A Deployable Data as a Service Architecture for Enterprises

Adri

an T

oth

and Mouzhi Ge

Faculty of Informatics, Masaryk University, Brno, Czech Republic

Deggendorf Institute of Technology, Deggendorf, Germany

Keywords:

Service Computing, Cloud Computing, as a Service, Data as a Service.

Abstract:

Nowadays, data have been considered as one of the valuable assets in enterprises. Although the cloud com-

puting and service-oriented architecture are capable of accommodating the data asset, they are more focused

on software or platforms rather than the data per se. Thus, data management in cloud computing is usually not

prioritized and not well organized. In recent years, Data as a Service (DaaS) has been emerged as a critical

concept for enterprises. It beneﬁts from a variety of aspects such as data agility and data quality manage-

ment. However, it is still unknown for enterprises why and how to develop and deploy a DaaS architecture.

This paper is therefore to design a deployable DaaS architecture that is based on the as-a-Service principles

and especially tackles data management as a service. To validate the architecture, we have implemented the

proposed DaaS with a real-world deployment.

1 INTRODUCTION

Big Data has received increasing attention in recent

years, as organizations and cities are dealing with

tremendous amounts of data with high complexity

and velocity (Ge et al., 2018). These data are fast

moving and can originate from various sources, such

as social networks, unstructured data from different

devices or raw feeds from sensors (Ge and Dohnal,

2018). Thus, cloud environment has been used to ac-

commodate the Big Data and signiﬁcantly facilitate

the development of service-oriented computing for

data management (Park et al., 2021). While service-

oriented computing not only provides a better perfor-

mance of service offerings, perception, and quality of

service delivery (Mishra and Kumar, 2018), it also

offers a foundation for the development, execution,

composition, and integration of business processes

that are distributed across the cloud computing net-

work and accessible via standard interfaces and pro-

tocols (Serhani and Dssouli, 2010).

Cloud storage and service-oriented computing al-

lows enterprises to focus more on effective cross-

platform application data exchange (Khan et al.,

2019). However, when an enterprise intends to mi-

grate or build their data center with service-oriented

computing, there is lacking of guidelines on how to

design and build a Data as a Service (DaaS). Beyond

understanding the well-known cloud service models

such as Infrastructure as a Service (IaaS), Platform as

a Service (PaaS), and Software as a Service (SaaS),

it is unclear that how the as a Service features can be

applied to data, and compared to other cloud service

models, what the speciﬁc components for DaaS are.

Thus, DaaS can on the one hand beneﬁt from the gen-

eral as a Service features, on the other hand, data can

be managed as a service to potentially enhance the

resilience and reusability for data-driven applications

(Rao and Nayak, 2019).

This paper therefore proposes a DaaS architecture

that can guide users to deploy the data centre on the

cloud. It can enhance the data accessibility through

different channels, and eliminate the geographical and

scalability limitations (Rajesh et al., 2012). We will

provide a blueprint of how as a Service features can

be used for DaaS. The architecture is mainly focused

on speciﬁc components to construct the DaaS. The

proposed architecture is further validated by the phys-

ical implementations with real-world deployment.

2 AS A SERVICE

The term as a Service has been widely associated with

different service models while it sometimes can be

misleading or confusing to understand what actually

the as a service is (Duan et al., 2015). With respect to

the speciﬁc functionality, the as a Service on the cloud

278

Tóth, A. and Ge, M.

A Deployable Data as a Service Architecture for Enterprises.

DOI: 10.5220/0010470702780285

In Proceedings of the 6th International Conference on Internet of Things, Big Data and Security (IoTBDS 2021), pages 278-285

ISBN: 978-989-758-504-3

Table 1: Summary of features in as a service model.

IaaS PaaS SaaS DaaS

Monitoring computing resources software systems software applications data transactions

Metering load of resources system performance application usage transaction attributes

Security user’s responsibility combined responsibility

of user and provider

provider’s responsibility

to protect application

software

provider’s responsibility

to protect data assets

Tenancy Partitioning - system instances application instances data pools

Availability dependent on data cen-

ter(s)

dependent on infrastruc-

ture

dependent on infrastruc-

ture or platform

dependent on infrastruc-

ture or platform

Scalability on-deman resources allo-

cation/deallocation

on-deman resources allo-

cation/deallocation

on-deman resources allo-

cation/deallocation

on-deman resources allo-

cation/deallocation

Fault Tolerance task resubmission, stor-

age backup, etc.

system replicas, load

balancing, etc.

application replicas, or-

chestration, etc.

data protection, data reg-

ulation, etc.

