IRJET- Effective Technique for Optimizing Timestamp Ordering in Read-Write/Write-Write Operations

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 06 Issue: 09 | Sep 2019 www.irjet.net p-ISSN: 2395-0072
© 2019, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 1917
Effective Technique for Optimizing Timestamp Ordering in Read-
Write/Write-Write Operations
Obi, Uchenna M1, Nwokorie, Euphemia C2, Enwerem, Udochukwu C3, Iwuchukwu, Vitalis C4
1Lecturer, Department of Computer Science, Federal University of Technology, Owerri, Nigeria
2Senior Lecturer, Department of Computer Science, Federal University of Technology, Owerri, Nigeria
---------------------------------------------------------------------***----------------------------------------------------------------------
Abstract - In recent times, the use of big data has been the
trending technology, most enterprises tend to adopt large
databases, but the inherent problem is how to ensure
serializability in concurrent transactions that may want to
access the data in the database so as to maintain its data
integrity and not to compromise it. The aim of this research is
to develop an efficient Timestamp Ordering algorithm and
model in conflictingoperationsinread-write/write-writedata
synchronization. In read-write synchronization, one of the
operations to perform is read while the other is write
operation. While in the write-write synchronization, both
operations are trying to access the same data itemandboth of
them are write operations. If multiple transactions modify the
same data item, the integrity of the database might be
compromised if there is no proper control mechanism.
Timestamp and C4.5 machine learning algorithm was used to
develop an algorithm that ensures serialization and
consistency of database. We adopted object Oriented Analysis
and Design Methodology in the analysis of effective technique
for optimization timestamp ordering scheduler in RW/WW
synchronization. The proposed hybrid algorithm was able to
carry out the RW and WW transactions in minimal time. It
optimizes the cost of reading or writing in large datasets and
enable speed access in the database system with huge volume
of data; thus, eliminating the problem of coordinating
concurrent access to a database system.
Key Words: Timestamp, Concurrency, Transactions,
Serializable, Conflict, Read, Write, C4.5.
1. INTRODUCTION
A transaction is a set of instructions or user program that
comprises of set of read and write operations in a database.
In other words a transaction is an atomic unit of processing,
that is completed entirely or not done at all [1]. In a
multiuser system with large databases and large number of
users executing database transactionsatthesametime,such
as in airline reservations, railway reservations, electronic
fund transfer, stock markets, online retail purchasing,online
job applications and manyotherapplications,conflictsmight
arise among transactions and that may create problems of
database consistency. There are two kinds of
synchronizations that will require concurrency control
measures, these are: read-write (RW) synchronization and
write-write (WW) synchronization.
Concurrency control techniques are employedformanaging
concurrent access of transactions on a particular data item
by ensuring serializable executions or to avoid interference
among transactions andtherebyhelps inavoidingerrorsand
maintain consistency of the database. Various concurrency
control techniques have been developed by different
researchers and these techniques are distinct and unique in
their own methods and representations [2].
Over the last few decades, processing of transactions and
concurrency control issues have played a vital role in
modern databases and have been a crucial research area.
After studying relevant proposed algorithms, Arun and Ajay
(2010) found out that there are compositions of only a few
algorithms used by DBMS, two basictechniquesusedtoform
various concurrency control algorithmsare:2-phaselocking
(2PL) and timestamp ordering (TO) [3].
2-Phase Locking(2PL) algorithm employs the mechanism of
preventing multiple transactions from accessing data items
simultaneously by locking data items, 2PL protocols are
mostly used in commercial DBMSs but transactions
encounter problems like long waiting time (blocking) and
deadlock which requires deadlock detection,preventionand
avoidance mechanisms. While timestamps ordering (TO)
does not use locks, therefore deadlock cannot occur.
A timestamp is a distinct identifier for every transaction
created by the database management system (DBMS).Atthe
start of every transaction, Timestamps values are generated
base on system clock or a logical counter.
Elmasri and Navathe[4] states that timestamp ordering
protocol ensuresthattransactionswitholdertimestampsget
higher priority in the event of conflicting operations on a
particular data item and for every data item, two timestamp
are given: W-timestamp which is the last timestampofwrite
operation performed successfully on a data item and R-
timestamp which is the last timestamp of read operation
performed successfully on a data item [5]. Read/write
operations can only take place if the last update on that data
item was done by a transaction with an older timestamp
value. Else, transaction requestingread/writeisaborted and

