2013 International Conference on Communication Systems and Network Technologies

Seed Block Algorithm: A Remote Smart Data Back-up Technique for Cloud
Computing
Ms. Kruti Sharma Prof. Kavita R Singh
Department of Computer Technology, Department of Computer Technology,
YCCE, Nagpur (M.S), Nagpur, YCCE, Nagpur (M.S),
441 110, India 441 110, India
[email protected] [email protected]

Abstract— In cloud computing, data generated in electronic complete state of the server that takes care of the heavily
form are large in amount. To maintain this data efficiently, generated data which remains unchanged during storing at
there is a necessity of data recovery services. To cater this, in main cloud remote server and transmission. Integrity plays
this paper we propose a smart remote data backup algorithm, an important role in back-up and recovery services.
Seed Block Algorithm (SBA). The objective of proposed In literature many techniques have been proposed
algorithm is twofold; first it help the users to collect HSDRT[1], PCS[2], ERGOT[4], Linux Box [5], Cold/Hot
information from any remote location in the absence of backup strategy [6] etc. that, discussed the data recovery
network connectivity and second to recover the files in case of process. However, still various successful techniques are
the file deletion or if the cloud gets destroyed due to any
lagging behind some critical issues like implementation
reason. The time related issues are also being solved by
proposed SBA such that it will take minimum time for the
complexity, low cost, security and time related issues. To
recovery process. Proposed SBA also focuses on the security cater this issues, in this paper we propose a smart remote
concept for the back-up files stored at remote server, without data backup algorithm, Seed Block Algorithm (SBA). The
using any of the existing encryption techniques. contribution of the proposed SBA is twofold; first SBA
helps the users to collect information from any remote
Keywords-Central Repository; Remote Repository; Parity location in the absence of network connectivity and second
Cloud Service; Seed Block; to recover the files in case of the file deletion or if the cloud
gets destroyed due to any reason.
I. INTRODUCTION This paper is organized as follows: Section II focuses on
the related literature of existing methods that are successful
National Institute of Standard and Technology defines as to some extent in the cloud computing domain. In Section
a model for enabling convenient, on-demand network access III, we discuss about the remote data backup server. Section
to a share pool of configurable computing service (for ex- IV describes the detailed description of the proposed seed
networks, servers, storage, applications and services) that can block algorithm (SBA) and Section V shows the results and
be provisioned rapidly and released with minimal experimentation analysis of the proposed SBA. Finally, in
management effort or services provider [1]. Today, Cloud Section VI conclusions are given.
Computing is itself a gigantic technology which is
surpassing all the previous technology of computing (like II. RELATED LITERATURE
cluster, grid, distributed etc.) of this competitive and In literature, we study most of the recent back-up and
challenging IT world. The need of cloud computing is recovery techniques that have been developed in cloud
increasing day by day as its advantages overcome the computing domain such as HSDRT[1], PCS[2], ERGOT[4],
disadvantage of various early computing techniques. Cloud Linux Box [5], Cold/Hot backup strategy [6] etc. Detail
storage provides online storage where data stored in form of review shows that none of these techniques are able to
virtualized pool that is usually hosted by third parties. The provide best performances under all uncontrolled
hosting company operates large data on large data center and circumstances such as cost, security, low implementation
according to the requirements of the customer these data complexity, redundancy and recovery in short span of time.
center virtualized the resources and expose them as the Among all the techniques reviewed PCS is comparatively
storage pools that help user to store files or data objects. reliable, simple, easy to use and more convenient for data
As number of user shares the storage and other resources, recovery totally based on parity recovery service. It can
it is possible that other customers can access your data. recover data with very high probability. For data recovery, it
Either the human error, faulty equipment’s, network generates a virtual disk in user system for data backup,
connectivity, a bug or any criminal intent may put our cloud make parity groups across virtual disk, and store parity data
storage on the risk and danger. And changes in the cloud are of parity group in cloud. It uses the Exclusive–OR ( ) for
also made very frequently; we can term it as data dynamics. creating Parity information. However, it is unable to control
The data dynamics is supported by various operations such the implementation complexities.