Distribution - enabled enabled enabled

and normal services may have the same functionality

but can be organized in different forms. For example,

the difference between software and software as a ser-

vice does not depend on the functionality but on how

the functionality is provided. Implementing the as a

Service method requires to meet a set of requirements

that guarantee the functionality provision (Moorthy

and Pabitha, 2019). Given that as a Service is con-

sidered as a subset of cloud service model, there are

features that are inherited, meanwhile there are some

speciﬁc features for data management. By reviewing

the common features from IaaS, PaaS and Saas. We

have derived the following features for as a Service.

• Monitoring - it covers the operations and actions

connected to the provided service that are later to

be observed, checked, and processed for meter-

ing. Monitoring results are assets, logs, and audit,

which grant repudiation and accountability char-

acteristics to the service itself (Bouasker et al.,

2020). These monitoring data are considered as

a trusted source of recorded events, operations,

and transactions that are substantially signiﬁcant

for further calculations such as billing for the used

resources.

• Metering - it regulates the resource usage analy-

sis. This analysis takes into account the resource

restrictions concerning the rules in the service

level agreement (Narayan et al., 2017). The ser-

vice provider may deliver different service qual-

ities to the customers e.g. lower price for less

resources, higher price for unlimited operations.

The customers should be familiar with the terms

of service and these limitations. Metering can be

observed as a special type of monitoring that is

required for business and marketing purposes.

• Security - It is a fundamental feature in the cloud

computing. Each interaction from outside or

within the service represents a potential security

risk that might lead to exploitation of a security

vulnerability (Mthunzi et al., 2020). The data se-

curity is critical and it has been moved to the pri-

mary security stipulation. as a Service introduced

dozens of challenging security problems that re-

sulted in a continuous security incident monitor-

ing, reporting, and resolving.

• Tenancy Partitioning (Multitenancy) - It en-

sures that every tenant (user) is isolated from each

other and prevents violations between them while

each user is using a single instance of the appli-

cation (Aljahdali et al., 2014). Each tenant has

its own operational environment and he is able to

work in the space excluding the ability to use oth-

ers operational environment. For instance, the ap-

plication database can be conﬁgured to serve just

for a single tenant, which would prevent storing

multiple users’ data in the same database (Liu,

2010). There may be an exception when the ten-

ants share spaces consciously with respect to their

needs such as collaboration. Tenancy partition-

ing requires to implement access control and user

isolation. From the aspect of SaaS application,

this feature has been considered as highly impor-

tant (Aleem et al., 2019).

• Availability - It can be seen as a warranty from

a cloud perspective. It is usually expected that

the as a Service derivations meet criteria of high

availability and reliability (Li et al., 2020). As as

a Service is adopted on the basis of cloud comput-

ing, the availability is guaranteed partially by the

cloud, which means that the cloud hosted service

itself should be also resilient to errors and fault-

tolerant.

• Scalability - This is the ability to increase or de-

crease IT resources regarding the actual demand

A Deployable Data as a Service Architecture for Enterprises

279

(Bellavista et al., 2017). Scalability can be dif-

ferentiated into two types - horizontal and vertical

scalability. In vertical scaling we change the com-

puting resources (such as CPU, storage, RAM,

etc.) on an existing machine while in horizon-

tal scaling we change the pool of resources by

adding a new or removing an existing computing

machine. The resource addition is called scaling

up and resource removal is called scaling down.

Horizontal scalability mostly includes clustering,

load balancing fencing and other tools as well as

techniques.

• Fault Tolerance - It is the capability of a sys-

tem that can operate continuously despite errors.

Most of the applications hosted on a cloud require

a high level of fault tolerance that is solved by

transparent replication of applications (Moham-

madian et al., 2020). We distinguish two types of

fault tolerances - proactive and reactive fault toler-

ance. Proactive fault tolerance means that the ap-

plication is preemptive by nature by using preven-

tion techniques before the appearance of the error,

such as precautionary migrations. On the other

hand, reactive fault tolerance reduces the conse-

quences of the already occurred errors using prac-

tices for instance backups and rollbacks.

• Distribution - A service is usually not a stan-

dalone service that provides all the functionalities

(Suresh and Varatharajan, 2019). The service may

rely on other integrated services used for speciﬁc

operations and additional resources. The speciﬁc

responsibilities can be shifted to other services in

order to improve efﬁciency of the business and in-

crease the service quality.

