The Evolution of Cloud Computing Towards a Vendor Agnostic Market

Place Using the SKY CONTROL Framework

Henry-Norbert Cocos

, Christian Baun

and Martin Kappes

Frankfurt University of Applied Sciences, Nibelungeplatz 1, Frankfurt am Main D-60318, Germany

Keywords:

Sky Computing, Multi-Cloud, Vendor-Agnostic Framework, Workload Placement.

Abstract:

Multi-cloud environments offer beneﬁts like vendor diversiﬁcation and resilience but pose challenges such

as increased management complexity, lack of cost transparency, and compliance. This concept paper intro-

duces SKY CONTROL, a vendor-agnostic framework for small and medium-sized enterprises (SMEs). SKY

CONTROL integrates cost control and risk management into multi-cloud setups, providing static and dynamic

resource analyses, a cost calculator, and risk assessment tools. By leveraging the Sky Computing paradigm,

SKY CONTROL simpliﬁes resource orchestration and enhances security. This novel framework is the ﬁrst

implementation of the innovative Sky Computing concept. It aims to improve cost efﬁciency, regulatory com-

pliance, and strategic IT planning for SMEs, offering a uniﬁed approach to managing hybrid infrastructures.

1 INTRODUCTION

Cloud Computing, based on distributed systems (van

Steen and Tanenbaum, 2017), integrates virtualiza-

tion and modern web technologies to provide scalable

IT infrastructure, platforms, and applications as on-

demand services. The National Institute of Standards

and Technology (NIST) deﬁnes Cloud Computing by

ﬁve key properties (Mell et al., 2011):

• On-demand Self-Service: Automatic resource

provisioning.

• Broad Network Access: Accessibility via stan-

dard network interfaces.

• Resource Pooling: Shared, scalable resources.

• Rapid Elasticity: Dynamic scalability.

• Measured Service: Usage-based billing.

These features facilitate seamless cloud integra-

tion, reducing costs and effort. Figure 1 illustrates

the NIST model, covering service properties, models,

and deployment types.

Since the late 2000s, IT infrastructures have

shifted from on-premise to multi- and hybrid cloud

systems (Hong et al., 2019; Gundu et al., 2020).

https://orcid.org/0009-0001-7573-0361

https://orcid.org/0009-0004-9955-3752

https://orcid.org/0000-0002-8768-8359

By the mid-2010s, multi-cloud strategies, combin-

ing services from multiple providers, became com-

mon (Jamshidi et al., 2016). Beneﬁts include (Hong

et al., 2019; Georgios et al., 2021; Petcu, 2013):

• Reduced provider dependency

• Cost optimization

• Load balancing

• Business continuity through redundancy

• Free service selection

• Enhanced security via data diversiﬁcation

Figure 1: NIST Deﬁnition of Cloud Computing.

Multi-cloud strategies foster innovation by in-

tegrating diverse platforms. Hybrid models use

infrastructure-as-a-service (IaaS) to minimize admin-

istrative tasks and software-as-a-service (SaaS) for

more straightforward conﬁguration and integration.

Cocos, H.-N., Baun, C. and Kappes, M.

The Evolution of Cloud Computing Towards a Vendor Agnostic Market Place Using the SKY CONTROL Framework.

DOI: 10.5220/0013361200003950

In Proceedings of the 15th International Conference on Cloud Computing and Services Science (CLOSER 2025), pages 211-218

ISBN: 978-989-758-747-4; ISSN: 2184-5042

211

However, managing multi-cloud environments re-

mains complex, with interoperability a signiﬁcant

challenge. Sky Computing aims to simplify this by

abstracting cloud resources across providers, enabling

optimal service selection. Though still being re-

searched, it has the potential to enhance multi-cloud

adoption.

This paper explores Sky Computing and its role in

advancing Cloud Computing.

2 BACKGROUND AND RELATED

WORKS

Cloud computing has expanded signiﬁcantly in recent

years, leading to diverse public and private offerings.

Based on the hybrid cloud model, multi-cloud en-

vironments have become common, enabling compa-

nies to avoid vendor lock-in and adopt ﬂexible strate-

gies (Mulder, 2020). These environments distribute

services across multiple providers, offering economic

and organizational beneﬁts, such as ﬂexibility and in-

creased resource availability. However, challenges in-

clude service transparency, interoperability issues be-

tween APIs, and the complexity of managing diverse

cloud services (Ardagna, 2015; Barker et al., 2015).