given a new timestamp. The rule is that whenever we have a
data item (X) and some transactions T1, T2 and T3 try to
issue a read operation on data item R(X)ora writeoperation
on data item W(X), the timestamp of T1 will be compared
with the R-timestamp of data item R-TS(X) and the W-
timestamp of data item W-TS(X) to make sure that the
transaction timestamp order is not violated. If the order is
violated, then transaction T1 will beaborted,rolledback and
restarts as a fresh transaction with a new generated
timestamp value, the next transaction T2 which might have
used the value written by transaction T1 will also be rolled
back. On the same note, T3 which might have made use of a
value written by transaction T2 will also be rolled back, and
so on. This resulting effect is called cascading rollback and it
is one of the major problems of basic timestamp [2].
Timestamp ordering algorithm can give efficient results for
concurrent control problems when provided with relevant
information of transaction of the database [6]. Therefor a
hybrid algorithm will be developed using intelligent query
optimizer and timestamp ordering techniques to control
concurrency problems in a database system. The intelligent
query optimizer will be developed with C4.5 Machine
Learning Algorithm. The intelligent query optimizer will
provide timestamp ordering algorithm with previous
information about the transaction in the database system
needed to efficiently control concurrent problems.
1.1 Overview of Timestamps and their origin
Generally, timestamps are digital representation of
specific process recorded in time and are created at the start
of executing codes running on a processor, the executed
codes obtains the value based on the local clock assigned to
the processor andare stored ontheharddiskorcanbeadded
in the packets sent through the network. Time identification
of when the enclosing code was carried out is normally the
function of the timestamp [7]. Timestamp uniquelyidentifies
a specific process on computer systems. Large numbers of
timestamps are stored in the computer system memory and
typicallystore them in manner thattheprocesscanbeclearly
identified.
Sources of timestamps include the following:
(i) File systems: Each file inacomputersystemisassociated
with a timestamp base on the user actions on the file
either when the file was created, last read,lastwrittenor
otherwise modified.
(ii) Email: SMTP severs add timestampstoemailsthatwillbe
transmitted, therefore timestamps are associated with
email messages and some messaging protocols, such as
social networkand SMS/MMS used in GSMalsogenerate
timestamps to each messagesentacrossthenetwork[8].
(iii) Logs: The logging mechanism keeps the record of every
activity taking place in the system. System logging
facilities usually log events from system processes in
system logs. Every event has its own timestamp value.
(iv) Database: Timestampare usedtocoordinateconcurrent
transactions accessing data item in database
management system (DBMS) so as to avoid conflicts
among transactions, timestamps are given to every
transaction at the start of execution by the DBMS to
identify each transaction. Typically, transactions obtain
timestamp values in the order in which they are
submitted to the system and priority of getting access to
the data item is based on the timestamporder.Oneofthe
possible way timestamps are created is by the use of the
current date/timevalue of the localclock assignedtothe
system processor and ensure that no two transactions
are assigned with the same timestamp value. Another
way to implement timestamps is the use of a logical
counter that increases each time a value is assigned to a
transaction. The maximum value of the logical counter
may be finite. So periodically, whenever transactions
stop executing for short period of time, the system reset
the counter to zero [4].
1.2 Distributed Database Management System
(DDBMS)
Large databases that are accessed from remote machines
or remote areas are to be place in sections and stored on
different machines or sites for fast andreliabledata retrieval
and access. Distributed Database Management System
comprises of different number of sites interconnected
together by a communication network [3]. Each of the sites
is a centralized database that consists of:
1. Transactions (T)
2. Transaction Manager (TM)
3. Data Manager (DM)
4. Data.
Users communicate with the DBMS by the execution of
transactions that may beembeddedqueriesinanapplication
programs written in a high level programming language,
while TM supervises and manages transactions between
users and the DBMS while DM manages the actual data
(database).
InDistributedDatabaseManagementSystem,transactions
interact with TMs and TMs interact with DMs while DMs
interact and manage the Data. Meanwhile the interaction
between TMs with each other andDMswitheachotheris not
possible. Fig - 1.1 shows the components of DDBMS.