as insertion, deletion and block modification. Since services On the contrary, HSDRT has come out an efficient
are not limited for archiving and taking backup of data;
technique for the movable clients such as laptop, smart
remote data integrity is also needed. Because the data
phones etc. nevertheless it fails to manage the low cost for
integrity always focuses on the validity and fidelity of the
the implementation of the recovery and also unable to
control the data duplication. It an innovative file back-up Table-I Comparison between various techniques of Back-up and
concept, which makes use of an effective ultra-widely recovery [20]
distributed data transfer mechanism and a high-speed
encryption technology
The HS-DRT [1] is an innovative file back-up concept,
which makes use of an effective ultra-widely distributed
data transfer mechanism and a high-speed encryption
technology. This proposed system follows two sequences
one is Backup sequence and second is Recovery sequence.
In Backup sequence, it receives the data to be backed-up
and in Recovery Sequence, when some disasters occurs or
periodically, the Supervisory Server (one of the components
of the HSDRT) starts the recovery sequence. However there
are some limitation in this model and therefore, this model
is somehow unable to declare as perfect solution for back-up
and recovery.
Rather, Efficient Routing Grounded on Taxonomy
(ERGOT) [4] is totally based on the semantic analysis and
unable to focus on time and implementation complexity. It
is a Semantic-based System which helps for Service
Discovery in cloud computing. Similarly, we found a unique
way of data retrieval. We made a focus on this technique as
it is not a back-up technique but it provide an efficient
retrieval of data that is completely based on the semantic
similarity between service descriptions and service requests.
ERGOT is built upon 3 components 1) A DHT (Distributed
Hash Table) protocol 2) A SON (Semantic Overlay
Network), 3) A measure of semantic similarity among
service description [4]. Hence, ERGOT combines both these
network Concept. By building a SON over a DHT, ERGOT Additionally [10], it shows how service imposed
proposed semantic-driven query answering in DHT-based maximum outage requirements that have a direct effect on
systems. However does not go well with semantic similarity the setting of the SBRR architecture (e.g. imposing a
search models. minimum number of network-wide shared router resources
In addition, Linux Box model is having very simple locations). However, it is unable to include optimization
concept of data back-up and recovery with very low cost. concept with cost reduction.
However, in this model protection level is very low. It also With entirely new concept of virtualization REN cloud
makes the process of migration from one cloud service focuses on the low cost infrastructure with the complex
provider to other very easy. It is affordable to all consumers implementation and low security level. Another technique
and Small and Medium Business (SMB). This solution we found in the field of the data backup is a REN (Research
eliminates consumer’s dependency on the ISP and its Education Network) cloud. The lowest cost point of view
associated backup cost. It can do all these at little cost we found a model “Rent Out the Rented Resources” [17]. Its
named as simple Linux box which will sync up the data at goal is to reduce the cloud service’s monetary cost. It
block/file level from the cloud service provider to the proposed a three phase model for cross cloud federation that
consumer. It incorporates an application on Linux box that are discovery, matchmaking and authentication. This model
will perform backup of the cloud onto local drives. The data is based on concept of cloud vendor that rent the resources
transmission will be secure and encrypted. The limitation from venture(s) and after virtualization, rents it to the clients
we found that a consumer can backup not only the Data but in form of cloud services.
Sync the entire Virtual Machine[5] which somehow waste All these techniques tried to cover different issues
the bandwidth because every time when backup takes place maintaining the cost of implementation as low as possible.
it will do back-up of entire virtual machine. However there is also a technique in which cost increases
Similarly, we also found that one technique basically gradually as data increases i.e. Cold and Hot back-up
focuses on the significant cost reduction and router failure strategy [6] that performs backup and recovery on trigger
scenario i.e. (SBBR). It concerns IP logical connectivity that basis of failure detection. In Cold Backup Service
will be remain unchanged even after a router failure and the Replacement Strategy (CBSRS) recovery process, it is
most important factor is that it provides the network triggered upon the detection of the service failures and it
management system via multi-layer signaling. will not be triggered when the service is available. In Hot