Based on the derived features above, we have

compared the IaaS, PaaS, SaaS and DaaS, where how

the as a Service features can be used on the data is

highlighted in Table 1.

The service computing and its as a Service model

have a beneﬁcial impact on IT related business. The

increasing amount of divergent as a Service conﬁrms

the usefulness and effectiveness (Prasad et al., 2014).

When the enterprises intend to use additional ser-

vice features, these services can be implemented on-

demand. This can inﬂuence the costs per resources or

usage. The underlying architecture of as as Service

model facilitates the extensibility and adaptive scal-

ing. It reduces processing time and costs for enter-

prises. Furthermore, the ubiquitous access makes the

service widely accessible to a wide range of users.

3 DATA AS A SERVICE

ARCHITECTURE

An architecture provides foundation components that

reveals the implementation requirements. Thus, the

DaaS architecture can be considered as a general tem-

plate, which includes fundamental concepts and pat-

terns from various disciplines to facilitate the adop-

tion and implementation of DaaS. The DaaS architec-

ture deﬁnes an architecture at a conceptual level while

the physical levels are designed by the enterprise it-

self based on their available technical resources. In

order to construct the DaaS, data architecture design

needs to conforms with the service design. First of

all, the architecture needs to ensure that heteroge-

neous distributed data are provided from its author-

itative source, which requires a data acquisition com-

ponent. Next component processes the acquired data,

while it applies the necessary data policies, rules, and

regulation. A separate component is allocated for se-

curity because there is nothing more vulnerable than

the data itself. Afterward, as the data are processed,

the data (including meta data) are delivered to the re-

quester in a speciﬁc format. There is also a particu-

lar need for service management, because the DaaS

should meet the service-level agreement criteria and

service quality expectations, which are committed be-

tween a provider and a client.

We propose the DaaS architecture that includes

the following components. All those components are

organized in Figure 1.

• Data Acquisition - it formulates an interconnec-

tion layer between the system and each registered

data sources.

• Data Management - it deﬁnes data policies that

guarantee consistent data manipulation manners.

A set of standardized techniques are required in

such an environment, which manage interoper-

ability between diverse and heterogeneous data

ﬂows.

• DaaS Engine - this is the core of the DaaS that

processes the data request accordingly to the data

policies, rules and further data regulations.

• Data Regulations - applies the rules and laws from

the area of authority and legislation. The DaaS

needs to comply with the established legal entitle-

ments, which are formally deﬁned.

• Security - this is responsible for the continuous

protection of DaaS data assets. Cyberattacks are

still evolving by the time using more sophisticated

methods than before. It is necessary to protect and

ensure high security level for each component of

DaaS during the entire operation of the service.

IoTBDS 2021 - 6th International Conference on Internet of Things, Big Data and Security

280

• Service Management - it ensures that DaaS will

satisfy all the customer expectations and re-

quirements and maximize the business value and

progress efﬁciency of the service. A service

needs to continuously improve its capabilities

to assure achieving a high customer satisfaction

even though the customer data requirements may

change from time to time.

3.1 Data Acquisition

The integration of heterogeneous data on distributed

data sources introduces two principal challenges -

integration of different data sources into one data

model, and performing manipulating operations on

the data (Khan et al., 2019). Data acquisition rep-

resents an interface layer between data sources and

the DaaS. This component of DaaS is the foundation

layer in DaaS, which ensures proper data acquired

from physical data sources.

A data source is an self-organized entity that sup-

plies the data. It can be a database, IoT devices, cloud

storage, data warehouse, distributed ﬁle system, etc;

which might support further operations beyond this

data (Weik, 2001). Each data source can have its own

interface that implements an access protocol. To in-

tegrate the data, the data sources need to support one

commonly shared data mapping protocol in order to

deal with the problem of data divergence with differ-

ent communication protocols. Nowadays, there are

some promising technologies such as message broker

that provides a translation from sender’s formal mes-

saging protocol into the receiver’s formal messaging

protocol, which mitigates the complexity of such a

problem.

Before data operations, data discovery is con-

ducted to select the corresponding data sources that

includes the necessary data. The data discovery is

responsible to select and prepare in advance the re-

quired data sources that includes pertinent data, which

will be required for the later search execution. The

output of the data discovery phase is a particular list

of data sources, which also includes supplementary

metainformation (Terzo et al., 2013).

