Yogesh Chimate et al, / (IJCSIT) International Journal of Computer Science and Information Technologies, Vol.

5 (6) , 2014, 8101-8102

Review on Preserving Cloud Consistency Using

CAAS Model
Yogesh Chimate, Prof. P. B. Kumbharkar
Department Of Computer Engineering , Siddhant college of engineering,
Savitribai phule university Pune, India.

Abstract- Cloud storage services are commercially more document that can check whether the data cloud provide a
popular due to their amount of advantages. Most of the cloud promised level of consistency or not. Two-level auditing
service provider provides services like infrastructure structure which require only a loosely synchronized clock
management, data storage services on 24/7 through any for ordering operation in an audit cloud then perform
devices at anywhere. To provide this ubiquitous always on
service most of the cloud service provider (CSP) maintains
global auditing with a global trace of operations
each piece of data on geographically distributed servers. The periodically an auditor is elected from an audit cloud. Local
main key problem with this technique is that, it is very auditing is concentrate on monotonic-read and read-your-
expensive and some to fail to provide required consistency of write consistencies, which can be performed by an online
service. To overcome this problem, we propose to use a new light-weight algorithm while Global auditing focuses on
approach of service (i.e. Consistency as a Service (CaaS)) this causal consistency, in which construct a directed graph. If
paper, firstly concentrate on a consistency as a service (CaaS) the constructed graph is a directed acyclic graph also called
model, which has a large data cloud and multiple small audit as precedence graph, we claim that causal consistency is
clouds. In the CaaS model, a data cloud is formed by a CSP, preserved. We determine the severity of violations by two
and a group of users form an audit cloud that can verify
whether the data cloud provides the promised level of
metrics for the CaaS model: commonality of violations and
consistency i.e. quality of service or not, for that make use of staleness of the value of a read, as in. Finally, we propose a
two-level auditing strategy which require loosely synchronized heuristic auditing strategy (HAS) which adds appropriate
clock for ordering operations in an audit cloud. Then perform reads to display as many violations as possible to determine
global auditing by global trace of operations through cloud consistency and also actual cost per transaction.
randomly electing an auditor from an audit cloud. Finally, use
a heuristic auditing strategy (HAS) to display as many II. LITERATURE SURVEY
violations as possible. This section reviews some of the related work followed by
discussing their connections and differences with the
Keywords— Cloud Storage Systems, consistency as a service
(CaaS), two-level auditing and heuristic auditing strategy
proposed approach.
(HAS). A. TANENBAUM AND M. VAN STEEN [2] proposes the
two classes of consistency models i.e. data centric
I. INTRODUCTION consistency and client centric consistency. Data centric
Clouds computing is become more popular as it provides consistency focuses on internal state of storage system
guaranteed services like data storage, virtualized while the client centric consistency focuses on what specific
infrastructure etc. e.g. Amazon, SimpleDB etc. By using the customer want. To verify these levels of consistency
cloud services, the customers or user can access data stored provided by CSP is done by two ways: trace based
in a cloud anytime and at anywhere using any device, and verification and benchmark based verification.
customer ensure about less capital investment. To provide W. VOGELS [3] sates that strict consistency is never
promised always on 24/7 access, the cloud service provider required in practice also it is considered to be harmful. Thus
(CSP) stores data replicas on multiple geographically The Amazon.com technology provides a set of more
distributed servers. The main drawback of using the advanced business and infrastructure services that are
replication technique is it is very expensive to achieve implemented using scalable distributed systems. In this
strong consistency, and user is ensured to see the latest environment we can analyse a number of particular data
updates. Many CSPs (e.g., Amazon S3) provide only access patterns, each with their own consistency
eventual i.e. updates are visible definitely but not requirements. Thus to provide a collection of more diverse
immediately. E.g. Domain name system (DNS), but the business processes the different patterns are provided.
eventual consistency is not interesting for all applications E. Brewer [4] and Pushing the CAP [5] states that Modern
and which require strong consistency. Some applications distributed Systems have adopted new types of data store
like social networking sites require causal i.e. strong that are not secure to provide strong Consistency. Thus the
consistency. Thus the different applications require CAP theorem and its evolution will influence on these
different level of consistency. We propose novel systems which provide scarifies strict consistency and
consistency as a service (CAAS) model. The CaaS model provide weak consistency with high availability.
consists of, A large data cloud formed by CSP and multiple E. Anderson, X. Li, M. Shah, J. Tucek, and J. Wylie[6]
audit clouds formed by group of users worked on project or propose efficient algorithm that analyzing the trace of
interactive operations between the client machines and key

