Cloud Computing is one of the most important areas in the current IT industries. It provides different type of services (SaaS, PaaS, and IaaS) depending on the customers’ requirements. Instead of owning, installing and maintaining software in IT industries can use Cloud computing. In this paper, needs of ACID properties for cloud databases, advantages and limitations of cloud databases are discussed. It also portrait the major issues and challenges of Database Architectures and Data Transaction Management in Cloud Environment. There are different procedures used to handle the data, but each one of them has its own limitations in cloud environment. Finally this paper provides an exhaustive study on Transaction Processing System, Cloud DataBase as a Service (DBaaS), Cloud RDBMS and Cloud data storages in order to design a novel architecture for cloud databases with the support of traditional ACID properties.
Comments
Content
International Research Journal of Engineering and Technology (IRJET)
Volume: 02 Issue: 04 | July-2015
www.irjet.net
e-ISSN: 2395 -0056
p-ISSN: 2395-0072
Issues and Challenges of Data Transaction Management in
Cloud Environment
J.Antony John Prabu1, Dr.S Britto Ramesh Kumar2
1Research
Scholar and Assistant Professor, Department of Computer Science
St. Joseph’s College, Trichy, Tamilnadu, India- 620 002
2Assistant Professor, Department of Computer Science
St. Joseph’s College, Trichy, Tamilnadu, India- 620 002
Abstract - Cloud Computing is one of the most
important areas in the current IT industries. It provides
different type of services (SaaS, PaaS, and IaaS)
depending on the customers’ requirements. Instead of
owning, installing and maintaining software in IT
industries can use Cloud computing. In this paper, needs
of ACID properties for cloud databases, advantages and
limitations of cloud databases are discussed. It also
portrait the major issues and challenges of Database
Architectures and Data Transaction Management in
Cloud Environment. There are different procedures used
to handle the data, but each one of them has its own
limitations in cloud environment. Finally this paper
provides an exhaustive study on Transaction Processing
System, Cloud DataBase as a Service (DBaaS), Cloud
RDBMS and Cloud data storages in order to design a
novel architecture for cloud databases with the support
of traditional ACID properties.
cloud environment and provide features such as data
definition, data storage and data retrieval. The leading cloud
computing providers like Amazon, Google, IBM, Oracle and
Microsoft provide database –as- a-service with their cloud
DaaS solutions[2]. In Cloud resources are typically elastic,
with infinite amount of compute power and storage
available on demand and pay-only-for-what-you-use. In this
paper the overall issues and challenges in cloud data
transaction management has been discussed.
2. NEED OF ACID PROPERTIES FOR CLOUD TPS
During the data transaction management data are stored at
third-party and replicated from large geographic distance
[4]. It is also hard to maintain the ACID (Atomicity,
Consistency, Isolation and Durability) guarantees. Cloud
platform providers provide databases for developers to
maintain the data and each one of them has their own
advantages and limitations. For example Amazon SimpleDB
and Google’s Big table provide good databases for vendors
with different features [3].
The important concept of transaction is a sequence of
database events (reads/writes).
DBMS entrust atomicity (all-or-nothing property)
even if system crash in the central point of a
transaction.
Each transaction is executed entirely and it must
take the database between consistent states or
must not run at all.
DBMS entrust that concurrent transactions emerge
to run in isolation.
DBMS entrust durability of committed transactions
even if system smashes.
When a transaction completes successfully in Transaction
Processing System (TPS) database changes are said to be
committed and it should satisfy the ACID properties [8].
Page 123
International Research Journal of Engineering and Technology (IRJET)
Volume: 02 Issue: 04 | July-2015
www.irjet.net
3. ADVANTAGES AND LIMITATIONS IN CLOUD
DATABASES
e-ISSN: 2395 -0056
p-ISSN: 2395-0072
Increase scalability and
reliability
A suitable cloud database is needed to execute a successful
transaction in cloud TPS. Cloud uses different database
architectures for TPS and maintains the presentation,
application, database and data storage layers. Such a
database-as-a-service (DBaaS) is attractive for two reasons.
