Cloud-RA: A Reference Architecture for Cloud Based

Information Systems

Jalal Kiswani, Sergiu M. Dascalu and Frederick C. Harris, Jr

Department of Computer Science and Engineering, University of Nevada, Reno, 1664 N. Virginia Street, Reno, NV, U.S.A.

Keywords:

Software Reference Architecture, Cloud Computing, Cloud Applications, Software as a Service,

Microservices Architecture.

Abstract:

Software architecture is an essential phase of the software development process, as it signiﬁcantly increases the

success rate of software projects and enables achieving their quality attributes and goals. However, implement-

ing software architecture is not a straightforward process, and requires specialized expertise and knowledge

-in both domain and technology- to achieve its requirements. To overcome this complexity, many tools have

been developed to make the architecture process systemic, predictable and repeatable. These tools include

architectural styles, architectural patterns, and reference architectures. In fact, these tools encourage sharing

of experience and reducing the architecture process cost. In addition, tools such as reference architecture can

make non-expert architects and developers start with ready-made architecture templates ”as is,” or with mini-

mal customization. On the other hand, cloud computing is everywhere, and many applications are developed

as cloud applications in what is called Software as a Service delivery model. In this paper, we propose Cloud-

RA, a reference architecture for developing cloud-based multi-tenant information systems. In particular, it

includes the problem, motivation, and proposed architecture. We hope this proposed work can be the bases for

future cloud application reference architectures.

1 INTRODUCTION

Cloud Computing is one of the hottest trends in Infor-

mation Technology nowadays (Armbrust et al., 2009).

It changed how organizations from different domains

perform their business (Economist, 2008). Cloud

Computing service providers offer services in three

different delivery models: Infrastructure as a Service

(IaaS), Platform as a Service (PaaS), and Software as

a Service (Mell et al., 2011).

SaaS is based on developing software systems that

can be used by different tenants (i.e., customers) at

the same time. In fact, designing applications in a

way that enables all customers to share the same in-

stance of the deployed application is the preferred ap-

proach of developing cloud applications. This ap-

proach is preferred because it enables full utiliza-

tion of cloud resources and more efﬁcient scalabil-

ity. The approach of having all customers using the

same instance is called Single-Instance Multi-Tenant

approach (SIMT) (Guo et al., 2007).

Designing and developing SIMT applications is more

complicated than traditional monolithic applications

for many reasons. In particular, multi-tenant support,

billing, and monitoring are not easy to achieve. In

addition, and with high-competition in the startups’

ecosystem, the time to market and faster response to

changes are required to achieve more proﬁts and en-

sure more business sustainability. For these and many

other reasons, a new architectural style was devel-

oped: Microservices Architecture (MsA) (Fowler and

Lewis, 2018a).

MsA is based on having applications developed as

separate loosely-coupled services (i.e., components)

that are independently deployable, and able to com-

municate with other services over lightweight proto-

cols such as HTTP. In addition, every service man-

ages its own data storage; in fact, no direct access

to other services data-store is allowed (Fowler and

Lewis, 2018a). Also, MsA services are designed

around business capabilities (Richardson, 2018). Fur-

thermore, they can be developed in any technology

(e.g., Java, Python, C++), and can be added and re-

moved at any time without affecting the whole system

functionality (Killalea, 2016).

Designing software applications -and more speciﬁ-

cally cloud applications- based on MsA can exploit

the full beneﬁts of Cloud Computing of elasticity and

Kiswani, J., Dascalu, S. and Jr, F.

Cloud-RA: A Reference Architecture for Cloud Based Information Systems.

DOI: 10.5220/0006863608490854

In Proceedings of the 13th International Conference on Software Technologies (ICSOFT 2018), pages 849-854

ISBN: 978-989-758-320-9

849

global software engineering (Beecham et al., 2014).

Even though MsA has many advantages, it is not a

free lunch. Primary concerns are design and devel-

opment complexity, long learning curve, and lack of

expertise and tools (Richardson, 2018).