Sky computing aims to address these challenges

by introducing an abstraction layer over CSP of-

ferings, enabling uniform service provisioning re-

gardless of location (Fortes, 2010; Monteiro et al.,

2014; Yang et al., 2023). The SkyPilot project (Wei

et al., 2021) at Berkeley is a pioneering effort in

this domain. It features an intercloud broker opti-

mized for machine learning workloads across multi-

ple providers. Despite its innovative approach, SkyP-

ilot remains experimental, with limited cloud service

diversity and a focus on IaaS environments.

This raises questions about applying Sky Comput-

ing to diverse workloads and its potential to general-

ize across services. We explore Sky Computing op-

portunities in section 3 and critically assess its de-

mands on the current cloud computing landscape in

section 4.

3 SKY COMPUTING

Sky Computing is a new conceptual vision for Cloud

Computing that introduces an intercloud broker for

the mediation of services. The Sky Computing con-

cept (Stoica and Shenker, 2021) proposes an addi-

tional abstraction layer between the cloud services

from the providers (Amazon Web Services, Google

Cloud Platform, etc.) and the workloads of the end

users. Sky Computing aims to complete the abstrac-

tion of cloud resources from different providers so

that applications and users can access these resources

without worrying about where the resources or ser-

vices are located in the individual clouds. For this

reason, the term ”cloud of clouds” is also used, as the

additional abstraction of resources leads to creating

a uniform and interoperable cloud and thus includes

several individual cloud providers. This abstraction

layer is called an intercloud broker (Yang et al., 2023).

One idea for solving the problem of heterogene-

ity in cloud services and their respective APIs would

be an effort for standardization across cloud vendors.

As Chasins et al. (Chasins et al., 2022) describe in

their work, introducing standardization in the ﬁeld of

Cloud Computing poses risks to future developments

since it would slow down or even stop innovation by

locking the ecosystems into a set of interfaces that

may not be suitable for future applications. Many

examples in the past have shown that standardiza-

tion efforts have failed. One example is the introduc-

tion of the Common Object Request Broker Archi-

tecture (CORBA) by the Object Management Group

(OMG) (Siegel, 1999). In an attempt to standardize

the communication of different distributed objects on

various operating systems, this effort posed strict lim-

itations on the implementation of distributed systems

since it was designed as an interfacing technology

for other programming languages and imposed harsh

conventions on the usage of the standard. There-

fore, the standard adoption in the industry was poorly

met, and it was doomed to become a niche technol-

ogy. This example demonstrates that standardization

is not the solution to the interoperability problem. Sky

Computing aims to alleviate this challenge by intro-

ducing the intercloud broker, which acts as a common

platform for different cloud services. It focuses on the

workloads and types and places them in the appropri-

ate service on the vendor side.

Sky Computing redeﬁnes cloud usage by routing

jobs through an intercloud broker, which selects and

manages cloud services for execution. This creates

a two-sided market, linking user-submitted jobs with

cloud services. Some services are multi-cloud (e.g.,

Kubernetes, Apache Spark), while others are cloud-

speciﬁc (e.g., AWS Inferentia) (Chasins et al., 2022).

Unlike traditional multi-cloud setups, Sky Comput-

ing allows workloads to run on multiple clouds or

split across them, offering greater ﬂexibility and ben-

eﬁts. It focuses on partial compatibility, enabling

many—but not all—jobs to run across clouds, with

compatibility improving over time.

A characteristic of Sky Computing is its dis-

CLOSER 2025 - 15th International Conference on Cloud Computing and Services Science

212

Service

Catalog

Tracker

Service

Publisher

Service

Publisher

Compatibility Set Compatibility Set

Cloud A Cloud B

Intercloud

Broker

Price,

APIs,...

Executor

Provisioner

Optimizer

Private Cloud On-premise

Workload

Figure 2: Overview of the components of the inter-cloud broker (Yang et al., 2023).

tributed infrastructure, which dynamically allocates

workloads across providers. This ensures scalability

and independence from workload type or provider,

maximizing resource utilization. Figure 2 visually

represents the Sky Computing framework, illustrating

how it integrates cloud resources into a cohesive and

scalable ecosystem.