Data discovery is followed by data mapping that

maps proprietary source data into a standardized data

structure. The data mapping transforms the data set

into an understandable format and ensures consis-

tency for later processes. The format represents a pre-

deﬁned structure - a schema.

As certain data might be separated and stored in

different data sources, data need be aggregated cor-

respondingly. Data aggregation is a process of data

transformation with the intent to prepare combined

data for subsequent data processing. The data aggre-

gation can be described as a merge of multiple data

schema from the data mapping into one entity - more

complex schema that will be later processed.

As a fundamental component of DaaS, data acqui-

sition interconnects the whole system with each data

source that is managed by this system. Data acquisi-

tion solves the data access problems in DaaS such as

disparate data formats, access to the data by different

interfaces, and data source selection. This component

is closely connected to the DaaS engine, which inter-

acts with the data through this component.

3.2 Data Management

Data management includes all the data operations that

are more focused on the data values. As Gartner

deﬁnes the data management as “Data management

consists of the practices, architectural techniques, and

tools for achieving consistent access to and delivery of

data across the spectrum of data subject areas.”(gar, ),

DaaS requires to have a component that deals with the

data representation, data coordination, data manipula-

tion as well as data cleaning.

To understand data and provide more comprehen-

sive insights from the data, it is essential to conduct

data analysis. Data analysis in the context of DaaS

stands for knowledge and pattern discovering from

available data sets by using different approaches and

techniques such as data mining. The acquired knowl-

edge can lead to enrichment of the basic data set, for

instance, data augmentation can be implemented. The

results from data analysis can directly support the de-

cision making.

Data control is responsible for procedures related

to data permissions, which include access, privileges,

ownership, etc. Delivering data demands for data pro-

tection from source to destination as well as its value

protection from misuse. Data protection from the le-

gal point of view will be described in Section 3.4. On

the other hand, the data control in data managements

protects the data during its delivery that is followed by

its usage. Data control imposes the data source rules

and ingestion deﬁned by its owner.

Especially in a Big Data context, the data can be

inconsistent or with represented in different scales.

Therefore, data standardization is to deal with data

inconsistency problem. Different data may repre-

sents different obstacles and difﬁculties associated to

tasks. Data standardization deﬁnes an uniform for-

mat, which facilitates the work associated with such

inconsistent data.

Data governance deals with the permissions and

speciﬁc criteria granted by the data owner to their data

A Deployable Data as a Service Architecture for Enterprises

281

DaaS

DaaS Engine

Data Acquisition

Data Regulations

Interface

Registry

Data Management

Data Mapping

Data Control

Data Analysis

Data Aggregation

Service Management

Service Quality SLA Compliance

Request Processing

Data Discovery

Privacy

Data Standardization

Data Governance

Data Quality

Security

Legal

3rd Party

Service Integration

Figure 1: Conceptual architecture of DaaS.

sources. Some of the sensitive data may be prohib-

ited, allowed, or in a speciﬁc case provided under de-

termined restrictions and limitations such as partial

censorship. Overall, data governance regulates the

valuable information of data.

Data quality can directly affect the efﬁciency and

effectiveness of organizations and businesses (Ge and

Lewoniewski, 2020). DaaS is based upon data deliv-

ery that is associated with data risks and threats. To

minimize and eliminate data quality problems, it is

essential to focus on data quality improvement. Data

quality in DaaS provides a qualitative and quantitative

cleaning process for the available data, which leads to

the improvement of the overall data analysis provided

by DaaS.

3.3 DaaS Engine

DaaS engine represents the core component of the

DaaS. It is a system that implements the DaaS poli-

cies and regulates each data ﬂows. This component

in the DaaS is mainly focused on operating efﬁciently

data ﬂows between the data source and data consumer

that are managed and controlled in relation to the de-

ﬁned policies.

The DaaS engine consists of two fundamental

parts: the DaaS system and the DaaS registry. The

system is responsible for the interoperability - infor-

mation exchange between two end devices. The data

consumer can access the DaaS via an interface. The

interface requests are processed by the DaaS system.

The DaaS registry represents a catalog that includes

meta information about the data, which are used for

incoming requests in the DaaS. The registry contains

information that includes the data relation, location,

accessibility, etc., which facilitate to understand the

data. The DaaS system then provides authoritative

source data to the data consumer in a standardized

structured data that can be understood on a success-

fully processed request. The DaaS engine might be

integrated with other third party services, which en-

rich the feature set of the DaaS. The integration is

controlled and accompanied by the conditions of ser-

vice usage that entail particular responsibilities.