On the other hand, Information Systems (IS) is a par-

ticular category of software applications. IS enables

controlled access to a broad base of shared informa-

tion, such as library management systems and patient

record systems (Sommerville, 2015). In such appli-

cations, the relational database is the most commonly

used engine. However, due to the vast number of

database tables in these applications, having user in-

terface views for managing most of these tables is

mandatory, which includes Create, Read, Update, and

Delete (CRUD). Developing these functionalities in-

creases the development cost and time, and introduces

other risks. However, utilizing a metadata-driven ap-

proach for building such application can reduce these

risks dramatically. In fact, it can enable higher quality

and consistency (Kiswani et al., 2017).

Consequently, developing multi-tenant cloud-based

information systems by utilizing MsA is challenging

and requires specialized expertise. Thus, having an

approach for enabling a more efﬁcient way of design-

ing and developing such applications may reduce the

development time and cost, and be beneﬁcial for both

researchers and practitioners.

Many tools in the software architecture discipline

were developed to make it predictable, systematic,

and repeatable. Architectural Styles, Architectural

Patterns, and Reference Architectures are examples

of these tools (Len Bass, 2012).

Architectural styles and architectural patterns solve

partial problems of software systems. Client-Server,

Publish-Subscribe, and Layered Architecture are ex-

amples of architectural styles. While Model-View-

Controller (MVC) and State-Logic-Display (i.e.,

Three-Tier) are examples of architectural patterns

(Taylor et al., 2010).

In the other hand, in the reference architecture ap-

proach, a complete architecture is used for speciﬁc

types of solutions and domains. An example is the

Microsoft Industry Reference Architecture for Bank-

ing (Microsoft, 2012). Furthermore, lists of common

reference architecture are documented and explained

in literature (Humberto Cervantes, 2016).

For the preceding reasons, we propose Cloud-RA ref-

erence architecture. In particular, Cloud-RA is a tem-

plate architecture that can be used to build cloud in-

formation systems based on MsA. Furthermore, it in-

cludes multi-tenant support, billing, and monitoring.

Also, it includes cross-cutting services such as se-

curity, auditing, and logging. Moreover, it includes

a metadata-driven approach that can reduce develop-

ment time and increase software overall quality.

This paper is organized into 4 sections. This sec-

tion covers the introduction. Then, it is followed by

Section 2, which includes the background and related

work. Section 3 includes the proposed work. Finally,

Section 4 concludes the paper and identiﬁes several

directions of future work.

Figure 1: Examples of monolithic software systems.

SE-CLOUD 2018 - Special Session on Software Engineering for Service and Cloud Computing

850

2 BACKGROUND

Organizations from all domains implement software

applications with different scales for both internal and

external use. Governments implement software sys-

tems to enable internal administration management

with thousands of users. At the same time, they might

implement an online software system to enable smart-

government online services for citizens and residents.

Another example is that of a scientiﬁc community

may implement a group-scoped system for scientiﬁc

Big Data management while enabling visualization

features for external entities.

The monolithic approach of software development

was the dominant model, in particular, building soft-

ware applications as a single deployable unit (Fowler

and Lewis, 2018a). In the past, this approach was

practical since software size was relatively small and

consisted of a low number of software components

and functionality. In addition, this approach was com-

mon due to its convenience and ease of development.

Moreover, the systems were only used by a limited

number of users, with relatively long-term and sta-

ble requirements. Figure 1 shows some examples of

monolithic applications.

However, things have changed, with the Internet,

smartphones, Big Data, Cloud Computing, and the

startups’ ecosystem. Moreover, time to market and

response to new requirements became more critical.

Scalability has reached limits that were never pos-

sible and required before, with hundreds of millions

of concurrent users accessing the same service at the

same time. In addition, Internet of Things (IoT)

also affected this wave, which increased the num-

ber of devices accessing online services exponentially

(Economist, 2008).

With all these factors, the monolithic approach of

software applications development may be a severe

bottleneck. In fact, it may affect the organization’s

existence. The high-risks produced by adopting this

approach includes a high cost of implementation and

scalability, heavyweight testing and deployment pro-

cesses, and complexity of development. Moreover,

factors such as global software engineering (Ebert

et al., 2016), and the need to use different technol-

ogists in the same project (e.g., Angular for front-

end, Python for presentation tier, and Spring-Boot

for backend) increased the challenge of selecting the

monolithic approach, where the same technology is

the main theme for most of the applications. Fur-

thermore, various front-end technologies such as mo-

bile devices, browsers, desktop applications, and IoT

devices require lighter communication and full sepa-

ration between front-end and back-end technologies

(Laplante et al., 2008).