Fig - 1.1: DDBMS Architecture [9].
1.3 Conflicting Operations
Two transactional operations T1(A) and T2(A) are in
conflict if and only if they access the same data item (A) and
at least one of the operations is a write operation, i.e.
1. T1 = read (A) and T2 = write (A).
2. T1 = write(A) and T2 = read(A)
3. T1 = write(A) and T2 = write(A)
Here, read(A) is equivalent to the time of operation by
transaction T1 to read data item (A), write(A) means thetime
of write operation on data item (A) by transaction T1 etc.
The conflict between two operations shows that their order
of execution is important, because if there isconflictbetween
the operations of more than one transaction, their execution
may result to a database inconsistency state. Note, the
ordering of read operations does not matter,becausetheydo
not conflict with each other.
Fig -1.2: Conflict relation of read and write operations
(Design of an object-oriented framework for atomic
transactions).
1.4 Schedule
A schedule is a list of operations (read, write, abort, or
commit) ordered by time, performed by a set of transactions.
Fig-1.2 shows a Schedule involving two Transactions T1 and
T2. T1 contains two operations, read data item (A) and write
on data item (A), while T2 also contains two operations,read
data item (A) and write on data item (A). The Fig-1.3 shows
two transactions executed in serial order, while Fig- 1.4
shows all the possible order of executing T1 and T2. In
schedule1, the operations of T1 are executed entirely before
T2 start execution. In schedule2, the operations of T2 are
executed entirely before T1, which showsthatschedule1and
schedule2 are serial execution. While the operations in
schedule3 and schedule4 are interleaving, this means that
their execution are serializable.
So far, the most accepted criterion for correctness in
concurrent transactions in database system is serializabilty
[10]. The concept of serializability is based on the fact that
any sequential execution of transactions will leave the
database in a consistence state. Serializabilitymaybedefined
as the execution of certain schedules on the database which
will produce the same output and effect to at least one serial
execution of the same transactions [18].
Fig-1.3: A schedule involving two Transactions T1 and T2
[11]
Fig - 1.4: The sequence of execution T1 and T2 [12]