The Service Catalog captures the instances and

services available in each cloud, detailed information

about locations that offer them, and the APIs for allo-

cating and accessing them. It also stores the long-

term pricing for on-demand virtual machines, data

storage, egress, and services (typically, these prices

stay the same for months). The term refers to all

prices for a ﬁxed time horizon. Cloud computing

services are highly time-dependent, meaning that the

time interval of the service offered is speciﬁed in ad-

vance (Ibrahimi, 2017). The Service Catalog can pro-

vide ﬁltering and search functionality based on infor-

mation published by the cloud providers, listed by a

third party, or collected by the broker.

The tracker tracks spot prices (which may change

more frequently, e.g., hourly or daily) and the avail-

ability of resources across different cloud providers

and locations. This module is of central importance

as the prices of cloud providers and the associated ser-

vices are a decision metric for service placement.

The Optimizer processes the workload require-

ments and checks the availability of instances and ser-

vices and their prices, which the Service Catalog and

Tracker provide. This module then calculates the op-

timal placement of the services. In a broader con-

text, the optimality of workload placement is heavily

dependent on the used metrics. An algorithm using

multiple parameters for calculating optimality criteria

and decision-making is needed (see section 7). If re-

source availability and/or price change, the Optimizer

can perform a new optimization.

The Provisioner module manages the resources by

allocating the resources required for execution. The

Optimizer’s execution plan allocates the resources ac-

cordingly and releases them when each task is com-

pleted. The executor manages the application by ag-

gregating the functions of each workload and execut-

ing them based on the resources allocated by the cloud

provider and service provider.

Compatibility sets are an essential feature of Sky

Computing. The focus is on using existing services

and APIs from all cloud providers, intended to offer

transparent and standardized options for connecting

services without reimplementing them.

Sky Computing provides a reasonable basis for

The Evolution of Cloud Computing Towards a Vendor Agnostic Market Place Using the SKY CONTROL Framework

213

implementing our proposed framework since it adds

an abstraction layer between users and cloud service

providers. However, having a tool or component in

the Sky Computing concept is still essential for ana-

lyzing the costs drawn by the services used.

The extension of the cloud by constructing an in-

tercloud broker poses new challenges that need to be

tackled, especially in networking and security. In Sky

Computing, tackling the more ambitious task of au-

tomated assurance of suitable security levels for all

assets is a considerable challenge. A few years ago,

this would have been an almost impossible task in het-

erogeneous environments. However, the concept of

SASE - Secure Access Service Edge - was introduced

in 2019 and has since evolved into a solid technology

that offers new possibilities for our purposes (Islam

et al., 2021; MacDonald et al., 2019). SASE is a syn-

ergetic combination of techniques for controlling se-

curity - i.e., enforcing security policies - in complex,

heterogeneous infrastructures.

SASE generally includes the following function-

alities:

• SD-WAN(Yang et al., 2019) optimizes wide-area

networks by allowing organizations to use mul-

tiple transport services (MPLS, LTE, 5G, broad-

band) for secure connectivity(Islam et al., 2021).

• Secure Web Gateway (SWG) ﬁlters and moni-

tors web trafﬁc to protect users from threats and

ensure compliance.

• Cloud Access Security Broker (CASB) en-

forces security policies between cloud users and

providers.

• Firewall-as-a-Service (FWaaS) provides scal-

able, cloud-based ﬁrewall functionalities.

• Zero Trust Network Access (ZTNA) follows the

never trust, always verify principle, securing user

sessions inside and outside corporate networks.

SASE enables centralized policy management

with distributed enforcement points, ensuring local

decision-making when needed (van der Walt and Ven-

ter, 2022). For example, policies can be enforced

locally via Customer Premises Equipment (CPE) or

managed device agents, enhancing security ﬂexibil-

ity.

Integrating SASE into Sky Computing addresses

security challenges such as asset and risk-based man-

agement, though practical implementation is still

evolving. While these concepts promise seamless

workload distribution and cloud interoperability, their

adoption—especially for SMEs—requires further re-

search. The following section explores industry use

cases, potential developments, and open questions.

4 PROPOSITION FOR THE

TRANSFORMATION OF THE

CLOUD COMPUTING

MARKET PLACE

Sky Computing offers signiﬁcant potential for the in-

dustry, mainly by providing end-users with more op-

tions for optimizing workload operations. The inter-

cloud broker could shift the balance from large cloud

vendors to companies, beneﬁting small to medium-

sized enterprises (SMEs). However, Sky Computing

must become an attractive, incentive-driven technol-

ogy for widespread adoption.

A key focus is addressing end-users’ needs, es-

pecially when deploying multi-cloud environments.

