A Backup-as-a-Service (BaaS) Software Solution

Heitor Faria, Priscila Sol

ıs, Jarcir Bordim and Rodrigo Hagstrom

Department of Computer Science (CIC), University of Brasilia (UnB), Brasilia, DF, Brazil

Keywords:

Backups, Disaster Recovery, Cloud, Backup-as-a-Service, Software-as-a-Service.

Abstract:

Backup is a replica of any data that can be used to restore its original form. However, the total amount of digi-

tal data created worldwide more than doubles every two years and is expected to reach 44 trillions of gigabytes

in 2020, bringing constant new challenges to backup processes. Enterprise backup is one of the oldest and

most performed tasks by infrastructure and operations professionals. Still, most backup systems have been de-

signed and optimized for outdated environments and use cases. That fact, generates frustration over currently

backup challenges and leads to a greater willingness to modernize and to consider new technologies. Tradi-

tional backup and archive solutions are no longer able to meet users current needs. The ideal modern backup

and recovery software should not only provide features to attend a traditional data center, but also allow the

integration and exploration of the growing Cloud, including “backup client as a service” and “backup storage

as a service”. The present study proposed and deploys a Backup as a Service software solution. For that, the

cloud/backup parameters, cloud backup challenges, researched architectures and Backup-as-a-Service (BaaS)

system requirements are speciﬁed. Then, a selected set of BaaS desired features are developed, resulting in the

ﬁrst truly cloud REST API based Backup-as-a-Service interface, namely “bcloud”. Finally, this work conducts

an on-line usability poll with a signiﬁcant number of users. The analysis of results in an overall average ob-

jective zero to ten questions evaluation was 8.29%, indicating a very satisfactory user perception of the bcloud

BaaS interface prototype.

1 INTRODUCTION

In the words of (Guise, 2008), backup is the replica

of any data that can be used to restore its original

form. In other words, backup is a valid copy of data,

ﬁles, applications or operating systems that can serve

for the recovery purpose. That is why, customarily,

backup is often stored to lower cost of high capacity

removable media, such as magnetic tapes, normally

stored in ﬁreproof safes or another protected physical

environment.

Conforming to the International Data Corporation

(IDC, 2014), the total amount of digital data created

worldwide more than doubles every two years. It will

grow from 4.4 zettabytes in 2013 to 44 zettabytes by

2020. For example, one zettabyte is equal to 1 trillion

gigabytes. According to (Russell et al., 2016), en-

terprise backup is among the oldest most-performed

tasks for infrastructure and operations professionals.

Still, most backup systems have been designed and

optimized for outdated environments and use cases.

As stated by (Silva, 2014), traditional backup and

archive solutions are no longer able to meet users cur-

rent needs.

In line with (Kaiser et al., 2016), the massive in-

crease in data generation and processing in industry

and academia has dramatically increased the pressure

on backup environments. Data itself has become a

very valuable asset, and its protection requires the use

of reliable backup systems.

In the vision of (Russell et al., 2016), the ideal

modern currently backup and recovery software prod-

ucts should not only provide features to attend a tra-

ditional data center, e.g.: backup to conventional

random-access or sequential media (hard disk, solid-

state, tape drives), data reduction techniques (com-

pression, deduplication or single instancing) and sys-

tems interoperability. However, this software must

also allow the integration and exploration of the grow-

ing Cloud, including “backup client as a service” and

“backup storage as a service”.

Moreover, even many currently solutions are sold

as BaaS, some of them do not meet the new market re-

quirements for data center backup and recovery soft-

ware (Russell et al., 2016). Backup software for a

homogeneous environment is also excluded, such as

native tools from Microsoft or VMware for their spe-

ciﬁc platforms.

Faria, H., Solís, P., Bordim, J. and Hagstrom, R.

A Backup-as-a-Service (BaaS) Software Solution.