3.4 Data Regulations

Data regulation is responsible for deﬁnition of data

protecting policies in the DaaS. This component of

the DaaS is a connecting point between the cloud ser-

vice and the area of authority and legislation, which

IoTBDS 2021 - 6th International Conference on Internet of Things, Big Data and Security

282

inﬂuence the data and its processing. By implement-

ing a proper data regulation, the DaaS can minimize

the data privacy and data protection related issues in

the cloud environment such as illegitimate data dis-

semination.

With the rapid development of IT and globaliza-

tion, Big Data introduces new data challenges, and

results in newly deﬁned policy to protect the data and

its further processing in IT. DaaS needs to act in ac-

cordance with the established legal and privacy rights.

The private data stands for every piece of information

that is considered as personal. Data protection rights

have a high effect on how the service is provided by

data providers and how the data consumers use the

subscribed service. Data regulation is responsible for

the formal deﬁnition of each use case necessary from

the legal point of view that has to be implemented into

DaaS.

Obligations derived from data protection and data

privacy can signiﬁcantly inﬂuence the DaaS. For ex-

ample, in Europe there are several rule enforcement

that deﬁnes these obligations such as GDPR. These

rule enforcements mandate the enterprise to adhere to

these obligations, which process or collect any type of

information related to its subscribers. The obligations

are inﬂuencing the data provider as well as to the data

consumer.

3.5 Security

With the widespread presence of cyberattacks and se-

curity breaches, the data security plays a signiﬁcant

role in IT management (Michener, 2020). Data se-

curity protects data assets from undesired disclosure,

modiﬁcation, exploitation, and destruction, whether

accidental or intentional. DaaS is particularly vul-

nerable, and may involve sensitive and conﬁdential

data. Securing DaaS system requires to protect the

data ﬂow from source to destination.

DaaS is a cloud service model and most of its

components such as data sources are integrated via

network connection. Thus, it requires different secu-

rity operations, for instance, active data in the tran-

sitions, data in an established communication chan-

nel between endpoints, or persistent data stored in

the cloud data storage. Securing such a complex and

constantly changing environment needs to provision

a continuously innovated scheme in each application

area of such a DaaS system to ensure the security.

Moreover, data fraud is becoming more prevalent.

Enterprises may use different approaches such as wa-

termarking or censorship to protect the sensitive data

to not being exposed, copied, and stolen. However,

it will devalue the expected information. This raises

the issue of trust from both business and customer.

For example, when the data are delivered to users, the

consuming side also needs to have the protection re-

sponsibility.

3.6 Service Management

The increase of data services leads to a growing de-

mand for service management that is to control to

the overall service provision. Enterprise activities that

are performed to manage, control, deliver and operate

data services offered to customers can be managed ac-

cordingly. The service management has a crucial role

for the DaaS provider, as it ensures the customers’ ex-

pectations for the provided service.

as a Service instances are associated with service

delivery, as the name suggests, which delivers value

to the customers/subscribers. Service management

is fundamental because it drives the associated areas

such as service strategy, service goals (including ob-

jectives), service metrics (e.g., key performance indi-

cator), etc. The key role rests in a clear, well-deﬁned

roadmap that will reduce costs and improve the efﬁ-

ciency of the provided service.

The service management in DaaS also includes

service delivery and IT service management. Trans-

formation shift requires ﬁrstly to adopt a service de-

livery model, which deﬁnes roadmap and establish

service blueprint for data services. It is required to

have a clearly deﬁned goals, strategy, mission and vi-

sion of the data service. Service delivery deals with

the service subscribers respectfully. Furthermore, ser-

vice management also tackles the DaaS operation

management and continuous improvement, which al-

low productive and efﬁcient data delivery service.

4 IMPLEMENTATION

A physical architecture represents a deployable archi-

tecture instance that describes operational character-

istics of the designed concept. In the physical in-

stance, we show a physical architecture in Figure 2.

It is composed of real-world tools and services. This

deployment consists of existing technologies and ser-

vices that are optimized for the desired purpose in this

instance.