1.5 Serial schedule
A schedule is serial if all the operations of one
transaction are executed and then all the operations of
another transaction, and so on. That is, for a schedule to be
serial the sequence of transactions T1, T2, T3, ..., Tn, the
transaction T1 must be executedentirely(committed)before
the next transaction T2 startsexecution. Serial schedules are
naturally consistent and result in correct execution and the
operations of transactions do not interleave among each
other [13]. Let’s consider the example where two accounts X
and Y have the balances of 500 and 1000 naira, respectively,
and two active transactions T1 and T2, in which funds are
transferred from one account to another. Fig-1.5 shows a
schedule involving T1 and T2.
Fig-1.5: A schedule involving two Transactions T1 and T2
in a serializable form [12].
2. Review of Related Work
Arun et al[3] proposed a distributed architecture in
distributed database system for synchronization in
transaction. In their analysis, problems of concurrency
control were decomposed into read-write and write-write
synchronization. They stated that the performance of their
proposed system will be evaluated in their future work [3].
Concurrency control method based on commitment
ordering in mobile databases was proposed by Karami et al
[14]. They introduced OPCT concurrent control algorithm
based on optimistic concurrency control method. They
provide their proposed OPCT concurrent control algorithm
usingserialization graph.From the results, it wasshownthat
there was a decrease in abortion rate and waiting time when
using OPCT algorithm compared to 2PL and Optimistic
algorithm, but the problem of their proposed system is
overhead of timestamps and their calculation.
Dagar et al. [2] in his analysis of effectiveness of
concurrent control techniques in databases, they analyzed
different techniques of concurrency control such as locking,
timestamp and optimistic-based mechanism. Theyproposed
optimistic-based mechanism isdeadlockfreetechniques.Itis
prior to other techniques as it assumes that not too many
transactions will conflicts with each other, but its waiting
time is less than locking and more than timestamp ordering.
Kamal [12] analyzed transaction concurrent control for
resource constrained application. They explore the previous
techniques relating to timely transactional systems for
remote clients and centralized database. They used the first
techniques developed to decrease disk access time via local
caching of state to tackle the problems of prevalent in real-
timedatabases. They results bringsgaveefficientthroughput
to improve batterylifeformobile deviceand minimizedtime
complexity of the system. They suggested that it would be
worth investigating the proposed approach with the new
transactional phase order in non-blocking software
transaction memory, order to achieve further advance in the
field.
Saeid et al [15] in his proposal analyzed the modeling of
timestamp ordering method with the use of coloured Petri
Net. The analysis using state space shows that timestamp
ordering methodsillustratedinsometextbooksmaybefaced
with starvation. They also analyzed model explosion of state
space and they observed that timestamp ordering method
falls in infinite loop and therefore has starvation. They
proposed concept of precedence graph produces an efficient
throughput of the system. They stated that it is an open
problem to start bringing up new variations of Timestamp
Ordering method that inherently eliminates starvation.
In Shefali et al [16] work, issues of performance in
concurrent transaction execution in DDBMS was revisited.
They stated that for a distributed data, it is very difficult to
find an absolute solution with respect to the following
challenges: Network cost, usage of memory, processingtime,
response time,accesstime,resources,etc.Theyreviewshows
that to obtain optimum solution these challenges should be
minimized which would be done with the use of better
algorithm for different principles of DDBMS.
Jaypalsinh et al. [6] made a proposal of a study and
comparative analysis of basic Optimistic and Pessimistic
management system. They highlighted on the pros and cons
of pessimistic and optimistic approaches for concurrency
controls. They observed that Pessimistic locking based
approach is suitable for update-intensive for read operation
and the optimistic method is a backward oriented
concurrency control method which works on three different
phases: Read, WriteandValidation.Theydidnotcomparethe
performance of all basic methods of concurrent control to
provide optimum performance of their proposed optimistic
concurrency control methods.
3. Analysis of the Existing System
The problem of coordinating concurrentaccessina database
system has been area of interest by many researchers and