DOI: 10.5220/0007250902250232

In Proceedings of the 9th International Conference on Cloud Computing and Services Science (CLOSER 2019), pages 225-232

ISBN: 978-989-758-365-0

225

This work in this paper aims to propose and de-

ploy a BaaS solution, under the Remote Backup to

the Cloud architecture. For that, we determined the

cloud/backup parameters, cloud backup challenges,

researched architectures and BaaS system require-

ments. Then, a set of features were selected, devel-

oped and implemented to attend the architecture. Fi-

nally, we applied a user-poll technique with a signif-

icant amount of users to validate the developed fea-

tures and analyzed their answers.

The paper is organized as follows : in Section 2 we

present the related work of this research. In Section

3, we present the Proposal of the BAAS Solution, in

which we used as a starting point some previous work

of the authors (Bacula The Open Source Backup Soft-

ware (book) (Faria, 2016). Storage Growing Forecast

with Bacula Backup Software Catalog Data Mining

(Faria et al., 2017a). Backup Storage Block Level

Deduplication with DDUMBFS and BACULA (Faria

et al., 2017b), A Hadoop Open Source Backup Solu-

tion (Faria, 2018)). . In Section 4 shows the Evalua-

tion Methodology and Results. Section 5 draws some

conclusions and future work

2 RELATED WORK

In correspondence with (Armbrust et al., 2010; Buyya

et al., 2011), cloud computing is a long-held idea of

computing as a general utility. It promises to shift data

and computational services from individual devices to

distributed architectures. As reported by (Columbus,

2017; Ried et al., 2011; Arean, 2013), cloud com-

puting services global total market revenues value is

expected to grow from U$ 67 billion by 2015 to U$

241 billion by the end of 2020. Still in 2013, near

to 61% of United Kingdom businesses were relying

on some cloud services. Cloud computing is becom-

ing more popular (Khoshkholghi et al., 2014) in large-

scale computing because of its ability to share glob-

ally distributed resources.

Pursuant to (Khoshkholghi et al., 2014), Disaster

Recovery (DR) is a persistent problem in Information

Technology (IT) platforms, and even more crucial in

cloud computing. A Cloud Service Provider (CSP)

must ﬁnd ways to provide services to their customers

even if the data center is down (v.g.: due to a disas-

ter). Disasters can lead to expensive service disrup-

tion (Khoshkholghi et al., 2014), regardless of the na-

ture of their causes. A CSP can adopt two different

DR models: traditional and cloud-based service mod-

els. The use of the traditional model can happen as

either a dedicated infrastructure or shared approach.

Nevertheless, as claimed by IBM (Raju et al.,

2012) the weather causes only 50% of disasters in its

cloud, and the rest because of miscellaneous causes

(e.g., cut power lines, server hardware failures, and

exploitation of security vulnerabilities). In this way,

disaster recovery is not only a mechanism to deal with

natural events but also for all severe disruptions that

may also happen with the modern cloud environment

services.

Along with (Khoshkholghi et al., 2014), there are

three deﬁned DR levels:

Data Level. The security of application data.

System Level. Reducing system recovery time as

short as possible.

Application Level. Application continuity.

As stated by (Alhazmi and Malaiya, 2013), Re-

covery Point Objective (RPO) and Recovery Time

Objective (RTO) there are two main parameters that

all recovery mechanisms should observe. If RPO and

RTO values are lower, the systems can achieve higher

business continuity. RPO might be interpreted as the

amount of lost data a disaster. RTO consists on the

time frame between disruption and restoration of ser-

vice. As demonstrated by equation 1, the Recov-

ery Point Objective value is inversely proportional to

the frequency of backups terminated along the time,

where FB represents the Frequency of Backup.

RPO ∝

(1)

On the other hand, as exhibited by equation 2,

Recovery Time Objective formula usually includes a

fraction of RPO, the readiness of the backup and ﬁve

failover steps delays, depending on backup capabili-

ties.