www.ijcsit.com 8101
 Yogesh Chimate et al, / (IJCSIT) International Journal of Computer Science and Information Technologies, Vol. 5 (6) , 2014, 8101-8102

value store i.e. data store to report weather the trace is safe , same order and concurrent writes may be seen in a different
regular or atomic, because number of data store provide order on different machine.
platform for always on globally distributed applications. To 3.4 Heuristic Auditing Strategy
meet their goals they scarify strong consistency and provide From the auditing process it is clear that only reads can
high availability. display violations by their values. Therefore, the basic idea
While some of the tools used to verify the violations. And behind the heuristic auditing strategy (HAS) is to add exact
conclude whether system is good or bad and provide reads for displaying as many violations as possible and call
promised level of consistency or not. these additional reads as auditing reads. Under the CaaS
W. GOLAB, X. LI, AND M. SHAH [7] provides an online model, consistency becomes a part of the Service Level
verification algorithms by using GK algorithms for several Agreement and the users can get something from the CSP,
known consistency properties. Also we consider how to by displaying consistency violations and determine the
determine the severity of the violations, if consistency severity of the violations. The CaaS model will help both
violations? And this is check by two matrices: value of old the CSP and the users adopt consistency as an important
reads, and another is the commonality of violations. For not aspect of cloud services.
fresh data again consider time-based staleness and
operation-count-based staleness. Finally present an efficient III. CONCLUSION
algorithm that calculates these quantities. In this paper, we argued that strong consistency
requirements should be adopted only for data objects
III. INTRODUCTORY crucial for application correctness, otherwise weaker forms
This section consist of three models i. e. consistency as a of data consistency should be adopted. We presented a
service (CaaS) model, user operation table (UOT) with consistency as a service (CaaS) model and a two-level
which each user records his operations and two-level auditing structure that helps users to verify whether the
auditing structure. cloud service provider (CSP) is providing the promised
3.1 Consistency as a Service (CAAS) Model consistency, and to quantify the severity of the violations, if
An audit cloud consists of a group of users that work any. With the CaaS model, the users can assess the quality
together on a job, e.g., a document or a program. We of cloud services and choose a right CSP among various
consider that each user in the audit cloud is identified by a candidates, e.g., the least expensive one that still provides
unique ID. Before assigning job to the data cloud, an audit adequate consistency for the users’ applications.
cloud and the data cloud will engage with a service level
agreement (SLA), which demands the promised level of ACKNOWLEDGE
consistency should be provided by the data cloud. The audit I would like to express my sincere thanks to my guide
cloud exists to verify whether the data cloud violates the Prof. P. B. Kumbharkar for his motivation and useful
SLA or not, and to analyze the severity of violations. suggestions which truly helped me in improving the quality
3.2 User Operation Table (UOT) of this paper. I take this opportunity to express my thanks to
Each user maintains his own User Operation Table (UOT) my teacher, family and friends for their encouragement and
for recording his trace of operations. Each record in the support.
UOT is described by elements like Operation, logical vector,
and physical vector. While issuing an operation, a user from REFERENCE
an audit cloud will record his operation in UOT, as well as [1] in Liu, Guojun Wang, IEEE, Member, and Jie Wu, IEEE Fellow
"Consistency as a service: Auditing cloud consistency", IEEE Vol 11
his current logical vector and physical vector. Each user No.1, July 2014.
will maintain a logical vector and a physical vector to track [2] A. TANENBAUM AND M. VAN STEEN, Distributed Systems:
the logical and physical time when an operation happens, Principles andParadigms. Prentice Hall PTR, 2002
respectively. [3] W. VOGELS, “Data access patterns in the Amazon.com technology
platform,” in Proc. 2007 VLDB.
3.3 Two-level auditing structure [4] E. BREWER, “Towards robust distributed systems,” in Proc. 2000
Local Auditing: Each user independently performs local ACM PODC.
auditing with his UOT with two consistencies; Monotonic- [5] Pushing the CAP: strategies for consistency and availability,”
read consistency, which requires that a user must read either Computer, vol. 45, no.2, 2012
a new value or same value Read-your’s-write consistency, [6] “What E. ANDERSON, X. LI, M. SHAH, J. TUCEK, AND J. WYLIE,
consistency does your key-value store actually provide,” in
which require a user, always read his latest update. Proc. 2010 USENIX HotDep.
Global Auditing : Global auditing is performed by global [7] W. GOLAB, X. LI, AND M. SHAH, “Analyzing consistency
trace of operations of all users operations with following properties for fun and profit,” in Proc. 2011 ACM PODC.
consistency Causal Consistency Causal consistency writes
that are causally related must be seen by all process in the

www.ijcsit.com 8102