First, due to economies of scale, the hardware and energy
costs incurred by users are likely to be much lower when
they are paying for a share of a service rather than running
everything themselves. Second, the costs incurred in a welldesigned DBaaS will be proportional to actual usage (“payper-use”)—this applies to both software licensing and
administrative costs. The latter are often a significant
expense because of the specialized expertise required to
extract good performance from commodity DBMSs. By
centralizing and automating many database management
tasks, a DBaaS can substantially reduce operational costs
and perform well. The advantages and limitations of various
database architectures are listed below [7].
Not possible to
Provide full scalability
achieve consistency
and elasticity.
and availability.
Distributed
Cheap hardware can be Weak to achieve
used in all tiers.
the atomicity.
TABLE -1: ADVANTAGES AND LIMITATIONS IN CLOUD DB
TABLE -2: ISSUES RELATED TO CLOUD TPS
Cloud DB
Classic
Advantages
Increase scalability and
reduce the cost.
Caching
Add
or
remove Complex design
memory cache machine
is trival.
4. ISSUES RELATED TO CLOUD TPS
Cloud computing is a resource delivery and usage model. It
gets resource through network “On-demand” and “at scale”
in a cloud environment. The different types of issues related
to Cloud Transaction Processing System [TPS] are [5]:
Limitations
Scalability
and Potential bottleneck
elasticity in the web in the database
and application layers. server.
Security
Technical Issues
Non-technical Issues
Financial Issues
Limitations in Cost.
Partitioning
transparent to
Partitioning
application
programmer.
is Add or remove
the machines in order
to workload it must
involve repartition.
Distributed Business Level
Operational and
Organizational Issues
Add or remove satellite
server depends on the
work load.
ROWA (read once
write all) based on
master copy.
Cheap hardware can be
used to run the DB
For Updateserver.
intensive workloads
Replication Scale – out and scalethe master becomes
down nicely with the bottle neck.
work load.
Hardware and Software
expertise
Consumer Requirements
Service Compared and
tested
Integrity and Infrastructure
Security and Reliability
Issues
Portability
High quality and Low price
Interface with business
process
Speed change
Page 124
International Research Journal of Engineering and Technology (IRJET)
Volume: 02 Issue: 04 | July-2015
www.irjet.net
All the cloud services are cost effective but they have lot of
issues regarding the security and backup [6]. So the cloud
environment needs a secure Transaction Processing System
to maintain the ACID properties.
e-ISSN: 2395 -0056
p-ISSN: 2395-0072
5.2.2) Bottlenecks:
One of the key challenges they face in moving to the cloud is
due to the bottlenecks that occur within their data
architecture. In cloud TPS every transaction is an update
process, so challenging to design database to avoid
bottlenecks in database layer.
5. CHALLENGES OF CLOUD TPS AND DATABASES
5.1. Cloud TPS
5.1.1) Performance:
Industries are challenging with the difficulty of ensuring the
online transaction applications. It has many performance
necessities including transaction speed and near real-time
responsiveness.
5.1.2) ACID guarantees:
It is hard to maintain ACID guarantees in the face of data
replication over huge geographic distances [8]. If a
transaction committed successfully in cloud transaction
processing system that means it satisfied all the ACID
properties.
5.1.3) Security:
There are massive risks in storing transactional data on
third party host. Service providers must make the proper
investments in providing; proving and ensuring appropriate
levels of security over time. Malicious users can utilize weak
spot in the data security model to increase unauthorized
access to data. The security of the enterprise data stored at
the vendor is validating through the subsequent
assessments test.
Cross-site scripting [XSS]
Access control weaknesses
OS and SQL injection flaws
Cross-site request forgery[CSRF]
Cookie manipulation
Hidden field manipulation
Insecure storage
Insecure configuration.
Whichever vulnerability detected during these tests can be
oppressed to gain access to responsive enterprise data and it
may leads to a financial failure [13].
5.2.3) Data Security:
In cloud computing security of data stored at the Cloud
Service Provider (CSP) is a major concern. In traditional
application data are stored on the computer which is reside
at the customer premises. The physical and logical security
should be provided by the customer. In the cloud computing
environment the data are stored on shared environment
which is not under the control of user. Data security and
privacy cannot be guaranteed by physical boundaries of
machines or networks.
5.2.4) Data Privacy:
Privacy is one of the major factor in cloud TPS. It provide the
facility of an individual or group to seclude themselves or
information about themselves and there by revels
themselves selectively. The data privacy is also a matter of
concern in the cloud computing. Companies need to setup a
group of employee to look into the privacy concern of data.