Even with the existence of techniques and concepts

such as Object Oriented Programming (OOP), de-

sign patterns (Gamma, 1995), reusable components

development, application frameworks, and software

product lines (Linda M. Northrop, 2012), building the

modern requirements of software systems is still chal-

lenging (Garlan, 2014).

Service Oriented Architecture (SOA) is an approach

of software architecture and development driven by

the academia and practitioners (Turner et al., 2003).

In particular, it is a software constructions model that

aims to change how the software systems are built. Its

central idea is based on separating an application into

smaller self-contained components, which are encap-

sulated, independently deployable, and communicate

over a standard protocol. The common used protocol

for SOA is Simple Object Access Protocol (SOAP)

based web services. However, SOA did not get high

traction because of its heavy-weight nature, in par-

ticular, communication, development and conﬁgura-

tion complexity, tools and technologies. Moreover,

the frequent use of SOA applications was for medium

to large-scale information systems that require a rela-

tional database as the central persistence mechanism

of application data. In fact, a uniﬁed instance of the

database was used, which limits the ﬂexibility, where

developers were constraints in terms of taking design

decisions, and created a bottleneck on architectural

and database design decisions.

Consequently, a new architectural style has evolved:

Microservices Architecture (MsA). In MsA, a soft-

ware application is built as small, lightweight,

reusable and self-deployable components that com-

municate over a lightweight protocol. In fact, these

components can be developed using any technology.

Moreover, every service has its business logic, stor-

age engine, and persistent mechanism. Therefore, de-

velopers have full decisions control over the design,

technology, and implementation of their services. The

common used protocol in MsA is the Representa-

tional State Transfer (REST). REST is a lightweight

version of SOAP web-services. In particular, it works

as a layer that operates directly over the HTTP proto-

col (Fowler and Lewis, 2018a).

The microservices architecture overcomes many is-

sues, and challenges over the monolithic and SOA

approaches. However, it requires a particular early

architecture and design, mainly for the services and

their interfaces. Notably, these decisions may affect

the overall success of the project. In fact, arguments

about whether to start monolithic or microservices are

increasing day after day (Tilkov, 2018) (Fowler and

Lewis, 2018b).

Cloud-RA: A Reference Architecture for Cloud Based Information Systems

851

Even though MsA can be implemented in traditional

on-site applications, its full advantage can directly be

gained if implemented in Cloud Applications. Cloud

Applications are the applications deployed on Cloud

Computing environments such as Infrastructure as a

Service (IaaS) or Platform as a Service (PaaS) and are

publicly accessible over the Internet by their intended

users (Gorelik, 2013). As explained earlier, adopting

MsA on the cloud is more efﬁcient than Monolithic or

SOA. In addition, implementing the quality attributes

of scalability, high availability, auditability, monitora-

bility, and traceability, will achieve the full beneﬁts of

Cloud Computing. In fact, including these quality at-

tributes along with MsA in an application is termed as

Cloud Native Applications (Balalaie et al., 2016).

3 PROPOSED REFERENCE

ARCHITECTURE

This section presents Cloud-RA a proposed reference

architecture for developing cloud-based multi-tenant

information systems. In particular, it describes the re-

quired layers and their components, and the motiva-

tion for them.

3.1 Cloud-RA Layers

To achieve higher separation of concerns and modu-

larity, the layered architectural style was used to de-

sign the higher level architecture of Cloud-RA. As

shown in Figure 2, Cloud-RA consists of 5 layers:

Figure 2: Layers of Cloud-RA.

1. Edge Services: Edge services are the front-line

of cloud applications that are exposed to applica-

tion clients. In fact, no access to the layers below

is allowed from outside applications. In modern

cloud software systems, client applications can

be web browsers, rich client applications, smart-

phones or smart-device clients. Consequently, en-

abling universal devices access over a lightweight

standard protocol is crucial.

