Expert answer:database management challenges
Expert answer:Search paper file:Shelar,
M., Sane, S., Kharat, V., Database Management Challenges in Cloud Environment,
International Journal of Modern Trends in Engineering and Research, 2016.After
critical analysis of the given research paper, discuss the various database management challenges in cloud environment and the
proposed techniques to overcome them ???.Note: The
solution must be given in your own words with relevant examples not exceeding 2
pages. The given research paper is for your reference only and your answer
should not be limited to this research paper only. You may refer various other
resources and provided references for the same.Plagiarism
please its zero or 3%
database_management_challenges_in_cloud_environment.pdf
Unformatted Attachment Preview
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/308994340
DATABASE MANAGEMENT CHALLENGES IN
CLOUD ENVIRONMENT
Article · September 2016
DOI: 10.21884/IJMTER.2016.3067.9NEO0
CITATIONS
READS
0
561
3 authors:
Madhukar Shelar
Shirish Sane
kthm college
K. K. Wagh Institute of Engineering Educatio…
8 PUBLICATIONS 8 CITATIONS
12 PUBLICATIONS 21 CITATIONS
SEE PROFILE
SEE PROFILE
Vilas Kharat
Savirtibai Phule Pune University
52 PUBLICATIONS 57 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Ph.D. Research Scholar at Department of Computer Science in Savitribai Phule Pune University. View
project
All content following this page was uploaded by Madhukar Shelar on 11 October 2016.
The user has requested enhancement of the downloaded file.
DATABASE MANAGEMENT CHALLENGES IN CLOUD
ENVIRONMENT
Madhukar Shelar1 , Shirish Sane 2 and Vilas Kharat3
Ph.D. Research Scholar, Department of Computer Science, Savitribai Phule Pune University
2
Department of Computer Engg, K.K.Wagh Institute of Engineering Education and Research, Nashik
3
Department of Computer Science, Savitribai Phule Pune University
1
Abstract— Recently cloud computing is widely used technology in delivering computing resources
as a service. There are variety of web-based applications hosted on cloud computing platforms,
majority of them are data driven. Therefore database management is now become critical component
and challenging task for the DBMS designers and researchers. Scalability, elastic load balancing,
pay-per-use pricing and self- managing are the major reasons for the successful cloud database
management. This is the review of research published/presented on databases in cloud platform.
Various challenges of managing databases in the cloud and various techniques proposed by
researchers to face these challenges are presented in this review. It is based on various parameters
like database systems in the cloud, scalability and elasticity, autonomy or self- managing database
systems and preserving consistency of database in cloud.
Keywords—Cloud Computing, Cloud Database Management, Scalability, Elasticity, Autonomic
Database Management
I. INTRODUCTION
Cloud computing is the latest trend in distributed computing. Majority of cloud applications
are data driven, so DBMS software powering these applications form critical component in the cloud
software stack [4]. Traditional DBMS software’s are not cloud friendly, because cloud services such
as web servers and application servers which can scale from few machines to hundreds. DBMS
cannot be scaled very easily [1]. However, existing databases resulting in a heavy performance
impact during elastic scaling [2]. In Cloud computing environment, DBMS must follow the
properties such as scalable, elasticity, autonomic and load balancing in additio n to the other wellknown properties like high level functionality, consistency, performance and reliability. This review
surveyed the existing research papers and literatures using systematic approach. We have searched
the major research on database management in cloud computing environment in IEEE-Xplore, ACM
Digital Library, Springer Link and Google Scholar. We have decided to select limited papers on the
desired properties of cloud databases such as scalability, elasticity, autonomy, reliability, availability
and consistency.
II. DATABASES IN THE CLOUD
In cloud environment, data is stored on multiple dynamic servers at data centers rather than
on dedicated servers like in traditional data storage. When storing a database, user sees a virtual
server, but in reality the data is stored on any one or more of the servers at data centers. Currently
cloud platforms have very little support for database design related virtualization enhancement, but
in future designing databases specific for cloud is sure possibility [11]. Divyakant Agrawal et.al [12]
presented a tutorial which broadly described the challenges faced by the application developers and
DBMS designers in developing and deploying internet scale applications.
The components for the cloud service like application servers, web servers are scaled
dynamically as need arises. However DBMSs and RDBMSs cannot be scaled very easily, so they are
not cloud friendly [1]. Many data management key-value store technologies such as Google’s
AppEngine, Microsoft’s Azure, Amzon’s AWS are used today to maintain databases in cloud
DOI:10.21884/IJMTER.2016.3067.9NEO0
208
International Journal of Modern Trends in Engineering and Research (IJMTER)
Volume 03, Issue 09, [September– 2016] ISSN (Online):2349–9745; ISSN (Print):2393-8161
platform. In these systems, data is represented as Key-Value pairs, and atomic access is provided
only at the granularity of single keys. While these properties are insufficient for the web applications
such as online gaming, social networks, collaborative editing, and many more which emphasize
collaboration[20], because they requires consistent access to groups of keys, scalable and consistent
multi key access is critical for such applications. S udipto Das et.al [20] design and implement GStore which uses a key- value store as an underlying substrate to provide efficient, scalable, and
transactional multi key access.
