Multiuser Privacy and Security Conflicts in the Cloud

Title of the Paper

Multiuser Privacy and Security Conflicts in the Cloud

Publication Information

• Authors: Eman Alhelali, Kopo M. Ramokapane, Jose Such
• Conference: CHI ’23
• Year: 2023
• Field: Multiuser Privacy and Cloud Computing Security
• Keywords: Multiuser Conflicts, Multiparty Privacy Conflicts, Interconnected Privacy, Cloud Security, Collaborative Platforms

Research Background and Problem

• Identified Problems or Challenges:
  The use of cloud collaboration platforms (e.g., Google Drive, Dropbox) facilitates teamwork but also triggers multiuser privacy and security conflicts (MCs). These conflicts may involve data breaches, privacy issues, and access problems. However, there has been little in-depth analysis of multiuser conflicts in cloud environments in the existing literature.

• Significance:
  The impact of multiuser conflicts can lead to privacy breaches, data loss, and relationship breakdowns. In workplace and educational settings, such conflicts can severely affect productivity. As the volume of cloud data continues to grow (projected to exceed 100 zettabytes by 2025), related privacy and security issues will become increasingly critical.

• Research Motivation and Related Work:
  While multiuser privacy conflicts have been extensively studied in domains such as social media, research in cloud computing environments is limited and often restricted to very specific scenarios (e.g., conflicts caused by renaming or relocating files). This study aims to systematically investigate multiuser conflicts in cloud platforms, covering the entire process from conflict occurrence to resolution, and providing substantial empirical data to fill this gap.

Solution

• Research Methodology:
  Using the Critical Incident Technique (CIT), the researchers designed four questionnaires targeting both "Initiators" and "Affected" parties in conflicts, covering scenarios such as file and folder sharing, editing, and permission settings. A mixed-methods approach (survey, qualitative, and quantitative analysis) was employed to collect and analyze 1,050 participants' instances of multiuser conflicts.

• Innovations:

  • This study provides the first comprehensive investigation of multiuser conflicts in the cloud computing domain, examining conflict characteristics, root causes, impacts, and resolution strategies.
  • It explores the mechanisms of conflict occurrence, including scenarios such as data loss, privacy breaches, and file corruption, while distinguishing between participant roles (Initiator/Affected).
  • The study proposes specific recommendations for improving cloud platform privacy and security design, such as sensitive content detection, collaboration interface optimization, and conflict communication mechanisms.

• Implementation Steps:

  • Conduct surveys to collect conflict scenarios and data.
  • Perform quantitative analysis to assess the severity and frequency of conflicts; conduct qualitative analysis of user-submitted conflict descriptions.
  • Define conflict characteristics and propose both technical and non-technical resolution strategies.
  • Compare cloud computing conflicts with similar conflicts in social media to identify similarities and differences.

Research Findings

• Specific Findings:

  • Conflict Characteristics: The vast majority of participants had experienced at least one multiuser conflict, with most occurring within the past six months. The primary actions leading to MCs included sharing (44%), permission settings (19%), file editing (17%), and deletion (13%).
  • Conflict Consequences: The most common consequences included privacy and security losses, such as confidentiality breaches (37%), data unavailability (17%), and data integrity issues (15%). Conflicts sometimes resulted in decreased productivity, relationship breakdowns, and negative emotions (e.g., anger, frustration).
  • Resolution Strategies: Non-technical strategies included dialogue, explanation, and negotiation; technical strategies involved adjusting permissions, deleting files, recovery, and protective actions. The resolution rate for conflicts was 84%, though severe conflicts were harder to resolve.

• Comparison with Existing Solutions:

  • Cloud conflicts in workplace settings were more frequent, whereas social media conflicts were more concentrated among friends and family.
  • The types and scope of conflicts in cloud platforms were broader, requiring solutions that address productivity and efficiency needs.

• Experimental or Evaluation Results:

  • The severity of conflicts was directly related to whether communication occurred to resolve them. Communication by the affected party significantly increased the resolution rate (87% vs. 26%).
  • Statistical and qualitative analyses revealed that conflicts in cloud computing environments often have more complex technical dependencies, requiring solutions that integrate technical measures with user behavior patterns.

• Limitations and Future Directions:

  • Limitations: The sample may be biased toward tech-savvy users, potentially not representing all user groups; participants might exhibit subjective bias when describing conflicts.
  • Future Directions:
    (1) Further study the perceptual differences between Initiators and Affected parties and how to align perspectives during conflict resolution.
    (2) Explore the impact of cloud interface design on reducing user errors.
    (3) Develop machine learning algorithms to detect and classify the sensitivity of cloud files, enabling more precise understanding of data-sharing risks.

Conclusion

This study provides an in-depth exploration of multiuser privacy and security conflicts in cloud environments, revealing their complex interaction mechanisms and differences in user perceptions. It offers valuable insights and practical recommendations for designing secure and collaborative cloud platforms.