2. Application Services: This layer includes the

application business logic services. These ser-

vices include the domain logic and the functional

requirements.

3. Metadata Services: This layer contains the

services required for dynamic generation of user

interface and data access functionality (Kiswani

et al., 2017).

4. Cloud Application Infrastructure Services:

This includes the services required to enable cloud

applications features such as multi-tenant support,

monitoring, and analytics.

5. Cross-cutting Services: This layer includes the

services that are required for all the other layers,

such as conﬁgurations, security, and utilities.

3.2 Cloud-RA Microservices

As presented in the previous section, Cloud-RA con-

sists of four horizontal layers and one cross-cutting

layer. All the internal components of these layers are

based on the Microservices Architecture. In particu-

lar, every service is self-deployable and possesses its

own data storage. The microservices of Cloud-RA

(Figure 3) are:

1. API Gateway: It is the core component of the

Edge-Services and acts as the front-end proxy that

accepts requests from external clients. Its main

goal is to direct the requests to the appropriate

service instances, where multiple instances of the

same service could be projected to achieve the

beneﬁt of light-weight horizontal scalability of

cloud-applications. In addition, it might include

front-end security features.

2. Multi-tenant: This microservice handles the

multi-tenant support of clients transparently. In

particular, it is designed to enable transparent sup-

port for multi-tenant, with the illusion of having a

separate dedicated deployment for every tenant.

3. Billing: Manages the usage and billing for each

customer. Examples of billing criteria may be

based on calculating the usage based on the num-

ber of transactions, the number of users, or data-

size.

4. Monitoring: This microservice monitors the ap-

plication activity for the cases of normal load,

overload, and attacks. This service may be useful

for evaluating the infrastructure needs, detection

of security attacks, or identify software design is-

sues and bottlenecks.

SE-CLOUD 2018 - Special Session on Software Engineering for Service and Cloud Computing

852

Figure 3: Microservices of Cloud-RA.

5. Analytics: Analyses the application data dynami-

cally. This can be used for dashboard-like func-

tionality in the application. It may include an

internal data warehouse that contains aggregated

data collected from other services. This data may

be then exposed for further analysis and visualiza-

tion.

6. Security: This microservice manages the security

aspects of the applications. It includes the authen-

tication and authorization functionality. In addi-

tion, it may include the required encryption, de-

cryption, and hashing implementations.

7. Auditing: Provides auditing support for business

level transactions. In fact, having automated au-

diting features is signiﬁcant in most applications

(e.g., ﬁnancial and governmental) for compliance

and internal and external auditing.

8. Caching: Acts as a shared in-memory stor-

age across all services for performance reasons.

Caching for common required data that have a

high percentage of access can dramatically in-

crease system performance and scalability. Also,

having a dedicated service for caching man-

agement makes it easier to enhance the soft-

ware design to upgrade from single-host cache to

distributed-caching technique.

9. Conﬁguration: Provides common conﬁguration

for all Cloud-RA components.

10. Utilities: Common utilities that can be called

from any service, such as getting server date, time,

and their formats.

4 CONCLUSIONS

The work presented in this paper is a proposed ref-

erence architecture for cloud-based information sys-

tems. This could be the base architecture that could

be modiﬁed according to the speciﬁc software project

requirements.

However, since the microservices architectural style

is relatively new, and the cloud computing trends are

still evolving, there will be a need for new reference

architectures to cover the new features and require-

ments.

In addition, reference architectures for other cate-

gories of cloud-based applications are essential. Ex-

amples of such categories are Big Data and IoT appli-

cations.

ACKNOWLEDGMENT

This material is based upon work supported by the

National Science Foundation under grant number

Cloud-RA: A Reference Architecture for Cloud Based Information Systems

853

IIA-1301726. Any opinions, ﬁndings, and conclu-

sions or recommendations expressed in this material

are those of the authors and do not necessarily reﬂect

the views of the National Science Foundation.