The enterprise needs to obtain the requirements

for DaaS provisioning such as cloud platform, cloud

storage, etc. The platform ensures essential features

in the cloud. In our deployment, we have chosen the

Google Cloud Platform as a PaaS provider from the

available alternatives such as Amazon Web Service,

Microsoft Azure, Oracle Cloud, IBM Cloud etc.

A Deployable Data as a Service Architecture for Enterprises

283

Figure 2: Deployable physical architecture of DaaS.

The usage of enterprise service bus (ESB) pro-

vides a efﬁcient method for multiple divergent sys-

tem integration and management. Nowadays man-

ufactures released and announced message-oriented

middleware appliances such as Apache ServiceMix

that can realize the ESB communication system. The

ESB is responsible for information exchange in real

time between subsystems of DaaS. The usage ESB

has the advantage to build a loosely coupled system.

Furthermore, DaaS Engine delivers the requested

data to consumers or applications. The engine can be

implemented in various ways, we have chosen Flask

web framework and Gunicorn as our WSGI

while

both of them are implemented in a Python language,

which is a general object oriented imperative pro-

gramming language. The Gunicorn implements two

roles - server side (requests handling) and application

side (request delegation). The Gunicorn invokes the

corresponding python callable on the incoming re-

quest. In our deployment, the Flask is the callables

and responsible for processing requests in the data

management. Also, using Flask, there is a possibil-

ity to implement the data-related validation rules and

policies in a custom way. Flask is a framework writ-

ten in Python and supports the object mapping (in our

case, object document mapping) of database items.

The speciﬁc rules can be isolated and separated from

the application logic and stored in a persistent data

storage, for instance, in a database - MongoDB.

One of frequently used approaches to integrate

data to other application is via application protocols

such as HTTP or HTTPS. The access to the DaaS is

enabled through a RESTful API, which is based on

OpenAPI standard and implemented via Connexion

that maps each endpoint to the Python callables.

Third-party service integration is to enrich the

A framework composed of Apache ActiveMQ,

ApacheCamel, and Apache CXF, and Apache Karaf.

Web Service Interface Gateway

provided basic feature set of the DaaS. In most cases,

the external service is added as an encapsulated ser-

vice entity that is managed with respect to the in-

ternal policies and terms of service. The integration

process will connect the speciﬁc service by mapping

the service interface to the message-oriented middle-

ware. The integration result is a connected service,

which can exchange data with other components of

the DaaS.

Current cloud storage providers offer a wide range

of storage services such as backup and versioning.

There is a large amount of options on how to im-

plement cloud storage within a DaaS. Regarding the

price pool and available resources, the cloud storage

may be an integrated service provided by a cloud stor-

age provider or a custom cloud storage. For large

scale data volume, using existing cloud storage may

be a cheaper option to the enterprise in contrast to

re-implement the cloud storage. In our example, we

use an already existing cloud storage - Google Cloud

Storage, that is used for storage purposes of the DaaS.

5 CONCLUSIONS

In this paper, we have proposed a Data as a Service ar-

chitecture to guide users to build, migrate and deploy

data management on the cloud. The features of the as

a Service are derived from the IaaS, PaaS and SaaS.

Those features are then applied as a foundational set-

ting for DaaS. The DaaS architecture further deal with

speciﬁc data functionalities along with data ﬂow in

the cloud computing. In order to validate the pro-

posed DaaS architecture, we have demonstrated how

to instantiate the DaaS architecture to a deployable

physical architecture. During the implementation, we

have reported the lessons learned from the DaaS im-

plementation. It can be seen that the proposed DaaS

can be applied in real-world deployment and can sig-

niﬁcantly help the enterprises to build their DaaS.

As future works, the proposed implementation can

be further developed and enriched by performance

benchmarks of the deployed instance including dif-

ferent tools and technologies.

REFERENCES

Gartner glossary. https://www.gartner.com/

en/information-technology/glossary/

dmi-data-management-and-integration.

Aleem, S., Ahmed, F., Batool, R., and Khattak, A. (2019).

Empirical investigation of key factors for saas archi-

tecture dimension. IEEE Transactions on Cloud Com-

puting, pages 1–1.

IoTBDS 2021 - 6th International Conference on Internet of Things, Big Data and Security

284

Aljahdali, H., Albatli, A., Garraghan, P., Townend, P., Lau,

L., and Xu, J. (2014). In 8th IEEE International

Symposium on Service Oriented System Engineering,

pages 344–351. IEEE Computer Society.