Key-Value stores have emerged as a preferred choice for scalable and fault tolerant data
management, but lack the rich functionality, and transactional guarantees of RDBMS [22]. Sudipto
Das et.al [22] presented ElasTraS, an Elastic Transactional relational database, designed to scale out
using a cluster of commodity machines while being fault-tolerant and self- managing. Robin Bloor
[8] suggested the need of CDBMS for database management in cloud environment. CDBMS is a
distributed database that can deliver a query service across multiple distributed database nodes
located in multiple data centers, including cloud data centers. Robin has focused on the point that,
traditional RDBMS and column store databases have a centralized architecture and such architecture
would encounter a scalability limit at some point within as well as between data centers. Hence, they
could not fulfill the role of CDBMS. He examined the suitability of Algebraix Data’s technology
product A2DB to fulfill the role of a cloud database.
Most database systems have scalability limit. Once users reach these scalability limits, data
migration and load balancing are the only recourse [5]. Relational cloud, Databases as a Service
(DBaaS) model proposed by Carlo Curino et.al [5,9], which has three components – database
partitioning, live migration, workload analysis and allocation. It overcomes the challenges like
efficient multi-tenancy, elastic scalability and database privacy
III. DATABASE SCALABILITY AND ELASTICITY
Scalability of system indicates its ability to either handle growing amounts of work in
graceful manner when additional resources are added. In cloud computing environment, there is need
to support virtually unlimited number of users for web-based applications by making it scalable.
Distributed database systems for cloud applications emphasize scalability, fault tolerance and
availability, sometimes at the expense of consistency or ease of development [4]. Database
scalability in cloud can be achieved in two ways- Data Fusion and Data Fission [1]. First approach
Data Fusion architecture maintains multi-key atomicity while ensuring scalability. D. Agrawal et.al
[1, 20] designed a scalable data store, G-Store which gives multi-key access guarantees than Megastore designed by Google. Second approach is a Data Fission, where database schema is partitioned
instead of keeping individual tables. This approach tries to minimize the distributed transactions. A
prototype system ElasTras [1, 22] uses the concept of data-fission with static-partitioning approach.
Luis et.al [13] suggested to combine three mechanisms for handling the database scalability
in PaaS – distributed caching, NoSQL databases, and database clustering. Distributed caching
provides the same cache across several nodes, which is useful in clustered applications. NoSQL
databases refer to the storage solutions for structured data that are different from the traditional
relational database systems. Finally clusters can be built to provide better scalability, availability and
fault tolerance to DBMS system.
A scalability, in which a system can be scaled-up dynamically by adding more nodes or can
scaled down by removing nodes as per the load fluctuations, is referred to as Elasticity [1]. Google’s
Bigtable and Yahoo’s PNUTS are Key-Value store elastic databases. In cloud computing
environment, Database Elasticity and pay-per-use pricing model are major factors for the success of
the cloud [1]. Database elasticity can be achieved in two ways, live database migration and virtual
machine cloning.
Many researchers have proposed techniques for live database migration o n different cloud
architectures. A technique Zephyr [1, 10] efficiently migrates live database in shared nothing
transactional database architecture. Zephyr minimizes service interruption for the tenants being
migrated by introducing synchronized dual mode [10]. An Iterative copy [1, 19] technique based on
@IJMTER- 2016, All rights Rese rved
209
International Journal of Modern Trends in Engineering and Research (IJMTER)
Volume 03, Issue 09, [September– 2016] ISSN (Online):2349–9745; ISSN (Print):2393-8161
shared disk architecture focuses on transferring the main memory state of the database partition.
S.Das et.al [2] proposed a technique Albatross for live data migration in shared storage database
architecture. This technique is efficient and low cost technique for live migration of tenant database
in multitenant DBMS. Migration is a key component for elasticity and load balancing, and hence,
migration should be supported as a first class notion in any multitenant DBMS [21]. Multitenant
DBMS uses different multitenancy models for resource sharing viz. shared machine, shared process
and shared table. Shared process model provides good balance of effective resource sharing, schema
diversity, performance and scaling. Albatross is based on Shared process model. It provides minimal
impact on tenant SLA by working in three phases – Initialization, iterative copying and atomic
handover.
Virtual Machine cloning is another way to achieve the elasticity scaling in clo ud databases.