RTO = fraction of RPO+jmin+S1+S2+S3+S4+S5

(2)

Each variable used in the equation 2 has the fol-

lowing description:

fraction of RPO. Computation time lost since the

last backup.

jmin. Depends on service readiness of the backup.

S1. Hardware setup time.

S2. Operating System initiation time.

S3. Application initiation time.

S4. Data or process state restoration time.

S5. Internet Protocol (IP) address switching time.

Therefore, as alleged by (Wood et al., 2010), DR

mechanisms must have ﬁve requirements for an efﬁ-

cient performance:

CLOSER 2019 - 9th International Conference on Cloud Computing and Services Science

226

• Minimum RPO and RTO.

• Minimal impact on the normal system operation.

• Should be geographically separated.

• The application shall be restored to a consistent

state.

• DR solution must guarantee integrity, privacy, and

conﬁdentiality.

These general requirements are technology inde-

pendent and can also affect the Cloud-Based Disaster

Recovery Models.

In (Khoshkholghi et al., 2014), a disaster is de-

ﬁned as an unexpected event in the lifetime of a sys-

tem. It may consist (1) of natural causes, such as

tsunami or earthquake; (2) software or hardware fail-

ures; (3) human error or sabotage. As claimed by

(Kashiwazaki, 2012), it can lead to signiﬁcant ﬁnan-

cial or human lives loss. However, only approxi-

mately 2% to 4% of the IT infrastructure budget in

huge companies are annually destinated to DR solu-

tions (Prakash et al., 2012).

As maintained by (Khoshkholghi et al., 2014),

cloud-based DR solution is an increasing trend be-

cause of its ability to tolerate disasters and to achieve

the reliability and availability.

As shown in Figures 1, 2, 3 and 4, different archi-

tectures were proposed to overcome the cloud-related

DR challenges: Remote Backup to the Cloud (Cama-

cho et al., 2014), Local Backup from the Cloud (Is-

mail et al., 2013; Javaraiah, 2011), Cloud Geograph-

ical Redundancy and Backup (GRB, (Pokharel et al.,

2010)), Inter-private Cloud Storage (IPCS, (Jian-hua

and Nan, 2011)) and Secure-Distributed Data Backup

(SDDB, (Ueno et al., 2010)).

Figure 1: Remote Backup to the Cloud.

Figure 2: Local Backup from the Cloud.

According to (Ismail et al., 2013) and under the

veriﬁed cloud backup architectures, the backup solu-

tion must support different backup data, since differ-

ent layers of operation originate them. Still, to offer

backup as an Anything-as-a-Service model, it must

Figure 3: Cloud Geographical Redundancy and Backup.

Figure 4: GRB and IPCS Cloud Backup Architectures.

inherit the cloud characteristics such as self-service,

on demand and ease-of-use.

3 PROPOSAL OF THE BaaS

SOLUTION

There are currently on-premise self-hosted data cen-

ter backup solutions that can evolve with some devel-

opment support towards a BaaS delivery solution for

CSPs, but other relevant open source based solutions

other than the Gartner Magic Quadrant (Russell et al.,

2016) are also considered.

The primary criteria to choose one of the solutions

would be the integration with cloud storage; the vari-

ety of plug-ins for speciﬁc application backup; stan-

dard cloud industry interfaces such as REST API and

open backup catalog format for more natural BaaS

development and the licensing costs. The commer-

cial versions of Arcserve, Bacula Enterprise, Bareos,

EMC solutions, IBM Spectrum Protect, Veeam and

Veritas Netbackup were evaluated. Due to the exclu-

sive REST API availability at the time, broad Cloud

S3 Object backup storage support, open catalog for-

mat, one of the widest range of applications backup

plug-ins, reasonable pricing and developer availabil-

ity to help on the BaaS interface development, Bacula

Enterprise was considered the most suitable backup

software to deploy a minimum BaaS product in a

short time frame.