Multi-cloud setups offer beneﬁts like preventing ven-

dor lock-in and enabling the best services from var-

ious providers. However, managing resources and

costs across multiple providers and meeting gover-

nance requirements remains a challenge, particularly

for smaller companies.

The complexity of multi-cloud environments in-

creases for SMEs, and understanding their challenges

is essential for developing a framework that solves

these issues. Although research on multi-cloud ar-

chitectures exists (Baryannis et al., 2013; Kavitha

and Radha, 2022), introducing an intercloud broker

can help reduce complexity by creating an abstraction

layer between end-users and cloud vendors.

In summary, we propose adding an abstraction

layer to create a new marketplace for end-users, en-

abling interaction with vendors through a new plat-

form for negotiation rather than changing the ven-

dors’ services or APIs.

5 OPEN QUESTIONS

The proposition in Section 4 raises key questions

requiring further research. One major challenge is

establishing an open and user-friendly marketplace

for end-users. While an intercloud broker is a step

forward, practical implementation remains unclear.

Adoption by cloud vendors is not critical, but end-

user acceptance is essential, necessitating a standard-

ized framework for multi-cloud setups.

Another challenge is designing an attractive

framework for stakeholders, including executives,

technicians, and security auditors. C-level execu-

tives require cost insights, technicians need perfor-

mance data and easy conﬁgurations, and auditors pri-

oritize governance and risk visualization. Although

the SkyPilot project (Stoica and Shenker, 2021; Yang

CLOSER 2025 - 15th International Conference on Cloud Computing and Services Science

214

et al., 2023) has demonstrated Sky Computing’s vi-

ability for ML pipelines, broader industry-driven use

cases must be explored. We, therefore, want to start

inspecting the placement of IaaS (Infrastructure as a

Service) workloads since many companies use this

type of service model for migrating legacy applica-

tions to the cloud (so-called ”Lift-and-Shift” opera-

tion) (Ahmad et al., 2018). A key research question

is: How can Sky Computing enhance workload place-

ment efﬁciency for SMEs compared to multi-cloud se-

tups?

Efﬁcient workload placement is crucial, requiring

automation to minimize manual intervention, espe-

cially as workload volumes grow. SMEs with lim-

ited budgets and resources would particularly bene-

ﬁt from productivity and cost improvements through

Sky Computing. Technologies such as SD-WAN and

SASE introduce security and network conﬁguration

automation, but integrating computing resources like

VMs and container runtimes remains an open re-

search area.

Adoption risks include resistance from cloud ser-

vice providers (CSPs), who may restrict API access

to maintain market control, hindering interoperability.

Addressing this requires careful research into busi-

ness and technical strategies.

This section’s open questions guide the conceptu-

alization of the SKY CONTROL framework, which

aims to optimize multi-cloud environments for end-

users. As an initial step, SKY CONTROL will be

applied to IaaS workloads in SMEs, evaluating its

performance against deployments using major CSPs

like AWS and Azure. Later, higher-abstraction mod-

els like PaaS will be explored. The next section de-

tails the SKY CONTROL framework and its core con-

cepts.

6 CURRENT DEVELOPMENT IN

THIS FIELD

As an outlook and initial attempt, we present the

conceptual frameworks to address the challenges dis-

cussed in the previous sections. We introduce SKY

CONTROL, a framework for SMEs to control and op-

timize multi-cloud setups. This provides a common

ground for companies to onboard Sky Computing.

Another concept discussed in this section is the anal-

ysis of methods and technologies for attaching work-

loads to the Sky Computing ecosystem (section 8).

In this project, the geographic distribution of work-

loads across WAN boundaries will be investigated to

identify suitable solutions for the technological im-

plementation of an intercloud broker.

7 SKY CONTROL: A

FRAMEWORK FOR SMES

The distributed nature of multi-cloud environments

offers many beneﬁts for SMEs, but it also brings sig-

niﬁcant challenges in managing workloads across dif-

ferent CSP platforms. The overview of the costs of

the companies’ services is challenging, and there are

substantial drawbacks in choosing such architectures.

Another critical aspect is the overview and analysis of

potential security risks and the governance of assets.

SKY CONTROL addresses the challenges SMEs face

when using multi-cloud deployments. The following

sections present the architecture of SKY CONTROL,

along with the desired functionalities needed to fulﬁll

the requirements of SMEs.

