Source-linked AI summary
Research Challenges for Enterprise Cloud Computing
Ali Khajeh-Hosseini, Ian Sommerville, Ilango Sriram
TL;DR
The paper examines enterprise research challenges arising from cloud computing’s shift from purchased computing products to internet-delivered services. It reviews existing literature across organizational, economic, security, legal, and privacy dimensions, concluding that these challenges are interdisciplinary and relevant to wider adoption. The paper also notes that organizational issues are difficult to study because real-world migration decisions remain complex and enterprises must take the plunge before they can be investigated in practice.
Problem
Enterprise cloud adoption involves organizational, economic, security, legal, and privacy challenges that are not adequately addressed by a purely technical view.
Method
The paper reviews existing cloud-computing literature from an enterprise perspective, identifies research challenges within each topic, and discusses them collectively.
Results
The paper identifies interdisciplinary enterprise research challenges involving organizational change, utility billing, costs, and security, legal, and privacy issues.
Takeaways & Limitations
Cloud-computing research and adoption decisions must consider organizational and usage-practice effects alongside technological changes.
Takeaways & Limitations
Organizational issues are difficult to investigate because enterprises must undertake real-world migration before these issues can be studied in practice.
Abstract
from arXiv · showhide
Cloud computing represents a shift away from computing as a product that is purchased, to computing as a service that is delivered to consumers over the internet from large-scale data centers - or "clouds". This paper discusses some of the research challenges for cloud computing from an enterprise or organizational perspective, and puts them in context by reviewing the existing body of literature in cloud computing. Various research challenges relating to the following topics are discussed: the organizational changes brought about by cloud computing; the economic and organizational implications of its utility billing model; the security, legal and privacy issues that cloud computing raises. It is important to highlight these research challenges because cloud computing is not simply about a technological improvement of data centers but a fundamental change in how IT is provisioned and used. This type of research has the potential to influence wider adoption of cloud computing in enterprise, and in the consumer market too.
1. INTRODUCTION
Cloud computing research is framed here as an enterprise and organizational challenge, not merely a technical improvement to data centers. The paper reviews existing literature and highlights challenges involving organizational change, utility billing, and security, legal, and privacy issues.
- Research gap: Existing cloud-computing research has largely focused on technical problems, leaving comparatively little work on enterprise-oriented research challenges.The paper positions its review as addressing this underdeveloped perspective.
- Research perspective: Cloud computing is presented as a fundamental change in how IT is provisioned and used, requiring enterprises to assess organizational effects alongside benefits and risks.The paper argues that adoption depends partly on organizational, cultural, and legislative maturity.
- Enterprise context: Enterprise cloud decisions must account for complex systems, dependencies, multiple stakeholders, and established organizational processes.Relevant stakeholders include technical, project, operations, and financial managers, as well as system engineers.
- Research scope: The paper addresses organizational change, the economic and organizational implications of utility billing, and security, legal, and privacy issues.These topics structure the paper’s subsequent sections.
- Paper approach: Each section reviews existing cloud-computing literature from an enterprise perspective before identifying research challenges for further discussion.The paper aims to provide a resource for researchers in this rapidly developing field.
2. ORGANIZATIONAL CHANGE
Cloud computing can redistribute control, alter IT procurement and support practices, and require organizational changes across enterprise stakeholders. The literature identifies these effects but leaves important questions about planning and implementing the changes unresolved.
- Organizational change: Cloud computing can deliver computing facilities as metered services, changing enterprise IT provision from owned infrastructure toward utility-style consumption.The utility model includes services such as storage, compute power, network infrastructure, and applications.
- Organizational change: Utility billing and shared infrastructures require cultural and organizational process maturity and significant changes to business processes and organizational boundaries.Enterprise IT departments are described as unaccustomed to utility billing across shared infrastructures.
- Organizational change: Cloud computing may separate software integration from hardware management through narrow interfaces such as virtual machine images supplied by IaaS providers.This example illustrates how cloud services can resolve conflicts between system-development roles.
- Power and roles: Cloud computing may erode central IT authority by turning users into choosers who can replace internally provided services with cloud services.Cloud providers are also beginning to offer support services, potentially reducing reliance on central IT for user support.
- Research challenges: Research has offered glimpses of enterprise change but has not substantially addressed how to plan and design the organizational changes required by cloud migration.The paper compares the prospective shift with the earlier redistribution of power caused by departmental PC purchasing.
- Power and roles: Migration can shift power beyond the organization to cloud providers, raising questions about the future roles of central IT, compliance departments, support personnel, and system administrators.The paper explicitly identifies uncertainty about control, compliance access, political resistance, and changing working practices.
- Support practices: Cloud adoption may reduce local expertise and leave users uncertain about where to obtain help when existing organizational coping strategies and workarounds disappear.Reducing IT support costs is identified as one reason local expertise may decline.