Emmanuel et.al [7] designed a virtual machine cloning technique – Dolly for database provisioning
in shared nothing architecture. It assumes that cloud platforms employ a virtualized architecture.
Each database replica runs in a separate virtual machine. Dolly clones entire virtual machine of an
existing replica. The cloned virtual machine is started on new physical server. It then synchronizes
the state of other replicas, which was before the processing of user request. Dolly triggers the rep lica
spawning process well in advance of the increase in workload. It uses the model to estimate the
latency to spawn a new replica based on virtual machine snapshot size and the database
resynchronization latency.
Carlo Curino et.al [5, 9] designed a Relational Cloud, a DBaaS(Database As A Service)
model from scratch to adapt to the peculiarities of the cloud computing environment. It achieves
elastic scalability using graph based partitioning method to spread larger databases across many
machines.
IV. AUTONOMIC DATA MANAGEMENT
Autonomic or self- managing is another requirement in cloud database management which is
related to the scalability and elasticity. In traditional database management system trained DBAs
look after the system and takes appropriate actions to improve the performance of database system.
In cloud computing environment database is maintained at data centers on many servers, the manual
approach of database administration is not feasible. Database autonomy is essential in the cloud for
monitoring the behavior and performance of the system, elastic scaling and load balancing
dynamically. D. Agrawal et.al [1] suggested for the need of more research on developing an
autonomous and intelligent system controller to manage large systems. Self- management of the data
by the cloud will be critical to support large number of users with no database expertise [4].
S.Das et.al [2] designed a technique Albatross for live data migration in shared storage
architecture to achieve elasticity. They planned to extend the design by adding an intelligent system
control that can model the cost of migration to predict its cost as well as the behavior of the entire
system. It will leads to the autonomy in database management.
Cloud database management services are part of PaaS (Platform-As-A-Service). PaaS service
providers generates revenue from clients as per SLAs (Service Level Agreements) and may pay cost
for the rent their resources from IaaS (Infrastructure-As-A-Service) service providers. The cloud
database service providers has to meet the clients SLA as well as to maximize own profit. Zephyr [1,
10] technique focuses on minimizing the service interruption and synchronizes the execution of
source and destination node transactions, whereas Albatross [3] technique not only minimizes the
service interruption but also minimizes the impact on performance of transaction.
P.Xiong et.al [3] presented a cost aware resource management system named SmartSLA to
intelligently manage the resources in shared cloud database syste m. The SmartSLA considers two
important issues- local analysis and global analysis. The former issue is to identify the right
configuration of system resources to meet the clients’ SLA. The later issue is to take decision on how
resources among clients are based on current system status. This system captures relationship
between system resources and database performance using machine learning technique.
@IJMTER- 2016, All rights Rese rved
210
International Journal of Modern Trends in Engineering and Research (IJMTER)
Volume 03, Issue 09, [September– 2016] ISSN (Online):2349–9745; ISSN (Print):2393-8161
Allocating resources accurately for tenants in cloud data center results in the reduction of
extra cost. Jing Bi et.al [18] presented an optimal autonomic virtual machine provisioning technique
for cluster based virtualized multi-tier applications. They have presented a self- managing
community mechanism to automate the VMs of configuring and tuning the virt ualized multitier
applications. Ashraf et.al [17] outlined some of the challenges such as placement, resource
partitioning, service level objectives, dynamically varying workloads etc. which are associated with
deploying database appliances as virtual appliances on IaaS and tuning their performance. To address
these challenges, tools and techniques viz. performance models, optimization and control algorithms,
tools for system administrators, co-tuning and hint passing are proposed.
V. CONSISTENCY IN CLOUD DATABASE
Data availability and durability is the main requirement of the cloud vendors. It can be
achieved by maintaining certain number of replicas of the data at different locations. However to
maintain the consistency between these replicas become a complicated issue. Md. Asfakul et.al [14]
propose a tree based consistency approach that maximizes the consistency and performance. The tree
is formed between primary server to all replica servers in such a way that the maximum reliable path
is ensured between them, which reduce the probability of transaction failure.
It is hard to maintain the ACID guarantees if data is replicated over large geographic
distances [15]. When data is replicated over a wide area, consistency part of ACID is typically
compromised to yield reasonable system availability. Ozsu etl.al [4] concluded that, current solutions
of the cloud data management are successful, but developed with specific simple applications in
mind. They have sacrificed consistency and ease of programming for the scalability and elasticity.
VI. CONCLUSION
Extensive research work has been published generally on cloud computing and particularly
on management cloud databases. Majority of the work is focused on dynamic scalability or elasticity
and load balancing using low cost live data migration on shared nothing as well as shared storage
architecture. Since DBMS and RDBMS have scalability limit, many researchers have preferred to
use key-value store databases. However due to lack of rich functionality and transactional
guarantees, some researchers designed self- managing and scalable database systems. Cloud
databases have sacrificed consistency and ease of programming for the scalability and elasticity.