7.1 Architecture of SKY CONTROL

This section describes the key components of the

SKY CONTROL framework:

• Cost Control: Analyzes, calculates, and visu-

alizes costs for both on-premise and cloud re-

sources. It performs static analysis (e.g., resource

IDs, hardware specs) and dynamic analysis (e.g.,

CPU/memory usage, network bandwidth). Pric-

ing trends and predictions are derived from this

data, considering the complexity of cloud service

integration. A control and planning tool provides

insights into resource usage across multiple cloud

providers, with visualization for easier compre-

hension.

• Risk Management: Manages customer assets,

collects detailed insights, and performs risk anal-

ysis. It evaluates risks based on asset critical-

ity, data sensitivity, and compliance standards like

C5 (Cloud Computing Compliance Criteria Cata-

logue) (Di Giulio et al., 2017) for German SMEs.

This helps businesses meet governance require-

ments and enhances compatibility with larger en-

terprises. The module also provides risk and asset

visualization for CIOs, aiding audits and risk mit-

igation.

Figure 3 illustrates the architecture of the pro-

posed framework.

This section highlights the necessity of analyzing

pricing, workload criticality, and governance infor-

mation to deﬁne decision criteria for optimal work-

load placement. To achieve this, an algorithm must be

developed to calculate and formalize these decisions

within the Optimizer module (as deﬁned in section 3).

Given that ”optimality” varies across different factors,

the SKY CONTROL framework must ﬁrst establish a

The Evolution of Cloud Computing Towards a Vendor Agnostic Market Place Using the SKY CONTROL Framework

215

Cost

Analysis

Tool

SKY CONTROL

Cost

Calculator

Static/

Dynamic

Analysis

Monitoring

System Monitoring

and Logging

Planning

and Control

Deployment/

Control

CFO

Data collection for

resources

and

Enterprise

Asset

Management

Information Security Data Analysis

Risk

Analysis

Governance

Control

Data collection for

Assets and

Governance

Visualization

Cost

Control

Risk

Management

Cloud

Service

Provider

On-Premise

API for On-Premise and Cloud resources

CEO CIO Auditor

SME

Control

Technician

Figure 3: SKY CONTROL architecture.

conceptual foundation for analyzing these criteria be-

fore integrating them into the optimization process.

At its core, SKY CONTROL aims to provide

transparency into the properties of the entire infras-

tructure and its components. The project is designed

to be highly adaptable, leveraging both existing and

emerging technologies. Some of these technologies

will inspire SKY CONTROL’s development, while

others may be directly integrated into its architecture

as functional components.

8 ATTACHING ON-PREMISES

WORKLOADS TO THE SKY

This ﬁeld of research explores methods for integrat-

ing workloads into the Sky Computing ecosystem,

focusing on geographic workload distribution across

WAN boundaries to develop an intercloud broker. It

investigates an ofﬂine-ﬁrst strategy, prioritizing local

workload operation while enabling cloud relocation

as needed, enhancing service usability and availabil-

ity.

The emphasis of this sub-project lies in the inves-

tigation of geographical scaling, often overlooked in

practice, to optimize service proximity to users and

enhance performance.

Key resources examined in this strategy include:

• Computing Capacity (via VMs, containers)

• Network Resources (leveraging SDN virtualiza-

tion)

• Storage Resources (for distributed data synchro-

nization)

• Software

• Services

The exploration of the outsourcing and coopera-

tion of cloud services with client-side resources, in-

vestigating which cloud services can be ofﬂoaded

to clients and under what conditions is a central

point of interest. Unlike the cloud-ﬁrst approach, an

ofﬂine-ﬁrst strategy will prioritize local network ser-

vice availability, ensuring network-independent cloud

services while still enhancing overall functionality

through cloud cooperation.

A critical aspect is the scaling of services:

• Vertical scaling (adding resources)

• Horizontal scaling (adding service instances)

• Geographical scaling (placing services closer to

users, or “scaling away”)

An inter-cloud broker could facilitate seamless

interoperability between cloud and on-premise envi-

ronments using SKY CONTROL mechanisms. This

would allow users to access nearby services inde-

pendently of the cloud, increasing autonomy and re-

silience without losing cloud beneﬁts.

Key objectives include:

CLOSER 2025 - 15th International Conference on Cloud Computing and Services Science

216

• Deﬁning criteria for outsourcing computing-

intensive tasks, an area with no standardized re-

search methodology.