3. COSTS
Cloud-computing costs extend beyond direct infrastructure expenses to include pricing models, elasticity, migration, procurement, and the costs of providing private-cloud services. Enterprise adoption therefore raises economic questions that existing literature addresses only partially.
- Consumer Perspective: Pay-per-use billing makes application design and resource usage directly visible in running costs.Providers use detailed costing models and metrics, while pricing may be tiered, per-unit, or subscription-based.
- Consumer Perspective: Elasticity can shift over-provisioning costs and under-provisioning risks from users to cloud providers.However, many enterprise systems assume scaling up rather than scaling out, making cloud-oriented redesign expensive or impossible.
- Consumer Perspective: Economic evaluation frameworks compare cloud adoption with conventional infrastructure, but some remain conceptual and lack results.Klems et al. identified economic and technical considerations, whereas Walker modeled leased versus purchased CPU-hour costs.
- Consumer Perspective: Cloud-computing cost research also evaluates hybrid capacity expansion, but Average Weighted Response Time may not capture the benefits users expect from instant allocation.Further research is needed to identify suitable metrics for public clouds complementing private infrastructure.
- Provider Perspective: Research should examine provider-side costs because enterprises will likely retain private clouds and may rent spare capacity.Relevant models include total cost of ownership and utilization cost, incorporating hardware, software, power, cooling, staff, real estate, and virtual-machine density.
- Research Challenges: Current enterprise cost discussions often omit application migration and procurement policies, while utility billing conflicts with budget ownership and approval requirements.Usage patterns and deployment options introduce uncertainty into procurement processes designed around known costs.
4. SECURITY, LEGAL AND PRIVACY ISSUES
Cloud computing raises interconnected security, legal, and privacy challenges because users lack control over physical infrastructure, data location, and shared environments. Existing work identifies technical and regulatory risks, while enterprises still need clearer, scenario-specific guidance and liability arrangements.
- Security and Privacy: Users’ limited control over physical infrastructure creates security and privacy concerns, while physical location determines legal jurisdiction.Multitenancy additionally requires new security and privacy solutions.
- Security and Privacy: Cloud incidents can create distinctive financial exposure, such as users paying for resource consumption caused by denial-of-service attacks.This exposure may exceed the application’s normal in-house peak usage.
- Security and Privacy: The Cloud Security Alliance and ENISA organize cloud risks across policy, technical, legal, and cloud-independent categories.Examples include vendor lock-in, loss of governance, data leakage, DDoS attacks, data protection, licensing, network problems, and natural disasters.
- Security and Privacy: Security can become cheaper at larger scale, allowing providers to offer measures such as hardened virtual machines and specialized expertise.This potential benefit coexists with the risks identified for cloud users.
- Legal and Privacy Issues: Privacy and legal compliance are difficult when organizations cannot fully specify how or where cloud data will be processed.This limits informed consent and complicates confidentiality, copyright, regulated-sector obligations, and data-transfer decisions.
- Legal and Privacy Issues: Detailed legal guidance remains difficult because each cloud-usage scenario differs, despite existing reviews identifying major issues.Enterprises are also cautious because complex systems may require some data to remain within specific jurisdictions.
- Research Challenges: Cloud adoption leaves unresolved questions about liability, regulatory interpretation, and controlling staff behavior across mixed in-house and cloud systems.The paper identifies these as areas requiring further investigation and possible negotiated or regulatory solutions.
5. DISCUSSION
Cloud computing may change how large organizations work, but realizing its value requires addressing complex technical, organizational, economic, and operational challenges. The paper emphasizes that these challenges are difficult to study and require interdisciplinary research.
- Enterprise challenges: Enterprise cloud adoption must address complex IT systems, diverse stakeholder requirements, organizational change, interoperability, and resilience.Large enterprises contain long-lived systems with dependencies across technologies, data, middleware, platforms, and business processes.
- Potential practice changes: Cloud computing may reduce the cost of disaster recovery and support easier prototyping, parallel operation, incremental delivery, and load simulation.External cloud facilities can support experiments before migration to internal servers, although data transfer may introduce performance problems.
- Migration decisions: Cloud adoption requires organizations to reconsider IT-department work, system-running costs, risks, benefits, and usage practices.These issues affect migration decisions, and the enterprise shift to cloud computing may take 10 to 15 years.
- Migration decisions: System-modeling frameworks could support migration decisions by representing application portfolios, data, infrastructure, and costs.Existing modeling notations such as UML and SysML may be extended, while legacy integration and data migration remain necessary concerns.
- Research implications: Cloud research cannot be approached purely technically because its challenges are interdisciplinary and require cooperation among researchers, users, and providers.The paper presents these challenges as relevant to wider cloud adoption rather than as a negative judgment on enterprise cloud use.