several concurrent controls have been introduced. The
intention of concurrency control is to ensure that the
integrity of the database is preserved while allowing the
execution of transactions in the system. Lomet [17]
proposed a consistent time stamping for transactions in a
distributed system. He stated that achieving timestamp
ordering in a distributed system is potentially troublesome.
In Dagar et al. [2] analysis of effectiveness of concurrent
control techniques in Databases, they analyzed different
techniques of concurrency control such as locking,
timestamp and optimistic-based mechanism. The result
showed that the latter was more efficient than lock and
timestamp ordering but its waiting time is less than locking
and more than timestamp ordering.
3.1 Limitations of the Existing System
The disadvantages of the existing system are as follows:
1. The traditional query engine ofthedatabasesystem
with timestamp mechanism is time wasting,ittakes
a lot of time to complete transaction processes.
2. These mechanisms lack intelligence to minimize
time of processing transaction in a distributed
database system.
3. The problem with timestamp ordering techniques
of Transaction Processing is the time complexity of
its execution, most times transaction keep on
waiting for the assigned timestamp of the other
jobs.
4. The whole database has to be re-evaluated when
data resides in main-memory, which makes the
system to be lazy.
The architecture of the existing system is shown in Fig- 3.1.
Fig-3.1: Architecture of the Existing System [2].
3.2 Analysis of the Proposed System
This work examines effective technique for optimizing
timestamp ordering scheduler in Read-Write/Write-Write
synchronization. We proposed a hybrid intelligent decision
rule algorithm for Read-Write/Write-Write in a distributed
database system using C4.5 and Timestamp ordering
scheduler algorithm. A distributed relational database is a
database that comprises of a number of relations and
multiple physical locations or sites. Relations may be
fragmented or replicated at different sites in the system.
With the idea that queryoptimizationina distributedsystem
should be systematic just like in a centralized system, in the
sense that the optimizers should be able to use cost models
to make smart and cost effective decisions such as, to
determine where to fetch data from, where to execute
evaluation operators and in what order. The intelligent
system will ensure that data are stored at or close to the
location where they are needed most. This will reduce the
cost of reading or writing in a large datasets. Which is the
limitation of Dagar et al. analysis and it will enable speed
access in the central database distributed system with huge
volume of data and complicated data structure to provide
necessary information of transaction for efficient control of
concurrent problems. The proposed hybrid system will
eliminate the problem of coordinatingconcurrentaccesstoa
database system [2].
3.3 Advantages of the proposed system
1. It will produce efficient serialization order among
the transactions which will be determined by the
intelligent query algorithm that is sorting the
corresponding timestamp vectors.
2. It will enforce efficient concurrency control baseon
optimal solution of intelligent query of C4.5
algorithm that provides precise dependency
information obtained from the operations of the
transaction.
3. The hybrid system consisting of C4.5 and
Timestamp algorithmwill enhancetheconventional
timestamp ordering eliminates the premature
determination of the serialization order.
3.4 Architecture of the Proposed System
The architecture of the proposed system comprises of the
transactions, transaction manager, timestamp operations,
C4.5 algorithm, the data manager and the database. The
diagram is shown below

Fig -3.2: Architecture of the proposed System
From the mechanism of the Architecture, it shows that data
manager ensures that starting process is given copies of all
the files which it has to access and files are only to be read
and not to be copied, but only set a pointer to the parents’
workspace. In the log, the read/write files are modified in
place, but before changing a block, recordisthen written toa
log. This log contains details of transaction that is making a
change, the file to be changed, then the old and new values.
Sometimes in the cases of an abort, the write-head logs
makes a simple rollback, then read the log from the end and
undo all recorded changes. The transaction manager is the
coordinator that allocates transaction IDs (TIDs), Assigns
TIDs with operations, Coordination of commitments, aborts
and recover; it begin transaction and end transaction. The
scheduler ensures concurrency control and timestamp
operation, it provides efficient serialized operations.
3.5 Algorithm for the proposed system.
A: Timestamp Algorithm
Case1: For read(X) of transaction T1
(a) If Ts(T1) < W-Ts(X), then T1 needs to read a value
of X that was overwritten bynewertransactionwith
larger timestamp value, hence, the readoperationis
rejected, and T1 is rolled back and start as a new
transaction.
(b) If Ts(T1) ≥ W-Ts(X), this implies that the read
operation is executed, and R-Ts(X) is set to the
maximum of R-Ts(X) and Ts(T1).
Case2: For write(X) of transaction T1
(a) If Ts(T1) < R-Ts(X), this implies that the value of X
that T1 is generating was needed initially, and the
system assumed that the value might not be
generated again. Then the write operation will be
rejected and T1 is rolled back.
(b) If Ts(T1) < W-Ts(X), then T1 is attempting to write
an obsolete value of X. Hence, the write operation is
rejected and T1 is rolled back.
If Ts(T1) ≥ W-Ts(X) and Ts(T1) ≥ R-Ts(X), then the write
operation can be performed and sets W-Ts(X) to Ts (T1).
B: Optimization Algorithm
1) Let ‘T’ be the set of training instances
2) Choose an attribute that will properlydifferentiates
the instances contained in T.
3) With the chosen attribute, create a tree node whose
value is the same with the attribute.
4) Form child links from this node where each
individual link represents a distinct value for the
chosen attribute.
5) Create subclasses with further subdivision of the
Instances with the use of a child link value.
6) Do the instances in subclass satisfy predefine
criteria? If yes go to step 7 else go to 2.
7) Following the decision path, specify the
classification for new instances.
Stop.
3.6 The Activity Diagram of the Proposed System
Fig - 3.3: Activity Diagram of the System