• Investigating service migration (Rodriguez et al.,

2021) to optimize resource distribution across

cloud, end devices, and services.

• Exploring vertical migration (between cloud and

end devices) and horizontal migration (between

end devices), assessing their impact and technical

feasibility.

The project will evaluate how local resource uti-

lization enhances service performance and availabil-

ity.

9 CONCLUSION AND OUTLOOK

Sky Computing is a new Cloud Computing paradigm

that abstracts end-users from cloud vendors via an in-

tercloud broker, simplifying multi-cloud management

by eliminating API complexities.

We analyzed this concept and proposed practical

implementation strategies, identifying open questions

that highlight the need for new frameworks, methods,

and technologies to develop a marketplace from an

end-user perspective.

Our proposed SKY CONTROL framework offers

cost control and risk management solutions tailored

for SMEs. By integrating dynamic analysis, real-time

cost tracking, and risk assessment, SKY CONTROL

enhances transparency and control over IT infrastruc-

ture. Its Sky Computing capabilities improve multi-

cloud ﬂexibility and efﬁciency, giving SMEs a com-

petitive edge.

Additionally, we explored methods for attaching

workloads to the Sky Computing ecosystem, focusing

on geographic workload distribution across WANs.

An ofﬂine-ﬁrst strategy will be investigated to ensure

local workload operation while enabling cloud relo-

cation as needed.

This concept paper presents foundational ideas

for Sky Computing implementation, acknowledging

that many proposed questions remain open. As SKY

CONTROL evolves, we aim to provide practical re-

sults in the near future.

ACKNOWLEDGEMENTS

This work was funded by the Federal Ministry for

Economic Affairs and Climate Action (Bundesmin-

isterium f

ur Wirtschaft und Klimaschutz) under the

Central Innovation Programme for SMEs (Zentrales

Innovationsprogramm Mittelstand).

We thank our partners at Systrade GmbH for their

support, especially Andreas Schmidt, Thorsten Luft,

Martin Schesny, Julian Hofmann, and Abo El Hage.

Special thanks to Wei Yin Shing for proofreading and

improving the paper’s quality.

REFERENCES

Ahmad, N., Naveed, Q. N., and Hoda, N. (2018). Strategy

and procedures for migration to the cloud computing.

In 2018 IEEE 5th International Conference on Engi-

neering Technologies and Applied Sciences (ICETAS),

pages 1–5.

Ardagna, D. (2015). Cloud and multi-cloud computing:

Current challenges and future applications. In 2015

IEEE/ACM 7th International Workshop on Principles

of Engineering Service-Oriented and Cloud Systems,

pages 1–2.

Barker, A., Varghese, B., and Thai, L. (2015). Cloud ser-

vices brokerage: A survey and research roadmap. In

2015 IEEE 8th International Conference on Cloud

Computing, pages 1029–1032.

Baryannis, G., Garefalakis, P., Kritikos, K., Magoutis,

K., Papaioannou, A., Plexousakis, D., and Zeginis,

C. (2013). Lifecycle management of service-based

applications on multi-clouds: a research roadmap.

In Proceedings of the 2013 International Workshop

on Multi-Cloud Applications and Federated Clouds,

MultiCloud ’13, page 13–20, New York, NY, USA.

ACM.

Chasins, S., Cheung, A., Crooks, N., Ghodsi, A., Goldberg,

K., Gonzalez, J. E., Hellerstein, J. M., Jordan, M. I.,

Joseph, A. D., Mahoney, M. W., Parameswaran, A.,

Patterson, D., Popa, R. A., Sen, K., Shenker, S., Song,

D., and Stoica, I. (2022). The sky above the clouds.

Di Giulio, C., Sprabery, R., Kamhoua, C., Kwiat, K., Camp-

bell, R. H., and Bashir, M. N. (2017). Cloud stan-

dards in comparison: Are new security frameworks

improving cloud security? In 2017 IEEE 10th Inter-

national Conference on Cloud Computing (CLOUD),

pages 50–57.

Fortes, J. A. B. (2010). Sky computing: When multi-

ple clouds become one. In Proceedings of the 2010

10th IEEE/ACM International Conference on Cluster,

Cloud and Grid Computing, CCGRID ’10, page 4,

USA. IEEE Computer Society.

Georgios, C., Evangelia, F., Christos, M., and Maria, N.

(2021). Exploring cost-efﬁcient bundling in a multi-

cloud environment. Simulation Modelling Practice

and Theory, 111:102338.