REFERENCES

Armbrust, M., Fox, A., Grifﬁth, R., Joseph, A. D., Katz,

R. H., Konwinski, A., Lee, G., Patterson, D. A.,

Rabkin, A., Stoica, I., et al. (2009). Above the clouds:

A berkeley view of cloud computing. Technical re-

port, Technical Report UCB/EECS-2009-28, EECS

Department, University of California, Berkeley.

Balalaie, A., Heydarnoori, A., and Jamshidi, P. (2016). Mi-

croservices architecture enables devops: Migration to

a cloud-native architecture. IEEE Software, 33(3):42–

52.

Beecham, S., O’Leary, P., Baker, S., Richardson, I., and

Noll, J. (2014). Making software engineering research

relevant. Computer, 47(4):80–83.

Ebert, C., Kuhrmann, M., and Prikladnicki, R. (2016).

Global software engineering: Evolution and trends.

In Global Software Engineering (ICGSE), 2016 IEEE

11th International Conference on, pages 144–153.

IEEE.

Economist (2008). Let it rise: A special report on corporate

IT. Special report, The Economist.

Fowler, M. and Lewis, J. (2018a). Microser-

vices: a deﬁnition of this new archi-

tectural term. retreived jan 2018, from

https://martinfowler.com/articles/microservices.html.

Fowler, M. and Lewis, J. (2018b). Microser-

vices resource guide. retreived jan 2018, from

https://martinfowler.com/microservices/.

Gamma, E. (1995). Design patterns: elements of reusable

object-oriented software. Pearson Education India.

Garlan, D. (2014). Software architecture: a travelogue. In

Proceedings of the on Future of Software Engineering,

pages 29–39. ACM.

Gorelik, E. (2013). Cloud computing models. PhD thesis,

Massachusetts Institute of Technology.

Guo, C. J., Sun, W., Huang, Y., Wang, Z. H., and Gao, B.

(2007). mework for native multi-tenancy application

development and management. In e-commerce Tech-

nology and the 4th IEEE International Conference on

Enterprise Computing, e-commerce, and E-Services,

2007. CEC/EEE 2007. The 9th IEEE International

Conference on, pages 551–558. IEEE.

Humberto Cervantes, . R. K. (2016). Designing Soft-

ware Architectures: A Practical Approach (SEI Se-

ries in Software Engineering). Addison-Wesley Pro-

fessional; 1st edition.

Killalea, T. (2016). The hidden dividends of microservices.

Communications of the ACM, 59(8):42–45.

Kiswani, J., Muhanna, M., and Qusef, A. (2017). Using

metadata in optimizing the design and development

of enterprise information systems. In Information and

Communication Systems (ICICS), 2017 8th Interna-

tional Conference on, pages 188–193. IEEE.

Laplante, P. A., Zhang, J., and Voas, J. (2008). What’s in

a Name? Distinguishing between SaaS and SOA. It

Professional, 10(3).

Len Bass, Paul Clements, . R. K. (2012). Software Ar-

chitecture in Practice (3rd Edition). Addison-Wesley

Professional.

Linda M. Northrop, Paul C. Clements Contributor Reed Lit-

tle, J. M. L. O. F. B. J. K. B. G. C. S. G. C. P. D. L. G.

J. R. W. K. J. (2012). A Framework for Software Prod-

uct Line Practice, Version 5.0. Software Engineering

Institute, Carnegie Mellon University.

Mell, P., Grance, T., et al. (2011). The NIST deﬁnition of

cloud computing.

Microsoft (2012). Microsoft Industry Reference Architec-

ture for Banking (MIRA-B).

Richardson, C. (2018). Microservices are not a silver bullet.

retreived jan 2018, from https://microservices.io.

Sommerville (2015). Software Engineering (10th Edition).

AddisonWesley.

Taylor, R. N., Medvidovic, N., and Dashofy, E. M. (2010).

Software Architecture: Foundations, Theory, and

Practice. Wiley.

Tilkov, S. (2018). Dont start with a monolith. retreived

jan 2018, from https://martinfowler.com/articles/dont-

start-monolith.html.

Turner, M., Budgen, D., and Brereton, P. (2003). Turning

software into a service. Computer, 36(10):38–44.

SE-CLOUD 2018 - Special Session on Software Engineering for Service and Cloud Computing

854