The activity diagram shows the connectivity of the various
components of the proposed system. It consists of
transactions to be executed, transaction manger, scheduler,
data manager, C4.5 algorithm, database and log. As
committed version, requests are been received by a
coordinator, it will always be able to execute it because
consistency check has been carried out for all operations.
The mechanism of the proposed hybrid system ensures that
the committed versions of an object must be created in
timestamp order. The proposed hybrid system is strict such
that each read operation is delayed until previous
transactions that had written the object have committed or
aborted.
4. Result Analysis
Table 4.1: Analysis of Job Completion Time of
Transactions
Table 4.1 shows the analysis of job completion time of
transactions. It has two columns. Column one contains
transaction processed in the database. Column two contains
Arrival time. Column three contains Operations performed
and column four and five contains the job completion Time
per milliseconds. The results are represented graphically in
Fig - 4.1.
Chart 4.1: Analysis of Job Completion Time of
Transactions
4.1 Discussion of Results
Transactions were simulated simultaneously to analyze the
performance of the proposed system, where Write/Write
operations were performed. Six (6) transactions were
simulated which we calculated the transaction completion
time. From the result of Read/Write operations of the
transactions, the Read and Write transactions arrived at the
same time. The system enables the Read transaction to read
the values of database and makecopytoa privateworkspace
place where it does the modificationsandmakeslocal copies
of its modifications in the private workplace only and not to
the database. This is done in the Read phase of the proposed
system. After the conflict checking is done in the read phase,
new Timestamp Ts(Tv) is given to the active transaction, it
enters into the write phase where the modified values in the
private workspace updates the database, once a transaction
is in the write phase, it is considered to be complete. In the
write phase, transaction makes all its updates permanent in
the database. This is applicable to the rest ofthesimulations.
The experimental results show significant performance of
the proposed system with respecttosystemthroughput, and
response time as the transaction operation of the Read-
Write-Validate approach is deployed inthedatabasesystem.
5. Conclusion
This work examines effective technique for optimization
timestamp ordering of transactions in RW/WW
synchronization. We proposed a hybrid intelligent decision
rule algorithm for RW/WW in a distributeddatabasesystem
using C4.5 and Timestamp ordering algorithm. The results
enabled speed access in the central database distributed

