A Real-Time Web of Things Framework with Customizable Openness Considering Legacy Devices

Shuai Zhao, Le Yu, Bo Cheng2016 · DOI

TL;DR

This paper introduces a novel multi-level Web of Things (WoT) framework designed to overcome the limitations of current "siloed" IoT solutions. It enables customizable sharing of user data and resources while ensuring real-time performance, even with heterogeneous legacy devices.

What Problem Does This Paper Address?

The Internet of Things (IoT) is rapidly expanding, but current solutions often suffer from tight coupling between devices and applications, leading to fragmented, proprietary systems. This makes it difficult to reuse resources and build interoperable applications. Existing WoT frameworks, while promising, often lack real-time guarantees, fine-grained data control, and explicit strategies for integrating older, diverse devices. This hinders widespread adoption and practical use, forcing developers to manually bridge complex gaps between the physical and digital worlds.

Methods

The proposed framework consists of two main components: a Public Platform (PP) for cross-domain resource sharing and a Local Platform (LP) for domain-specific, real-time resource management. Key innovations include:

  • Unified Device Access (UDA): Implemented using OSGi and dependency inversion, UDA shields device heterogeneity and exposes them via uniform resource interfaces, simplifying integration.
  • Multilevel and Multidimensional (MM) Model: This model decouples upper applications from underlying devices, allowing sensing data to be interpreted as context information. It also supports flexible control over data granularity and aggregation using a Data Manipulation Language (DML).
  • Customizable Openness: Users can define the degree of openness for their resources, controlling what data is shared, how it's shared (e.g., read-only, time-delayed), and with whom.
  • Multiple Interaction Modes: The framework supports centralized, peer-to-peer, and local modes to cater to different needs and optimize performance.

Datasets

Not discussed in the paper. The paper focuses on a framework and its validation through case studies and experimental evaluations rather than training on specific datasets.

Key Insights and Ideas

  • The necessity of a layered approach: Separating public sharing (PP) from local, real-time management (LP) is crucial for balancing openness with performance.
  • Legacy device integration is a major hurdle: The UDA's ability to adapt to diverse protocols and expose them uniformly is a significant step towards a truly interoperable IoT.
  • Customizable openness is key to adoption: Allowing users to precisely control data sharing, rather than offering binary public/private options, builds trust and encourages participation.
  • The MM model bridges the gap between raw data and context: By providing a structured way to represent resources and their relationships, the framework enables more meaningful data interpretation.

Key Results

The framework's practicability was validated through case studies (District Heating Control and Information System, Coal Mine Comprehensive Monitoring and Early Warning System) and experimental evaluations.

  • UDA Performance: The UDA demonstrated a parsing speed of up to 250 packets per second, with an average of 183.5 packets per second and an average parsing time of 5.45 ms per packet.
  • Response Times: The OSGi-based UDA showed significantly reduced response times compared to traditional Java approaches, especially as the number of bundles increased, indicating its suitability for time-critical applications.
  • Interaction Patterns: The publish-subscribe pattern achieved higher transactions per second (167.6 max, 130.4 average) and lower response times (42 ms max, 30 ms average) compared to request-response (74 TPS max, 64.42 average; 150 ms max, 134 ms average), highlighting its efficiency for event-driven IoT.
  • Data Uploading/Retrieval: The local mode offered stable and significantly lower uploading times compared to Cosm and Yeelink. The P2P mode also showed better performance than Cosm and Yeelink for retrieval, especially for scenarios where data is not readily available on the platform.

Limitations and Future Work

  • Security and Privilege-Based Access Control: These aspects were not discussed in the paper and are subjects of ongoing work.
  • Single Point of Failure: The current framework has a single public platform, which could be a single point of failure.
  • Future Work: The authors envision federating public platforms to create a hierarchical structure similar to domain name servers, distributing them across the internet. They also plan to address security and privilege-based access control.

Summary updated

BibTeX
@misc{zhao2016real,
  author = {Shuai Zhao and Le Yu and Bo Cheng},
  title = {{A Real-Time Web of Things Framework with Customizable Openness Considering Legacy Devices}},
  year = {2016},
  doi = {10.3390/s16101596},
}
Figure