Bellavista, P., Corradi, A., and Zanni, A. (2017). Integrat-

ing mobile internet of things and cloud computing to-

wards scalability: lessons learned from existing fog

computing architectures and solutions. Int. J. Cloud

Comput., 6(4):393–406.

Bouasker, T., Langar, M., and Robbana, R. (2020). Qos

monitor as a service. Softw. Qual. J., 28(3):1279–

1301.

Duan, Y., Fu, G., Zhou, N., Sun, X., Narendra, N. C., and

Hu, B. (2015). Everything as a service (xaas) on the

cloud: Origins, current and future trends. In 2015

IEEE 8th International Conference on Cloud Comput-

ing, pages 621–628.

Ge, M., Bangui, H., and Buhnova, B. (2018). Big data for

internet of things: A survey. Future Gener. Comput.

Syst., 87:601–614.

Ge, M. and Dohnal, V. (2018). Quality management in big

data. Informatics, 5(2):19.

Ge, M. and Lewoniewski, W. (2020). Developing the qual-

ity model for collaborative open data. Procedia Com-

puter Science, 176:1883–1892.

Khan, F. A., ur Rehman, M., Khalid, A., Ali, M., Imran,

M., Nawaz, M., and Rahman, A. (2019). An intel-

ligent data service framework for heterogeneous data

sources. Journal of Grid Computing, 17(3):577–589.

Li, C., Song, M., Zhang, M., and Luo, Y. (2020). Effec-

tive replica management for improving reliability and

availability in edge-cloud computing environment. J.

Parallel Distributed Comput., 143:107–128.

Liu, G. (2010). Research on independent saas platform. In

2010 2nd IEEE International Conference on Informa-

tion Management and Engineering, pages 110–113.

Michener, J. R. (2020). Security issues with functions as a

service. IT Professional, 22(5):24–31.

Mishra, S. and Kumar, C. (2018). Effort estimation for

service-oriented computing environments. Comput.

Informatics, 37(3):553–580.

Mohammadian, V., Navimipour, N. J., Hosseinzadeh, M.,

and Darwesh, A. (2020). Comprehensive and sys-

tematic study on the fault tolerance architectures

in cloud computing. J. Circuits Syst. Comput.,

29(15):2050240:1–2050240:40.

Moorthy, R. S. and Pabitha, P. (2019). Optimal provisioning

and scheduling of analytics as a service in cloud com-

puting. Trans. Emerg. Telecommun. Technol., 30(9).

Mthunzi, S. N., Benkhelifa, E., Bosakowski, T., Guegan,

C. G., and Barhamgi, M. (2020). Cloud computing se-

curity taxonomy: From an atomistic to a holistic view.

Future Gener. Comput. Syst., 107:620–644.

Narayan, A., Pillai, P. S., Prasad, A. S., and Rao, S. (2017).

Resource procurement, allocation, metering, and pric-

ing in cloud computing. In Research Advances in

Cloud Computing, pages 141–186.

Park, J. H., Younas, M., Arabnia, H. R., and Chilamkurti,

N. K. (2021). Emerging ICT applications and services

- big data, iot, and cloud computing. Int. J. Commun.

Syst., 34(2).

Prasad, A., Green, P. F., and Heales, J. (2014). On gover-

nance structures for the cloud computing services and

assessing their effectiveness. Int. J. Account. Inf. Syst.,

15(4):335–356.

Rajesh, S., Swapna, S., and Reddy, P. (2012). Data as a

service (daas) in cloud computing. Global journal of

computer science and technology.

Rao, K. R. and Nayak, A. (2019). Data residency as a ser-

vice: a secure mechanism for storing data in the cloud.

Int. J. Embed. Syst., 11(4):397–418.

Serhani, M. A. and Dssouli, R. (2010). Case study: Master

of science in service computing (msc sc). In 2010 6th

World Congress on Services, pages 80–83.

Suresh, A. and Varatharajan, R. (2019). Competent re-

source provisioning and distribution techniques for

cloud computing environment. Clust. Comput.,

22(5):11039–11046.

Terzo, O., Ruiu, P., Bucci, E., and Xhafa, F. (2013). Data

as a service (daas) for sharing and processing of large

data collections in the cloud. In Seventh International

Conference on Complex, Intelligent, and Software In-

tensive Systems, pages 475–480.

Weik, M. H. (2001). data source, pages 358–358. Springer

US, Boston, MA.

A Deployable Data as a Service Architecture for Enterprises

285