This work focused on one of the reviewed Cloud

backup architectures. According to (Heslin, 2014),

self-owned data centers and collocation are still the

most common IT infrastructure model used by the

companies. Therefore, the proposed solution aims

A Backup-as-a-Service (BaaS) Software Solution

227

to prototype a BaaS solution for the Remote Backup

to the Cloud, since it would serve to the majority of

enterprises. In this scenario and considering Cloud

is user-centric, the most notable and critical require-

ments are interface related, such as a multi-tenant

user-friendly operation. Now the backup Clients are

deployed and conﬁgured by the ordinary Cloud User

instead of the old backup administrator specialist.

The traditional backup software is not multi-tenant,

and as observed by (Amvrosiadis and Bhadkamkar,

2015) they are too complicated even for specialized

users since the majority of them uses stock backup

software conﬁgurations for schedules and backup re-

tentions.

A list of 17 requirements were deﬁned for an ideal

BaaS interface, but only the four more critical were

prioritized and evaluated. They were: reconﬁgura-

tion, personalization and customization capabilities;

multi-tenancy; easy new tenants creation; and user-

centered design. The prototype development stage

had four months of duration until the interface Alpha

version release, which involved the web interface de-

velopment and backup client wizards to ease and au-

tomate their deploy.

As shown in Figure 5, bcloud topology relies on

the Bacula REST API that enables control of Bacula

with high-level HTTPS/REST calls.

Figure 5: bcloud topology.

The REST API plays a crucial role on the BaaS

interface development, creating an abstraction layer

for Bacula operation, granting a faster programming

and making unnecessary to know or to change Bacula

source code.

We selected the Smarty PHP framework for de-

velopment. The use of a popular PHP framework will

allow the screens to be customized, and will also al-

low future addition of custom modules.

The Tray Monitor is a multi-platform QT program

that enables to monitor the Client (File Daemon) and

start backup jobs, especially when behind inaccessi-

ble ﬁrewalls or NATed networks. The end user will

be able to use the Tray Monitor program to start client

initiated backup jobs or to perform other Bacula re-

lated actions.

The bcloud offers a PHP module (class and tem-

plates) that will list a local server directory and the

sub-directories for self-service backup clients down-

load.

As displayed in Figure 6, the user will access

the Download Center to fetch multi-platform backup

client installation packages and conﬁguration wiz-

ards, to automatically attach its machines to the

backup system.

Figure 6: Client Download Center.

As exhibited in Figure 7, the backup client Regis-

tration Wizard can run in operating systems with and

without graphical interfaces. That is important espe-

cially for servers that usually doesn’t have a GUI and

are the principal Bacula service object.

Figure 7: Conﬁguration Wizard.

The interface saves all the backup server con-

ﬁguration in a temporary area (named work set).

The Client is conﬁgured with the server information

(Name, Password and TLS certiﬁcate). The Client’s

local daemon is started. The registration program

must be executed on the Client with the necessary

permissions to edit the daemon conﬁguration ﬁle,

the Tray Monitor conﬁguration ﬁle and control the

backup service (root on Unix and Administrator on

Windows).

By default, the CSP administrator will be notiﬁed

about the registration request and can commit, modify

or discard the request.

All Bacula Clients that are attached to a Director

will be located on networks that may be protected by

CLOSER 2019 - 9th International Conference on Cloud Computing and Services Science

228

ﬁrewalls or behind routers. A direct connection from

the Director to the File Daemon using the TCP port

9102 may not be possible without a VPN conﬁgura-

tion.

Once installed, conﬁgured, and approved by the

Cloud Administrator, the backup clients will appear

at the “My Data” bcloud screen. From this mo-

ment, as shown in Figure 8, the user can proceed with

backup job conﬁguration steps, such as choosing what

folders, ﬁles or applications will be backed up, and

scheduling. Ad hoc queuing is also supported.