377
Backup Service Replacement Strategy (HBSRS), a intentionally or unintentionally, it should be not able to
transcendental recovery strategy for service composition in access by third party or any other users/client’s.
dynamic network is applied [6]. During the implementation 3) Data Confidentiality
of service, the backup services always remain in the Sometimes client’s data files should be kept
activated states, and then the first returned results of confidential such that if no. of users simultaneously
services will be adopted to ensure the successful accessing the cloud, then data files that are personal to only
implementation of service composition. particular client must be able to hide from other clients on
Although each one of the backup solution in cloud the cloud during accessing of file.
computing is unable to achieve all the issues of remote data 4) Trustworthiness
back-up server. The advantages and disadvantages of all The remote cloud must possess the Trustworthiness
these foresaid techniques are described in the Table-I. And characteristic. Because the user/client stores their private
due to the high applicability of backup process in the data; therefore the cloud and remote backup cloud must play
companies, the role of a remote data back –up server is very a trustworthy role.
crucial and hot research topic. 5) Cost efficiency
The cost of process of data recovery should be efficient
III. REMOTE DATA BACKUP SERVER so that maximum no. of company/clients can take advantage
When we talk about Backup server of main cloud, we of back-up and recovery service.
only think about the copy of main cloud. When this Backup There are many techniques that have focused on these
server is at remote location (i.e. far away from the main issues. In forthcoming section, we will be discussing a
server) and having the complete state of the main cloud, technique of back-up and recovery in cloud computing
then this remote location server is termed as Remote Data domain that will cover the foresaid issues.
Backup Server. The main cloud is termed as the central
repository and remote backup cloud is termed as remote IV. DESIGN OF THE PROPOSED SEED BLOCK ALGORITHM
repository. As discussed in literature, many techniques have been
proposed for recovery and backup such as HSDRT[1],
PCS[2], ERGOT[4], Linux Box[5], Cold/Hot backup
strategy[6] etc. As discussed above low implementation
complexity, low cost, security and time related issues are still
challenging in the field of cloud computing. To tackle these
issues we propose SBA algorithm and in forthcoming
section, we will discuss the design of proposed SBA in
detail.

Fig.1 Remote data Backup Server and its Architecture A. Seed Block Algorithm(SBA) Architechture
And if the central repository lost its data under any This algorithm focuses on simplicity of the back-up and
circumstances either of any natural calamity (for ex - recovery process. It basically uses the concept of Exclusive–
earthquake, flood, fire etc.) or by human attack or deletion OR (XOR) operation of the computing world. For ex: -
that has been done mistakenly and then it uses the Suppose there are two data files: A and B. When we XOR A
information from the remote repository. The main objective and B it produced X i.e. X = . If suppose A data file
of the remote backup facility is to help user to collect get destroyed and we want our A data file back then we are
information from any remote location even if network able to get A data file back, then it is very easy to get back it
connectivity is not available or if data not found on main with the help of B and X data file .i.e. A = .
cloud. As shown in Fig-1 clients are allowed to access the Similarly, the Seed Block Algorithm works to provide the
files from remote repository if the data is not found on simple Back-up and recovery process. Its architecture is
central repository (i.e. indirectly). shown in Fig-2 consists of the Main Cloud and its clients
The Remote backup services should cover the following and the Remote Server. Here, first we set a random number
issues: in the cloud and unique client id for every client. Second,
1) Data Integrity whenever the client id is being register in the main cloud;
Data Integrity is concerned with complete state and the then client id and random number is getting EXORed ( )
whole structure of the server. It verifies that data such that it with each other to generate seed block for the particular
remains unaltered during transmission and reception. It is client. The generated seed block corresponds to each client
the measure of the validity and fidelity of the data present in is stored at remote server.
the server. Whenever client creates the file in cloud first time, it is
2) Data security stored at the main cloud. When it is stored in main server,
Giving full protection to the client’s data is also the the main file of client is being EXORed with the Seed Block
utmost priority for the remote server. And either of the particular client. And that EXORed file is stored at