Data in the cloud may be distributed across geographical
border and may not be satisfying privacy law of the
concerned region, in India Information Technology Act,
2000(“IT Act”) and its various sub section governs the data
security and privacy related issues.
5.2.5) Scalability:
Challenge to keep the scalability because traditional
databases do not scale effectively to number of nodes being
deployed in online transaction applications. Scalability
advantages are not without their own complexities. While
we can scale on demand, applications need to be able to
scale with the environment. Every component whether it is
processors, servers, storage or load-balancers that is to
be scaled need some kind of management overhead.
During scaling, it is important to note what percentage of
resource is actually scalable. It is called as scalability
factor [12].
6. CLOUD TRANSACTION PROCESSING SYSTEM
Data transaction management applications are the potential
candidates for deployment in cloud. Each cloud providers
give set of virtual machines to the cloud computing vendors
and they can install their own software for transaction
processing system.
Page 125
International Research Journal of Engineering and Technology (IRJET)
Volume: 02 Issue: 04 | July-2015
www.irjet.net
Elements of Transaction Processing System
Client
Cloud Service Providers
Nodes
Transaction Managers
Two Phase Commit Protocol
RDBMS
Data storage Architecture
e-ISSN: 2395 -0056
p-ISSN: 2395-0072
source databases such as MySQL or similar commercial
databases. These three approaches provide Government IT
leadership with a wide range of capabilities and potential
complexities.
Fig -2: Database as a Service
Fig -1: Cloud TPS
The clients send request to the web and application server.
Application server sends it to the transaction processing
system (TPS). Number of virtual nodes (machines) is
constructed in the TPS. The data entered in to TPS is called a
transaction. A single transaction is divided in to multiple
processes and sends it to sufficient number of nodes. One of
the involved nodes in the particular transaction acts as
Transaction Manager (TM) and it is responsible for other
nodes involved in the transaction. Two phase commit
protocol is used to maintain the ACID properties [10]. Cloud
data storage is used to store the data. It is hard to maintain
ACID properties and security in Cloud RDBMS and Cloud
Data Storage.
The client send request to the enterprise applications. Cloud
providers offer databases and storage space for enterprises
are called Database as a Service (DBaaS) [2]. The enterprise
applications send client data to database through cloud
provider. Data fetched from the cloud storage area and
update the data. Virtual machines are involving to maintain
the Cloud databases and storages.
8. OVERVIEW OF CLOUD RDBMS AND CLOUD DATA
STORAGE
Cloud providers provide different RDBMS for database
management and Data Storage to maintain data. Cost factor
for some of the leading cloud providers are as follows:
TABLE -3: COST FACTOR FOR CLOUD PROVIDERS
Cloud
Providers
RDBMS
Amazon
Amazon
RDS
Amazon
-------
Price
per
hour
$0.017
Storage
Price
per GB
Amazon
S3
$0.0300
-------
Amazon
EBS
$0.125
Page 126
International Research Journal of Engineering and Technology (IRJET)
Volume: 02 Issue: 04 | July-2015
www.irjet.net
e-ISSN: 2395 -0056
p-ISSN: 2395-0072
Google
Cloud SQL
$0.025
Cloud
Storage
$0.026
deployed in the cloud without a fuss. Google Cloud Platform
provides powerful databases that run fast. [15].
Microsoft
Azure
SQL
Database
$0.0067
Storage
$0.024
Rackspace
Cloud
Databases
$0.05
Block
Storage
$0.000164
8.2.2) Cloud Storage:
Google Cloud Storage provides developers and IT
associations durable and highly available object storage.
Google created three simple product options to help and
improve the performance of applications while keeping the
costs low. These three product options use the same API,
providing with a simple and consistent method of
access [15].
8.1. Amazon
8.1.1) Amazon RDS:
Amazon Relational Database Service (Amazon RDS) is a web
service that makes it simple to set up, function, and scale a
relational database in the cloud environment. Amazon RDS
offers cost-efficient and resizable capacity while managing
time-consuming database management responsibilities,
release up to focus on the applications and business [14].
Amazon RDS furnishes online access to the capabilities of a
familiar MySQL, Oracle, Microsoft SQL Server, or PostgreSQL
relational database management systems.
