There is a new concept of database management called cloud
database management system and the standard architecture concept of
cloud database management system is not been implemented yet. We are going to
present an architecture of cloud database management system consist of five
layers. The first layer is Interface Layer through which a user
can interact considering the issues of security, transparency and
manageability. The second layer is the Logical Middleware Layer, the major
issue that can be considered in this layer is interoperability due to
availability of heterogeneous databases in the market. The third layer is the Logical
Layer in which the issues of managing transaction, processing a query,
techniques used for programming and optimization should be considered. The
forth layer is the Database Middleware Layer and the biggest issue it has is; interoperability
between different environments because there are various platforms are
available. And the fifth layer is the Database Layer which can finally have
data so the issues that must be considered are security of data, creating
backup, data partitioning, fault tolerance, storage, replication, scaling and
In cloud the applications that manage data are possible applicants for
deployment. The cloud database naturally runs in a cloud computing environment.
The two reasons for concerning Database_as_a_service(DBaaS) are:
In economical way when the users have to pay for a part of a service, the
cost of the energy and the hardware is quite lower than doing everything
In a well-managed DBaaS the
cost will be directly related to the real usage (pay per use), this smear to
cost of administrative and the licensing of software.
Cloud database was considered for the determination of online database,
using the variability of distributed databases. Database_as_a_service(DBaaS) and
Application service provider(ASP) are two paradigms in which
cloud computing is meaningful. As DBaaS contributions are strongly
combined with further Platform_as_a_service(PaaS) therefore it bounces the
chance to organization to focus on evolving their goods rather than wasting of capitals
on management of the platform. For holding these services, the data centers are
used. Data centers use service hardware for storing and computation.
State of work
The three significant tasks like Well-organized multi tenancy, secrecy of
database and resistance of scalability has not been spoken in former databases.
Some points that are based on awareness of workload are follows:
Better performance and higher consolidation of multi
tenancy on a database server can be achieved by awareness of workload approach
than prevailing approaches.
Elasticity can be achieved by using partitioning
based on graph for even complex transactional workloads.
Use of some effective encryption algorithm that
enables structured query language queries to run over encrypted data.
Up to date information stored in the cloud often need to be combined
with the data stored in relational databases. An encrypt named “Big integrator” to approve queries that
associate data in the cloud-based data stores with relational
databases, there are several cloud-based systems available but with
limited query languages.
A principle “SQLMR”, which is a cross over approach to fill the
gap between SQL-based and MapReduce (Sea-chart synopsize)
data processing. Sea-chart synopsize provides
surroundings for broad data
processing and is shown to be scalable and responsibility tolerant on commodity machines. In any way it is complex to decide than SQL-like languages
and maintaining and reuse are shout
easy. On the transformation associate set SQL-based processing is not scalable but users are aware
To join the fault tolerating, inexact cluster and ease of use
out of the box capability of MapReduce with the efficiency, performance and
tool plug ability of shared nothing parallel database systems, Hybrid answer is needed. This Solution have a significant impact on
the cloud database
market. Cloud database organizes cloud architecture into a hierarchy of three
levels; cloud data center, cloud service provider, and client.
Layered Architecture of Cloud Database
There are different challenges in the
scalability of cloud applications like confidentiality of data, availability
of service, shared-nothing architecture. Cloud database system will be
successful if it following maximum goals like scalability, multitenancy, fault
tolerance, elasticity, availability, flexible query interface, capability to
run in heterogeneous environment. A standard architecture for the cloud-based
systems is needed to meet the goals of cloud and improve the challenges. Yet
the standard architecture is not introduced, so we are proposing the layered
architecture for cloud-based systems as shown in the figure 1.
Figure 1: Layered architecture of cloud database
All layers including their issues are
Users interact with this layer. Lot of
companies are shifting their applications of data management from high-end
servers, expensive to the cloud-based system which is
combination of cheaper, commodity machines. The main goal of the service
provider is to provide and manage the services with full security and
Manageability: This layer
manages the different users. Also keeps record of when (time) the specific user
uses the cloud database management system. And also generate report like
how long they used the service, and their number of transactions. Then transfer
the control to the next layer after validity and authenticity.
Security: In this layer
the authentication can be achieved by using passwords and user id’s. The
database can be only accessed by the legal one’s and the illegal one’s can be
Transparency: This layer
gives the transparency to the user which is a major advantage of cloud-based
systems, means the users will not know the actual placement of data.
Development of real-time applications are easier with transparency which
supports mobility, flexibility etc.
This layer ensures that different
databases like DB2, SQL, oracle hide the conceptual level heterogeneity. As
exposed in figure 2.
Figure 2: Logical middleware layer
Interoperability: This layer must provide interoperability means
process different types of communication between different databases. For
example, if one database wants to share information with another database, they
should be able to share without any difficulty.
This layer represents the logics of
whole database and dealing with data in internal processing. As cloud-based
systems deals with different types of data, users have to combine the old-style
data with the data placed on cloud-based system so different
types of systems are required which can have all these functionalities. Users
can do their tasks lacking knowledge of the logics behind. Different languages
are developed for cloud-based systems like SQLMR, Bigintegrator, SOQL (saleforce
object query language) which give results base on their analysis. This layer
deals with processing of efficient query, techniques of programming and query
Techniques of programming: This layer provide solution to the problem like
which technique of programming or query language is better for databases.
Query optimization and processing: The goal of this layer is to give output to the user
in minimum cost and time using query optimizer which takes the query and
discovers the execution plan which is cheapest among different execution plans
giving same solution.
Security: One of the major concern of this layer is that the
data should not be lost, copied, disclosed or altered by any illegal
(unauthorized) users by providing different methods of authorization control.
Database Middleware Layer
This layer ensures that the
heterogeneity among different platforms like MacOS, Windows, Linux must be
hided. As visible in figure 3.
Figure 3: Database middleware layer
Interoperability: In this layer interoperability means execution of
different types of communication between different platforms. User using
windows operating system should not find any type of difficulty having
connection with Man operating system.
In this layer actual data can be
represented. The responsibility of database layer is to configure and monitor
the database to achieve ideal scaling, effective allocation of resources and
multitenancy. This layer has different issues like privacy, shared nothing
architecture or shared disk architecture, partitioning and data security which
are described below.
Replication and backup: As data availability is most important for cloud-based
systems the companies should use different techniques for data duplication and
backups for restoring in any case of loss.
Partitioning: The partitioning techniques are used to share and
balance the load which improves the availability, scalability and performance
database systems in multitenant platforms.
Storage: Where the data should be stored that it can be retrieved in minimum
time. Shared disk architecture is suitable for the cloud-based systems. Sophisticated
caches should be used to maintain the metadata in memory.
Indexing: The each and every file in the database should be
indexed which increases scalability performance.
Load balancing: It must be intelligent enough to automatically
divide load among users to make the best utilization of the resources and avoid
any deadlock accruing situation.
Fault tolerance: Cloud-based systems should be intelligently
designed to solve or minimize any type of failure and maintains the system
operational. It includes deadlock detection and recovery techniques and
Security: Using encryption- decryption techniques this layer
can provide security to the raw data. These techniques can prevent different
threats like modification, privacy and fabrication.
In this work a layered architecture for cloud
database management system is proposed. Functions of each layer (interface,
middleware, logical, database middleware and database)
are discussed with different issues and challenges of each layer.