378
the remote server in the form of file’ (pronounced as File which can be extended as per the necessity. From Table-II, it
dash). If either unfortunately file in main cloud crashed / is observed that memory requirement is more in remote
damaged or file is been deleted mistakenly, then the user server as compare to the main cloud’s server because
will get the original file by EXORing file’ with the seed additional information is placed onto remote server (for
example- different Seed Block of the corresponding client
block of the corresponding client to produce the original file
shown in Fig-2).
and return the resulted file i.e. original file back to the Table-II System Environment
requested client. The architecture representation of the Seed
Block Algorithm is shown in the Fig.2.
B. SBA Algorithm
The proposed SBA algorithm is as follows:
Algorithm 1:
Initialization: Main Cloud: ; Remote Server: ;
Clients of Main Cloud: ; Files: ;
Seed block: ; Random Number: ; During experimentation, we found that size of original
data file stored at main cloud is exactly similar to the size of
Client’s ID:
Back-up file stored at Remote Server as depicted in Table-
Input: created by ; is generated at ;
III. In order to make this fact plausible, we perform this
Output: Recovered file after deletion at experiment for different types of files. Results tabulated in
Given: Authenticated clients could allow uploading, Table-III for this experiment shows that proposed SBA is
downloading and do modification on its own the files only. very much robust in maintaining the size of recovery file
same as that the original data file. From this we conclude that
Step 1: Generate a random number. proposed SBA recover the data file without any data loss.
int Table-III: Performance analysis for different types of files
Step 2: Create a seed Block for each and Store
att .
(Repeat step 2 for all clients)
Step 3: If creates/modifies a and stores at
, then create as

Step 4: Store at .
Step 5: If server crashes deleted from , Processing Time means time taken by the process
then, we do EXOR to retrieve the original as: when client uploads a file at main cloud and that includes
the assembling of data such as the random number from
Step 6: Return to . main cloud, seed block of the corresponding client from the
Step 7: END. remote server for EXORing operation; after assembling,
performing the EXORed operation of the contents of the
uploaded file with the seed block and finally stored the
EXORed file onto the remote server. Performance of this
experiment is tabulated in Table-IV. We also observed that
as data size increases, the processing time increases. On
other hand, we also found that performance which is
megabyte per sec (MB/sec) being constant at some level
even if the data size increases as shown in Table-IV.
Table-IV Effect of data size on processing time

