[Paper Review] When Cars Meet Distributed Computing: Data Storage as an Example

Lewis Tseng, Takamasa Higuchi+1 more|arXiv (Cornell University)|Nov 6, 2017

Vehicular Ad Hoc Networks (VANETs)10 references3 citations

TL;DR

This paper proposes a Macro-Micro Cloud (MMC) architecture to enable scalable, low-latency vehicular cloud services by organizing vehicles into stable, location-tied micro clouds for long-lasting, distributed data storage. By leveraging hierarchical coordination and dynamic handoff mechanisms, MMC reduces communication overhead and improves service continuity in mobile environments, demonstrating feasibility for location-based data storage using GPS-based atomicity and availability modeling.

ABSTRACT

As cars are ubiquitous they could play a major role in a next generation communication and computation framework. In the last years, the development of vehicle-to-vehicle communication and vehicle-to-infrastructure communication took huge steps forward and therefore gives us the tools to build "mobile computing service" on cars equipped with computation capabilities. Recently, several groups of researchers independently proposed the design of "vehicular clouds" that materializes the concept. In this paper, we introduce a new paradigm of the vehicular clouds, followed by a case study of data storage on top of the proposed cloud. Finally, we present several challenges and opportunities in the intersection of vehicular clouds and distributed computing.

Motivation & Objective

• To address the limitations of existing vehicular cloud models in supporting long-lasting, location-based services due to vehicle mobility and high communication overhead.
• To design a hierarchical vehicular cloud architecture that enhances service continuity and reduces reliance on V2I infrastructure.
• To enable efficient, scalable data storage in vehicular networks by modeling micro cloud availability and resource allocation.
• To explore incentive mechanisms and multi-radio communication strategies for robust and performant vehicular cloud deployment.

Proposed method

• Organizing vehicles into dynamic, location-specific micro clouds using a membership management protocol to form logically stable service clusters.
• Implementing a hierarchical cloud structure where micro clouds form part of a broader macro cloud for wide-area service discovery and provisioning.
• Using historical data to model resource availability (storage and bandwidth) per micro cloud for predictive data placement.
• Applying a GPS-based atomicity mechanism to approximate consistency in geo-distributed data storage, replacing traditional quorum systems.
• Designing a hand-off mechanism among micro clouds to maintain service continuity when vehicles leave a region.
• Adapting the GeoQuorum model for data replication and storage placement across micro clouds based on estimated availability.

Experimental results

Research questions

• RQ1 How can vehicular clouds efficiently support long-lasting, location-based services despite high vehicle mobility?
• RQ2 What architectural design enables stable service provisioning with minimal communication overhead in dynamic vehicular environments?
• RQ3 How can resource availability in micro clouds be modeled and used to guide data placement and replication?
• RQ4 What mechanisms ensure data consistency and atomicity in a GPS-synchronized, decentralized vehicular storage system?
• RQ5 How can incentives be designed to encourage vehicle participation in a decentralized vehicular cloud infrastructure?

Key findings

• The Macro-Micro Cloud (MMC) architecture reduces communication overhead by eliminating repeated service discovery for long-lived services through stable micro clouds tied to geographic regions.
• Micro clouds maintain service continuity via dynamic handoff mechanisms, allowing new vehicles to seamlessly take over tasks when existing providers leave.
• Resource availability modeling enables accurate estimation of storage capacity and bandwidth per micro cloud, supporting data placement decisions.
• GPS-based atomicity approximation provides a lightweight alternative to quorum-based consistency, reducing coordination cost in mobile environments.
• The hierarchical structure allows flexible deployment with region-specific protocols, adapting to varying traffic and mobility conditions.
• The framework demonstrates feasibility for location-based data storage in vehicular networks, with potential for integration into smart city and IoT applications.