Since cloud computing is gaining growing popularity in the IT industry, further research is
needed on efficient and effective management of cloud databases by achieving scalability, elasticity
with load balancing and autonomy without sacrificing database consistency.
REFERENCES
[1] D. Agrawal, A.E.Abbadi, S.Das and A.J.Elmore, “Database Scalability, Elasticity and Autonomy in Cloud”,
[2] S.Das, S.Nishimura, D.Agrawal, A.E.Abbadi, “A lbatross: Lightweight Elasticity in Shared Storage Databases for the
Cloud Using Live Data Migration”, Proc.VLDB Endowment, Vol.4 No.8, 2011.
[3] P.Xiong, Yun Chi, S.Zhu, H.J.Moon, Calton Pu, Hakan H., “Intelligent Management of Virtualized Resources for
Database Systems in Cloud Environment”, IEEE 2011, Proc. ICDE Conf.2011.
[4] Özsu M. Tamer, Valduriez, Patrick
[5] 3rd Edition. Springer, 2011
[6] Carlo Curino, Evan Jones, Yang Zhang, Eugene Wu, “Relational Cloud: The Case for a Database Service”, ACM,
2010.
[7] Herald Kllapi, Eva Sitaridi, M.M .Tsangaris, “Schedule Optimizat ion for Data Processing Flows on the Cloud”,
SIGMOD’11, ACM, 2011.
[8] Emmanuel Cecchet, Rahul Singh, Upendra Sharma, Prahant Shenoy, “Dolly : Virtualization – driven Database
Provisioning for the cloud”, VEE’11, ACM, 2011.
[9] Robin Bloor, “What is a cloud database? The suitability of Algebraix Data’s Technology to cloud co mputing”,
www.BloorGroup.com, The Bloor Group, 2011.
[10] Carlo Curino, Evan P.C.Jones, Raluca Ada Popa et.al., “Relat ional Cloud: A Database -as-a-Service for the Cloud”,
CIDR, 2011.
[11] Aaron J. Elmo re, Sudipto Das, Divyakant Agrawal, A mr El Abbadi: Zephyr: Live migration in shared nothing
databases for elastic cloud platforms, SIGMOD Conference 2011: 301-312.
@IJMTER- 2016, All rights Rese rved
211
International Journal of Modern Trends in Engineering and Research (IJMTER)
Volume 03, Issue 09, [September– 2016] ISSN (Online):2349–9745; ISSN (Print):2393-8161
[12] Arptia Mathur, Mridul Mathur, Pallavi Upadhyay, “Cloud Based Distributed Databases: The Future Ahead”, IJCSE,
ISSN:0975-3397, Vol.3 No.6 June 2011.
[13] Divyakant Agrwal, Suditpto Das, Amr El Abbadi, “Big Data and Cloud Co mputing: Current State and Future
Opportunities”, EDBT 2011, Mar 22-24, 2011 ACM.
[14] Lu is M.Vaquero, Lu is Rodero-Merino, Rajku mar Buyya, “Dynamically Scaling Applications in the Cloud”, ACM
SIGCOMM, Vol 41, Number 1, Jan 2011.
[15] Md. Asfakul Isalam, Susan V. Vrbsky: Tree-based Consistency Approach for Cloud Databases, 2nd Intl. IEEE
Conference on Cloud, 2011, 401-404.
[16] Daniel J. Abadi: Data Management in the Cloud : Limitations and Opportunities, Bulletin of IEEE Co mputer
Society, Technical Committee on Data Engineering, 2009.
[17] Sean Marston, Zhi Li, Subhajyoti, Anand Ghalasi : Cloud Co mputing : The Business Perspective, Proceedings 44th
Hawaii Intl. Conference, IEEE, 2011.
[18] Ashraf Aboulnaga, Kenneth Salem, Ahmed So ror, Umar F.M inhas, Peter Kokosielis, Sunil Kamath: Deploying
Database Appliances in the Cloud, Bulletin IEEE Computer Society Tech.Committee on Data Engineering, 200
[19] Jing Bi, Zhiliang Zhu, Ruixiong T …
Purchase answer to see full
attachment
How it works
- Paste your instructions in the instructions box. You can also attach an instructions file
- Select the writer category, deadline, education level and review the instructions
- Make a payment for the order to be assignment to a writer
- Download the paper after the writer uploads it
Will the writer plagiarize my essay?
You will get a plagiarism-free paper and you can get an originality report upon request.
Is this service safe?
All the personal information is confidential and we have 100% safe payment methods. We also guarantee good grades