Gundu, S. R., Panem, C. A., and Thimmapuram, A. (2020).

Hybrid it and multi cloud an emerging trend and im-

proved performance in cloud computing. SN Com-

puter Science, 1(5).

Hong, J., Dreibholz, T., Schenkel, J. A., and Hu, J. A.

(2019). An Overview of Multi-cloud Computing. In

The Evolution of Cloud Computing Towards a Vendor Agnostic Market Place Using the SKY CONTROL Framework

217

Barolli, L., Takizawa, M., Xhafa, F., and Enokido, T.,

editors, Web, Artiﬁcial Intelligence and Network Ap-

plications, pages 1055–1068, Cham. Springer Interna-

tional Publishing.

Ibrahimi, A. (2017). Cloud computing: Pricing model. In-

ternational Journal of Advanced Computer Science

and Applications, 8(6).

Islam, M. N., Colomo-Palacios, R., and Chockalingam, S.

(2021). Secure access service edge: A multivocal lit-

erature review. In 2021 21st International Confer-

ence on Computational Science and Its Applications

(ICCSA), pages 188–194.

Jamshidi, P., Pahl, C., and Mendonc¸a, N. C. (2016). Pattern-

based multi-cloud architecture migration. Software:

Practice and Experience, 47(9):1159–1184.

Kavitha, M. G. and Radha, D. (2022). Quality, Security

Issues, and Challenges in Multi-cloud Environment:

A Comprehensive Review, pages 269–285. Springer

International Publishing, Cham.

MacDonald, N., Orans, L., and Skorupa, J. (2019). The

future of network security is in the cloud. Gartner.

Mell, P., Grance, T., of Standards, N. I., and Division, T.

U. C. S. (2011). The NIST Deﬁnition of Cloud Com-

puting. NIST special publication. U.S. Department of

Commerce, National Institute of Standards and Tech-

nology.

Monteiro, A., Teixeira, C., and Pinto, J. S. (2014). Sky com-

puting: Exploring the aggregated cloud resources —

part ii. In 2014 9th Iberian Conference on Information

Systems and Technologies (CISTI), pages 1–6.

Mulder, J. (2020). Multi-Cloud Architecture and Gov-

ernance: Leverage Azure, AWS, GCP, and VMware

vSphere to build effective multi-cloud solutions. Packt

Publishing.

Petcu, D. (2013). Multi-cloud: expectations and current

approaches. In Proceedings of the 2013 international

workshop on Multi-cloud applications and federated

clouds, ICPE’13. ACM.

Siegel, J. (1999). An overview of corba 3. In Kutvonen, L.,

onig, H., and Tienari, M., editors, Distributed Appli-

cations and Interoperable Systems II, pages 119–132,

Boston, MA. Springer US.

Stoica, I. and Shenker, S. (2021). From cloud computing

to sky computing. In Proceedings of the Workshop

on Hot Topics in Operating Systems, HotOS ’21, page

26–32, New York, NY, USA. ACM.

van der Walt, S. and Venter, H. (2022). Research gaps and

opportunities for secure access service edge. In Inter-

national Conference on Cyber Warfare and Security,

volume 17, pages 609–619.

van Steen, M. and Tanenbaum, A. (2017). Distributed Sys-

tems. 3rd edition. Self-published, open publication.

Wei, X., Mohaimenur Rahman, A. B. M., and Wang, Y.

(2021). Data placement strategies for data-intensive

computing over edge clouds. In 2021 IEEE Interna-

tional Performance, Computing, and Communications

Conference (IPCCC), pages 1–8.

Yang, Z., Cui, Y., Li, B., Liu, Y., and Xu, Y. (2019).

Software-deﬁned wide area network (sd-wan): Archi-

tecture, advances and opportunities. In 2019 28th In-

ternational Conference on Computer Communication

and Networks (ICCCN), pages 1–9.

Yang, Z., Wu, Z., Luo, M., Chiang, W.-L., Bhardwaj,

R., Kwon, W., Zhuang, S., Luan, F. S., Mittal, G.,

Shenker, S., and Stoica, I. (2023). SkyPilot: An in-

tercloud broker for sky computing. In 20th USENIX

Symposium on Networked Systems Design and Imple-

mentation (NSDI 23), pages 437–455, Boston, MA.

USENIX Association.

CLOSER 2025 - 15th International Conference on Cloud Computing and Services Science

218