system with huge volume of data and provide necessary
information of transaction for timestamp efficient control of
concurrent problems in complicated data structure. The
proposed hybrid system eliminates the problem of
coordinating concurrent access in a database system.
5.1 Recommendation
An improvement on the optimistic concurrency control
algorithm (OCCA) will be recommended for similarprojects.
The basic aim of OCCA is to let transactions execute freely
with the presumption that most of its transactions will not
conflict, but in areas of high contention of data this
assumption might fail. An improvement in the OOCA
technique will be a good suggestion for future research.
REFERENCES
[1] A. Habes, and R. Hasan, “Multiversion Thomas' Write
Rule Timestamp-based Concurrency Control,”
International Journal of Computer Science Issues, Vol.
12, ISSN (Print): 1694-0814 | ISSN (Online):1694-0784,
2015.
[2] R. Dagar and B. Rachna, “Analysis of Effectiveness of
Concurrency Control Techniques in Databases,”
International Journal of Engineering Research &
Technology (IJERT), ISSN: 2278-0181. Vol. 1(5), 2012.
[3] K. Arun and A. Ajay, “A Distributed Architecture for
Transactions Synchronization in Distributed Database
Systems,” International Journal on Computer Science
and Engineering Vol. 02(6); 84-91, 2010.
[4] R. Elmasri and B. Navathe, “Fundamentals of Database
Systems,” sixth edition. Includes bibliographical
references and index. ISBN 13: 978-0-136-08620-8,
2011.
[5] K. Sonal, and R. Morena, “Analysis and Comparison of
Concurrency Control Techniques,”International Journal
of Advanced Research in Computer and Communication
Engineering Vol. 4(3); ISSN (Print): 2319-5940, 2015.
[6] A. Jaypalsinh, and M. Prashant , “ StudyandComparative
Analysis of Basic Pessimistic and Optimistic
Concurrency Control Methods for Database
Management System,”International Journal ofAdvanced
Research in Computer and Communication Engineering
Vol. 5(1), 2016.
[7] Y. Svein, “Methods for Enhancement of Timestamp
Evidence in Digital Investigations,” Norwegian
University of Science and Technology, Faculty of
Information Technology, Mathematics and Electrical
Engineering Department of Telematics, 2008.
[8] D. Brian, “A hypothesis-based approach to digital
forensic investigations,” CERIAS Tech Report, Purdue
University, 2006.
[9] C. Rinki, M Suman, R. Rathy and B. Preeti, “Recoverable
Timestamping Approach for Concurrency Control in
Distributed Database,” International Journal on
Computer Science and Engineering (IJCSE), ISSN: 0975-
3397, Vol. 3(7); 2707-2711, 2011.
[10] C. Papadimitriou, “The serializability of concurrent
database updates,” J, ACM Vol. 26(4): 131-153, 1979.
[11] R. Raghu and G. Johannes, “Database Management
Systems,” Third Edition, Includes index. ISBN 0-07-
246563-8, 2003.
[12] S. Kamal, “Transactional Concurrency Control for
Resource Constrained Applications,” School of
Computing Science, Newcastle University, 2014.
[13] U. Aydonat, “Relaxing Concurrency Control in
Transactional Memory,” Department of Electrical and
Computer EngineeringUniversityofToronto.Vol.6(45);
ISSN (Print). 2010, Pp.074-145.
[14] A. Karami, and B. Dastjerdi, “A Concurrent Control
Method Based on Commitment Ordering in Mobile
Databases,” International Journal of Database
Management System (IJDMS) vol.3, No. 4. 2011, pp. 39-
53.
[15] P. Saeid and R. Mahd, “Modeling Timestamp Ordering
method using Coloured petri net,” Indian Journal of
Science and Technology, Vol. 8(35); ISSN (Print): 0974-
6846, ISSN (Online): 0974-5645, 2015.
[16] N. Shefali and K. Samarat, “RevisitedPerformanceIssues
in Concurrent Transactions Execution in Distributed
Database Management System,” International Journal of
Current Engineering and Scientific Research(IJCESR),
Vol. 2, ISSN (PRINT): 2393-8374, 2015.
[17] D. Lomet, “The Consistent Timestamping for
Transactions in DistributedSystems,”Digital Equipment
Corporation Research Lab, 1990.
[18] P. Bernstein, V. Hadzilacos and N. Goodman,
“Concurrency Control and Recovery in Database
Systems,” Addison Wesley, Reading, MA, 1987.

IRJET- Effective Technique for Optimizing Timestamp Ordering in Read-Write/Write-Write Operations

Recommended

Recommended

More Related Content

Similar to IRJET- Effective Technique for Optimizing Timestamp Ordering in Read-Write/Write-Write Operations

Similar to IRJET- Effective Technique for Optimizing Timestamp Ordering in Read-Write/Write-Write Operations (20)

More from IRJET Journal

More from IRJET Journal (20)

Recently uploaded

Recently uploaded (20)

IRJET- Effective Technique for Optimizing Timestamp Ordering in Read-Write/Write-Write Operations