Figure 8: Job Conﬁguration.

As presented in Figure 9, a dashboard shows all

terminated and on-going jobs. It displays information

such as backup duration, number of ﬁles, size, trans-

fer rate, start and termination time. The user can ﬁlter

listed jobs by level, status, client, and time frame, us-

ing the same tab options.

Figure 9: Job Terminated Jobs Overview.

Finally, as displayed in Figure 10, the Cloud User

can select one or more terminated backup jobs, gener-

ally according to their termination time, and proceed

with ﬁle browsing and data restore. Differential and

incremental backups can be selected, manually or au-

tomatically, together with last full and other necessary

jobs to provide a complete restore to a given time.

Figure 10: Restore Job File Browsing.

4 EVALUATION

METHODOLOGY AND

RESULTS

The case study presents two basic bcloud roles: Cloud

Service Administrator and the Cloud User. The ﬁrst is

the backup-as-a-service provider, that access bcloud

to perform administrative tasks or another Bacula

CLI, GUI or other web-based interfaces. The second

one is the backup-as-a-service costumer, who receives

a tenant and executes the steps to set up the backup of

its data.

To validate the developed features, we set up a

small environment for user tests over the Internet. The

initial requirements are similar to a Bacula backup

server since bcloud is just one more hosted interface.

We used a machine with 64 bits virtualized CentOS

operating systems with 100 GB of disk space, and 16

GB of RAM. Also, a 512 GB direct-attached solid-

state drive was set up to host the global deduplica-

tion backup engine since it requires faster random ac-

cess for on-the-ﬂy processing. The backup server lo-

cal network trafﬁc relies on gigabit Ethernet connec-

tions, but the client backup trafﬁc data goes through

the Internet. This capacity considers a small work-

load, estimated to support up to 100 backup clients,

50 simultaneous bcloud web interface users and 50

TB of backups.

Since Cloud services are user-centric, focused on

user-experience and accessed by a large number of

customers with different proﬁles and technical lev-

els, we chose the individual inquiry technique for the

bcloud evaluation in this study.

The user inquiry was performed using documen-

tation, prototype user access, and an online question-

naire that contained six objective zero to ten eval-

uation questions. That scale was chosen for being

practical and for providing a granular perception of

A Backup-as-a-Service (BaaS) Software Solution

229

the evaluated features. Nevertheless, a general user-

feedback open question was made available, and the

variety of answers were categorized and presented.

A total of 278 ﬁllings were considered, correspond-

ing to a population of 1000 users, with an error mar-

gin of 5% and a conﬁdence level of 95%. We used

the Normal (Gaussian or Gauss or Laplace–Gauss)

continuous probability distribution, to represent real-

valued random variables whose distributions are not

known. The random variables observations sample

averages drawn from independent distributions con-

verge to the normal, in other words, they become nor-

mally distributed when the number of observations is

sufﬁciently large.

The statistical population is higher than most

small and medium company staff sizes, in a hypothet-

ical and extreme scenario where every employee be-

comes a Backup-as-a-Service user. Worldwide users

were invited to test the bcloud interface, follow the

Cloud User and Cloud Administrator workﬂow sug-

gestion, and provide feedback via questionnaires. The

users received a research description, the interface

user manual and the credentials to access the proto-

type. The poll sought heterogeneous user-proﬁles,

with different technical background and seniority lev-

els. Invites were sent to infrastructure, cloud, operat-

ing system, backup and even Bacula user groups.

We monitored the bcloud web engine logs to make

sure each invited customer was able to access the

solution, perform the deﬁned user roles and start a

backup. We evaluated the BaaS solution according

to the four cloud requirements that we chose earlier.

The “Interface Reconﬁguration Capabilities”, “Multi-

Tenancy Support”, and “Easy New Tenants Creation”

criteria had one question each. The “User-Centered

Design” was divided into three questions: “Reduced

User Operation Errors”, “Improved User Acceptance

and Satisfaction”, “and Improved Productivity”. We

present the questions as follows.

