[11] L.Lamport. "Times,Clocks,And The Ordering Of Events in a Distributed System".
Communications OF The ACM, Vol.21(No.7): 558-565, July 1978
[12]G.Ricart, and Ashok.K.Agrawala. "An Optimal Algorithm For Mutual Exclusion in
Computer Networks". Communication of The ACM, Vol.24(No.1):9-17, January 1981
[13] S.Lodha, and A.Kshemkalyani. "A Fair Distributed Mutual Exclusion algorithm".IEEE
Transactions On Parallel And Distributed Systems. Vol.11(No.6), June 2000
[14] R.Atreya , and N.Mittal. "A Dynamic Group Mutual Exclusion Using Surrogate-Quorums".
Proc,IEEE Int'l Conf. Distributed Computing System, June 2005
[15] M.Maekawa. "A Algorithm For Mutual Exclusion In Decentralized Systems". ACM
Transactions On Computer Systems , Vol.3 (No.2):145-159, May 1985
[16] R.Atreya , and N.Mittal. "a Quorum-based Group Mutual Exclusion Algorithm For A
Distrinbuted System With Dynamic Group Set". IEEE Transactions On Parallel And
Distributed Systems, Vol.18(No.10), October 2007

[17] G.Cao , and M.Singhal. "A Delay-optimal Quorum-based Mutual Exclusion Algorithm
For Distributed Systems". IEEE Transactions On Parallel And Distributed Systems,
Vol.12(No.12), December 2001
[18] H.Kakugawa , S.Kamei, and T.Masuzawa. "A Token-based Distributed Group Mutuaal
Exclusion Algorithm With Quorums", IEEE Transactions On Parallel And Distributed
Systems, Vol.19(No.9), 2008
[19] N.Mittal,and P.Mohan."An Efficient Distributed Group Mutual Exclusion Algorithm For
Non-uniform group access", proceedings Of The IASTED International Conference On
Parallel And Distributed Computing And Systems , Phoenix,Arizona,USA, 2005
[20] N.Mittal,and P.Mohan. "A Priority-based Distributed Group Mutual Exclusion Algorithm
When Group Access is Non-uniform". Journal Of Parallel And Distributed Computing,
No.67: 797-815, March 2007
[21] M.Bouillaguet,L.Arantes,and P.Sens. "Fault Tolerant K-mutual Exclusion Algorithm Using
Failure Detector" . International Symposium On Parallel And Distributed Computing,
2007
[22] P.Chaudhuri, and T.Edward. "An Algorithm for K-mutual Exclusion In Decentralized
Systems". Computer Communications 31: 3233-3235, 2008
[22] K.Raymond. "A Distributed Algorithm For Multiple Entries To A Critical Section",
Information Processing Letters, North-Holland, No.30, February 1989

A New Function-based Framework for Classification and Evaluation of Mutual Exclusion Algorithms in Distributed Systems

More Related Content

What's hot (18)

Viewers also liked (20)

Similar to A New Function-based Framework for Classification and Evaluation of Mutual Exclusion Algorithms in Distributed Systems (20)

Recently uploaded (20)

A New Function-based Framework for Classification and Evaluation of Mutual Exclusion Algorithms in Distributed Systems