8.1.2) Amazon S3:
Amazon S3 is a cloud storage system for the Internet. It is
designed to make web-scale computing easier for
developers. It provides an easy web-services interface that
can be used to store and retrieve any quantity of data, at
every time, from anywhere on the web. It gives any
developer access to the same highly scalable, reliable,
secure, fast, low-priced infrastructure that Amazon uses to
run its own universal network of web sites [14]. The service
aims to maximize benefits of scale and to pass those benefits
on to developers.
8.1.3) Amazon EBS:
Amazon EBS (Elastic Block Store) provides block level
storage volumes for use with Amazon EC2 instances.
Amazon EBS volumes are off-instance storage that persists
separately from the life of an instance. Amazon EBS volumes
offer consistent and low-latency act needed to run the user’s
workloads. With Amazon EBS, one can scale his usage up or
down within minutes [14].
8.2. Google
8.2.1) Cloud SQL:
Cloud SQL stores and manages data using a fully-managed,
relational MySQL database. Replication, patch management
and database management to ensure availability and
performance is handled by Google. MySQL databases are
8.3 . Microsoft Azure
8.3.1) SQL Database:
SQL Database is a relational database-as-a-service offering
with built-in mission-critical capabilities. It supports
existing SQL Server libraries and APIs, permitting to reuse
code and scripts. It can stand up a SQL Database on Azure in
seconds, without having to manage virtual machines or
infrastructure [16].
8.3.2) Storage:
Azure has four types of storage-blob, file (preview), table,
and queue. The total cost depends on how much quantity is
stored, the volume of storage transactions and outbound
data transfers, and which data redundancy options are
choosen [16].
8.4. Rackspace
8.4.1) Cloud Databases:
Cloud databases provide fast, scalable, fully managed MySQL
database service. It offers a performance-optimized
database for the applications. Cloud Databases give a simple,
on-demand provisioning and open APIs. So, it can deploy
MySQL, Percona Server, or MariaDB with minimal effort
[17].
8.4.2) Block Storage:
Cloud Block Storage offers reliable, high-performance, ondemand storage for applications hosted on Cloud Servers.
Even the most resource-eager workloads run faultlessly,
therefore users can always count on the application to
perform consistently [17].
9. CONCLUSION
Comfortable ACID guaranty is needed for perfect transaction
management system. Unfortunately, the cloud environment
is not fully reliable to maintain ACID properties. It requires a
better architecture to process the data transaction
Page 127
International Research Journal of Engineering and Technology (IRJET)
Volume: 02 Issue: 04 | July-2015
www.irjet.net
management and cloud storage system. Security is an
important factor to consider the transactional data. Because,
data are stored in the third party and replicated from large
geographic distance. The next step is to implement new
architecture to completely maintain ACID properties and
strengthen the security for data transaction management in
cloud environment.
Management Research, Vol.-3, Issue-4, August 2013
ISSN No.: 2250-0758 Pages: 9-12
[12] R. ANANDHI, K. CHITRA, “A Challenge in Improving the
Consistency of Transactions in Cloud Databases Scalability”, nternational Journal of Computer
Applications (0975 – 8887) Volume 52–No.2, August
2012
[13] S. Subashini, V. Kavitha, “A survey on security issues in
service delivery models of cloud computing”, Journal of
Network and Computer Applications, Accepted 11 July
2010.
[14] http://aws.amazon.com/
[15] https://cloud.google.com/
[16] https:// azure.microsoft.com/ en-us/ pricing/
details/ storage/
[17] http://www.rackspace.com/cloud/
BIOGRAPHIES
Prof. J. Antony John Prabu is working as
Assistant Professor and pursuing doctor of
philosophy in Department of Computer
Science, St. Joseph’s College, (Autonomous),
Tiruchirappalli, Tamil Nadu, India. He
received his M.Phil degree from Jamal
Mohamed
College,
Tiruchirappalli.
He received his MCA degree from St. Joseph’s
College, Tiruchirappalli. His area of interest is
Cloud Database Services and Distributed
Technologies.
Dr. S. Britto Ramesh Kumar is working as
Assistant Professor in the Department of
Computer Science, St. Joseph’s College
(Autonomous), Tiruchirappalli, Tamil Nadu,
India. He has published many research articles
in the National/International conferences and
journals. His research interests include Cloud
Computing, Data Mining, Web Mining, and
Mobile Networks.