4.1 Analysis of Results

In this section, we present the average questionnaire

evaluation for the ﬁrst three questions as follows:

1. Reconﬁguration, personalization and customiza-

tion of the BaaS interface

2. Bacula single instance of the object code and

database supports multiple customers

3. Easy new tenants creation

As shown in Figure 11, the average replies about

the investigated features deploy ranged from roughly

8 to an 8.5 score. The “Interface Reconﬁguration Ca-

Figure 11: Questions 1, 2 and 3 average evaluations.

pabilities” was the worse evaluated criteria and the

“Multi-Tenant Deploy” was the best.

We present the Design User-centered design ques-

tion results as follows:

1. Reduced backup operation error capacities

2. Improved user acceptance and satisfaction

3. Improved productivity

Figure 12: Questions 4, 5 and 6 average evaluations.

As shown in Figure 12, the average replies ranged

from 8.1 to an 8.35 score. The “Improved User Ac-

ceptance and Satisfaction” interface quality was the

worse evaluated, and the “Reduced Backup Opera-

tion Error Capabilities” was the best-evaluated crite-

ria. The user-feedback open question (7) received a

high number of different themes, and they were sum-

marized as follows:

1. Usability and user technical level concerns

2. Safety and performance considerations

3. More User experience work is needed

4. Innovative solution

5. Useful software

6. Adherent to the cloud trend

The three ﬁrst topics were considered more detri-

mental to the user interface perception and will be

taken into consideration for the future prototype im-

provement. The last three are more positive replies

which reafﬁrm the importance of the Backup-as-a-

Service solution development. The overall average

CLOSER 2019 - 9th International Conference on Cloud Computing and Services Science

230

objective questions mark was 8.29%, what indicates a

very satisfactory perception of the bcloud BaaS inter-

face prototype in such a short time-frame. We present

and comment each one of the objective user-inquiry

answers result from highest to the lowest as follows.

Multi-Tenancy Support - 8.52. This result is a little

bit surprising, because there are no gray shades of

multi-tenancy on the prototype deployment. By

default, each user only sees and manages its own

backup clients and jobs, even though bcloud has

multi-party capabilities. We report the score not

being perfect to the user lack of knowledge about

multi-tenancy, and to a fail in explain that to the

inquired subjects.

Easy New Tenants Creation - 8.4. We chose not to

create an ”add user” feature directly on the bcloud

interface, because it would elevate the complexity

considerably.

Improved Productivity - 8.37. As mentioned on the

literature review, many backup administrators just

want to use the backup software defaults. We

associate that fact to the good grade achieved

in this question, and automated features such as

the backup clients conﬁguration must have con-

tributed.

Reduced Backup Operation Error Capacities -

8.32.

The bcloud interface abstracts many of the com-

plexities of current data center backup software

ones. The users can select and download only the

backup clients and plug-ins that the administrator

provides, and the installation wizard makes

much easier to install the new clients. The user-

experience focused design abstracts complicated

backup elements such as storage capacity, type,

pools and volumes. Users of virtually any level

can operate bcloud.

Improved User Acceptance and Satisfaction - 8.07.

This result was a little disappointing, but it can

be explained on the lack of demonstrations, any

“help” content or built-in tutorial, due to the lack

of time.

Interface Reconﬁguration Capabilities - 8.04. We

report this lower result to the fact there was no

mention on the bcloud user manual on how to

customize the bcloud interface, even though it is

CSS based.

Despite of not being into the scope of the formal

analysis, the user submitted globally deduplicated

backups average transfer rate was 1.41 MB/s. The

deduplication is a fundamental technique for cloud

backups to use the minimum possible Internet link

bandwidth, so in this case lower values are better.