Fig.2 Seed Block Algorithm Architecture

V. EXPERIMENTATION AND RESULT ANALYSIS

379
 The Fig-3 shows the CPU utilization at Main Cloud and REFERENCES
Remote Server. As shown in Fig-3 the Main Cloud’s CPU [1] Yoichiro Ueno, Noriharu Miyaho, Shuichi Suzuki,Muzai Gakuendai,
utilization starts with 0% and as per the client uploads the Inzai-shi, Chiba,Kazuo Ichihara, 2010, “Performance Evaluation of a
file onto it then utilization increases; such that it has to Disaster Recovery System and Practical Network System
Applications,” Fifth International Conference on Systems and
check whether the client is authenticated or not, at the same Networks Communications,pp 256-259.
the time it send request to Remote Server for the [2] Chi-won Song, Sungmin Park, Dong-wook Kim, Sooyong Kang,
corresponding Seed Block. When request reached to 2011, “Parity Cloud Service: A Privacy-Protected Personal Data
Remote Server it started collecting the details as well as the Recovery Service,” International Joint Conference of IEEE
TrustCom-11/IEEE ICESS-11/FCST-11.
seed Block and gives response in form of the seed Block
[3] Y.Ueno, N.Miyaho, and S.Suzuki, , 2009, “Disaster Recovery
and during this period, load at Main Cloud decreases which Mechanism using Widely Distributed Networking and Secure
in return cause for gradual decreases in CPU utilization at Metadata Handling Technology”, Proceedings of the 4th edition of
main cloud. After receiving the requested data, CPU the UPGRADE-CN workshop, pp. 45-48.
utilization at main cloud increases as it has to perform the [4] Giuseppe Pirr´o, Paolo Trunfio , Domenico Talia, Paolo Missier and
EXORed operation. Again the Final EXORed file sends to Carole Goble, 2010, “ERGOT: A Semantic-based System for Service
Discovery in Distributed Infrastructures,” 10th IEEE/ACM
Remote Server. As compared to Table-IV the processing International Conference on Cluster, Cloud and Grid Computing.
time given can be compare with the time showing in Fig-3. [5] Vijaykumar Javaraiah Brocade Advanced Networks and
Telecommunication Systems (ANTS), 2011, “Backup for Cloud and
Disaster Recovery for Consumers and SMBs,” IEEE 5th International
Conference, 2011.
[6] Lili Sun, Jianwei An, Yang Yang, Ming Zeng, 2011, “Recovery
Strategies for Service Composition in Dynamic Network,”
International Conference on Cloud and Service Computing.
[7] Xi Zhou, Junshuai Shi, Yingxiao Xu, Yinsheng Li and Weiwei Sun,
2008, "A backup restoration algorithm of service composition in
MANETs," Communication Technology ICCT 11th IEEE
International Conference, pp. 588-591.
Fig.3 Graph Showing Processor Utilization [8] M. Armbrust et al, “Above the clouds: A berkeley view of cloud
computing,”http://www.eecs.berkeley.edu/Pubs/TechRpts/2009//EEC
The Fig-4 shows the experimentation result of proposed S-2009-28.pdf.
SBA. As fig-4 (a) shows the original file which is uploaded [9] F.BKashani, C.Chen,C.Shahabi.WSPDS, 2004, “Web Services Peer-
by the client on main cloud. Fig-4 (b) shows the EXORed to-Peer Discovery Service ,” ICOMP.
file which is stored on the remote server. This file contains [10] Eleni Palkopoulouy, Dominic A. Schupke, Thomas Bauscherty, 2011,
the secured EXORed content of original file and seed block “Recovery Time Analysis for the Shared Backup Router Resources
content of the corresponding client. Fig-4 (c) shows the (SBRR) Architecture”, IEEE ICC.
recovered file; which indirectly sent to client in the absence [11] Lili Sun, Jianwei An, Yang Yang, Ming Zeng ,2011,“Recovery
Strategies for Service Composition in Dynamic Network,”
of network connectivity and in case of the file deletion or if International Conference on Cloud and Service Computing, pp. 221–
the cloud gets destroyed due to any reason. 226.
[12] P.Demeester et al., 1999, “Resilience in Multilayer Networks,” IEEE
Communications Magazine, Vol. 37, No. 8, p.70-76.
[13] S. Zhang, X. Chen, and X. Huo, 2010, “Cloud Computing Research
and Development Trend,” IEEE Second International Conference on
Future Networks, pp. 93-97.
[14] T. M. Coughlin and S. L. Linfoot, 2010, “A Novel Taxonomy for
Consumer Metadata,” IEEE ICCE Conference.
Fig.4 Sample output image of SBA Algorithm [15] K. Keahey, M. Tsugawa, A. Matsunaga, J. Fortes, 2009, “Sky
Computing”, IEEE Journal of Internet Computing, vol. 13, pp. 43-51.
VI. CONCLUSION [16] M. D. Assuncao, A.Costanzo and R. Buyya, 2009, “Evaluating the
Cost- Benefit of Using Cloud Computing to Extend the Capacity of
In this paper, we presented detail design of proposed SBA Clusters,” Proceedings of the 18th International Symposium on High
algorithm. Proposed SBA is robust in helping the users to Performance Distributed Computing (HPDC 2009), Germany.
collect information from any remote location in the absence [17] Sheheryar Malik, Fabrice Huet, December 2011, “Virtual Cloud:
of network connectivity and also to recover the files in case Rent Out the Rented Resources," 6th International Conference on
Internet Technology and Secure Transactions,11-14 ,Abu Dhabi,
of the file deletion or if the cloud gets destroyed due to any United Arab Emirates.
reason. Experimentation and result analysis shows that [18] Wayne A. Jansen, 2011, “Cloud Hooks: Security and Privacy Issues
proposed SBA also focuses on the security concept for the in Cloud Computing, 44th Hawaii International Conference on
back-up files stored at remote server, without using any of System Sciences.Hawaii.
the existing encryption techniques. The time related issues [19] Jinpeng et al, 2009, “Managing Security of Virtual Machine Images
are also being solved by proposed SBA such that it will take in a Cloud Environment”, CCSW, Chicago, USA.
minimum time for the recovery process. [20] Ms..Kruti Sharma,Prof K.R.Singh, 2012, “Online data Backup And
Disaster Recovery techniques in cloud computing:A review”, IJEIT,
Vol.2, Issue 5.

380