Even though, small links are still a bottleneck for

data that was never backed up (e.g., ﬁrst machine full

backup) since there are no similar blocks in the dedu-

plication engine to avoid the transfer.

5 CONCLUSIONS AND FUTURE

WORK

The present study developed and prototyped a BaaS

solution, under the “Remote Backup to the Cloud”

architecture. The cloud/backup parameters, cloud

backup challenges, researched architectures and BaaS

system requirements were determined. A set of fea-

tures were selected to be developed and implemented

to attend the chosen architecture. We surveyed the

backup software alternatives according to consultancy

indicated leaders and the most popular solutions. Ac-

cording to the BaaS macro-requirements and cost, we

chose Bacula’s Enterprise version for the BaaS In-

terface prototype deployment. The already existing

REST API feature, open catalog format, cloud backup

storage capabilities, a variety of speciﬁc applications

backup plug-ins and reasonable cost were the main

reasons for this decision. We deﬁned a list of 17

requirements for an ideal BaaS interface, we prior-

itized and evaluated the four feasible within the lim-

ited time frame. They were: reconﬁguration personal-

ization and customization capabilities; multi-tenancy

support; easy new tenants creation; and user-centered

design. Considering these features, we developed a

new web interface named bcloud. As the result of this

work, the users received a research description, the in-

terface user manual and the credentials to access the

prototype interface. A sample was determined to at-

tend to most Small and Medium Business in an ex-

treme scenario where every corporate user would be a

BaaS user, rendering a total of 278 ﬁllings. That cor-

responds to a thousand users, with an error margin of

5% and a conﬁdence level of 95%. The overall aver-

age objective zero to ten questions mark was 8.29%,

indicating a very satisfactory perception of the bcloud

BaaS interface prototype. The “Multi-Tenancy Sup-

port” and “Easy new tenants creation” were the best

evaluated criteria according to the user feedback. The

“Improved User Acceptance and Satisfaction” qual-

ity and the “Interface Reconﬁguration Capabilities”

were the worse evaluated topics, and we are going to

improve the interface according to these feedbacks.

As an improvement to the prototype deployment, we

are going to add interface help content and demonstra-

tions, so even most unexperienced user will be able to

use the interface. That might improve the end-user

A Backup-as-a-Service (BaaS) Software Solution

231

acceptance and satisfaction perception. Still, we plan

to incorporate a CSS code editor to the bcloud admin-

istrator interface, so it will be easier for the admin-

istrator to perform design conﬁguration changes. We

are going to assist, study and evaluate a bcloud pro-

duction deployment, and will conduct another inquiry

to verify the product evolution.

REFERENCES

Alhazmi, O. and Malaiya, Y. (2013). Evaluating Disaster

Recovery Plans Using the Cloud.

Amvrosiadis, G. and Bhadkamkar, M. (2015). Identify-

ing Trends in Enterprise Data Protection Systems. In

USENIX Annual Technical Conference, pages 151–

164.

Arean, O. (2013). Disaster recovery in the cloud. Network

Security, 2013(9):5–7.

Armbrust, M., Stoica, I., Zaharia, M., Fox, A., Grifﬁth, R.,

Joseph, A. D., Katz, R., Konwinski, A., Lee, G., Pat-

terson, D., and Rabkin, A. (2010). A view of cloud

computing. Communications of the ACM, 53(4):50.

Buyya, R., Broberg, J., and Go

scinski, A. (2011). Cloud

computing: principles and paradigms. Wiley, Hobo-

ken, N.J. OCLC: ocn606774387.

Camacho, H., Brambila, A., Pe

na, A., and Vargas, J. (2014).

A Cloud Environment for Backup and Data Storage.

Columbus, L. (2017). Roundup Of Cloud Computing Fore-

casts.

Faria, H. (2016). Bacula The Open Source Backup Soft-

ware.

Faria, H. (2018). A Hadoop Open Source Backup Solution.

Faria, H., Carvalho, R., and Solis, P. (2017a). Storage

Growing Forecast with Bacula Backup Software Cata-

log Data Mining. In Computer Science & Information

Technology (CS & IT), pages 185–196. Academy &

Industry Research Collaboration Center (AIRCC).

Faria, H., Luiz Bordim, J., and Solis Barreto, P. (2017b).

Backup Storage Block Level Deduplication with

DDUMBFS and BACULA. International Journal of

Advanced Information Technology, 7(4):1–9.

Guise, P. d. (2008). Enterprise Systems Backup and Recov-

ery: A Corporate Insurance Policy. CRC Press, USA.

Google-Books-ID: 2OtqvySBTu4C.

Heslin, K. (2014). 2014 Data Center Industry Survey.

IDC (2014). Executive Summary: Data Growth, Business

Opportunities, and the IT Imperatives | The Digital

Universe of Opportunities: Rich Data and the Increas-

ing Value of the Internet of Things.

Ismail, B. I., Mydin, M. N. M., and Khalid, M. F. (2013).

Architecture of scalable backup service for private

cloud. In Open Systems (ICOS), 2013 IEEE Confer-

ence on, pages 174–179. IEEE.

Javaraiah, V. (2011). Backup for cloud and disaster recov-

ery for consumers and SMBs. In Advanced Networks

and Telecommunication Systems (ANTS), 2011 IEEE

5th International Conference on, pages 1–3. IEEE.

Jian-hua, Z. and Nan, Z. (2011). Cloud Computing-based

Data Storage and Disaster Recovery. pages 629–632.

IEEE.

Kaiser, J., S

uß, T., Nagel, L., and Brinkmann, A. (2016).

Sorted deduplication: How to process thousands of

backup streams. In Mass Storage Systems and Tech-

nologies (MSST), 2016 32th Symposium on. IEEE.

Kashiwazaki, H. (2012). Practical uses of cloud computing

services in a Japanese university of the arts against af-

termath of the 2011 Tohoku earthquake. In Proceed-

ings of the 40th annual ACM SIGUCCS conference on

User services, pages 49–52. ACM.

Khoshkholghi, M. A., Abdullah, A., Latip, R., Subrama-

niam, S., and Othman, M. (2014). Disaster Recovery

in Cloud Computing: A Survey. Computer and Infor-

mation Science, 7(4):39.

Pokharel, M., Lee, S., and Park, J. S. (2010). Disaster Re-

covery for System Architecture Using Cloud Comput-

ing. pages 304–307. IEEE.

Prakash, S., Mody, S., Wahab, A., Swaminathan, S., and

Paramount, R. (2012). Disaster recovery services

in the cloud for SMEs. In Cloud Computing Tech-

nologies, Applications and Management (ICCCTAM),

2012 International Conference on, pages 139–144.

IEEE.

Raju, M. R., Prakash, J. P., and Rao, G. R. (2012). Disaster

Recovery of Servers using Virtualized Cloud Comput-

ing.

Ried, S., Kisker, H., and Bartels, A. (2011). Sizing The

Cloud.

Russell, D., Rinnen, P., and Rhame, R. (2016). Magic

Quadrant for Data Center Backup and Recovery Soft-

ware.

Silva, T. J. (2014). Uma arquitetura de cloud storage para

backup de arquivos.

Ueno, Y., Miyaho, N., Suzuki, S., and Ichihara, K.

(2010). Performance Evaluation of a Disaster Recov-

ery System and Practical Network System Applica-

tions. pages 195–200. IEEE.

Wood, T., Cecchet, E., Ramakrishnan, K. K., Shenoy, P. J.,

van der Merwe, J. E., and Venkataramani, A. (2010).

Disaster Recovery as a Cloud Service: Economic

Beneﬁts & Deployment Challenges. HotCloud, 10:8–

15.

CLOSER 2019 - 9th International Conference on Cloud Computing and Services Science

232