Dolphin: An Actor-Oriented Database for Reactive Moving Object Data Management DOI: 10.1145/1122445.1122456 Conference: Woodstock ’18: ACM Symposium on Neural Gaze Detection; June 03–05, 2021; Woodstock, NY Woodstock ’18: ACM Symposium on Neural Gaze Detection, June 03–05, 2018, Woodstock, NY Price:
Yiwen Wang, Vivek Shah, Marcos Antonio Vaz Salles, Claudia Bauzer Medeiros, Julio Cesar Dos Reis, Yongluan Zhou · 2025
Read Dolphin: An Actor-Oriented Database for Reactive Moving Object Data Management DOI: 10.1145/1122445.1122456 Conference: Woodstock ’18: ACM Symposium on Neural Gaze Detection; June 03–05, 2021; Woodstock, NY Woodstock ’18: ACM Symposium on Neural Gaze Detection, June 03–05, 2018, Woodstock, NY Price: by Yiwen Wang, Vivek Shah, Marcos Antonio Vaz Salles, Claudia Bauzer Medeiros, Julio Cesar Dos Reis, Yongluan Zhou (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-10063.txt.
Introduction
Yiwen Wang, Vivek Shah, Marcos Antonio Vaz Salles, Claudia Bauzer Medeiros, Julio Cesar Dos Reis, and Yongluan Zhou present Dolphin, a database system for data about moving objects that must also react. Applications from traffic to logistics need objects that sense, move, and trigger responses, yet reactive behavior is usually hand-built by each application on top of spatio-temporal storage. Their answer is a moving actor abstraction with reaction built in.
Moving actors extend the actor model with reactive sensing, movement, and spatial querying, plus defined concurrency semantics. Dolphin layers this design, non-intrusively, on Microsoft's Orleans distributed actor framework. Experiments with realistic reactive scenarios showed the system scaling across machines while keeping reaction latency near real time, sparing developers from balancing performance and consistency by hand.
Catalog record
| CREATOR | Yiwen Wang, Vivek Shah, Marcos Antonio Vaz Salles, Claudia Bauzer Medeiros, Julio Cesar Dos Reis, Yongluan Zhou |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 73 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-10063
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Dolphin: An Actor-Oriented Database for Reactive Moving Object Data Management DOI: 10.1145/1122445.1122456 Conference: Woodstock ’18: ACM Symposium on Neural Gaze Detection; June 03–05, 2021; Woodstock, NY Woodstock ’18: ACM Symposium on Neural Gaze Detection, June 03–05, 2018, Woodstock, NY Price: 15.00 ISBN: 978-1-4503-XXXX-X/18/06 Yiwen Wang, Vivek Shah, Marcos Antonio Vaz Salles, Claudia Bauzer Medeiros, Julio Cesar Dos Reis, Yongluan Zhou Abstract . Novel reactive moving object applications require solutions to support object reactive behaviors as a way to query and update dynamic data. While moving object scenarios have long been researched in the context of spatio-temporal data management, reactive behavior is usually left to complex end-user implementations. However, it is not just a matter of hardwiring reactive constraints: the required solutions need to satisfy tight low-latency computation requirements and be scalable. This paper explores a novel approach to enrich a distributed actor-based framework with reactive functionality and complex spatial data management along with concurrency semantics. Our approach relies on a proposal of the moving actor abstraction, which is a conceptual enhancement of the actor model with reactive sensing, movement, and spatial querying capabilities. This enhancement helps developers of reactive moving object applications avoid the significant burden of implementing application-level schemes to balance performance and consistency. Based on moving actors, we define a reactive moving object data management platform, named Moving Actor-Oriented Databases (M-AODBs) , and build Dolphin – an implementation of M-AODBs. Dolphin embodies a non-intrusive actor-based design layered on top of the Microsoft Orleans distributed virtual actor framework. In a set of experimental evaluations with realistic reactive moving object scenarios, Dolphin exhibits scalability on multi-machines and provides near-real-time reaction latency. 1. Introduction In recent years, we are witnessing a growth in research in data-intensive Internet-of-Moving-Things scenarios involving mobile devices and sensors ( 23 ; 19 ) . For instance, applications such as collaborative transportation systems ( 46 ; 24 ) , fleet management and traffic monitoring ( 26 ; 4 ; 50 ; 69 ; 40 ; 9 ) , tracking of mobile objects and crowd sensing platforms ( 20 ) , etc. , concern objects that continuously move in space and react online to their surroundings. Reactivity is a crucial new requirement for these applications, e.g. , as objects change their states and share information to affect other “surrounding” moving objects. We name such objects reactive moving objects and these applications reactive moving object applications – to characterize the emphasis on surroundings-sensitive object behavior, and the collective participation of such objects in a spatially-sensitive decision-making process. Figure 1 . C-ITS Example Running Example. In Cooperative Intelligent Transportation Systems (C-ITS) ( 54 ; 17 ; 57 ) , vehicles communicate and cooperate towards improving overall transportation effectiveness, driving behaviours, security and safety ( 58 ) . While regular ITS ( 48 ; 41 ; 73 ) concentrate on local information sharing, optimization, and coordination ( e.g. , involving drivers, traffic controllers, or transportation system operators) to make better decisions, in C-ITS, global information ingestion is required and vehicles themselves are active participants in transportation actions and decisions. Figure 1 presents an example of C-ITS. Reactive vehicles (shown as red vehicle icons) start reactive sensing within a specific spatial region (shown as a semi-transparent red range). A reactive vehicle is aware of other vehicles if those vehicles’ movements satisfy its predefined spatial predicates ( e.g ., cross, cover, or overlap) against its spatial sensing area. Sensing areas and predicates may vary from vehicle to vehicle. In addition, reactive vehicles can take reactive actions, such as communication with approaching vehicles to send warnings to help them avoid hazards ( 74 ) . We envision cooperation will be a critical ingredient for self-driving or assisted-driving. ∎ According to 65 as well as 14 , a reactive system needs to update and react to the changes of inputs correspondingly in tight real-time computation constraints. This is challenging to meet in data-intensive settings. In conventional multi-tier distributed data architectures, frequent data shipment is required between the stateless application tier running the application codes and the database tier storing the application state. In addition, system performance would be constrained by the network bandwidth and latency. One might suggest to push the whole logic of a reactive moving object application down into a spatial DBMS. However, reactive features for moving objects are lacking in these systems. Triggers are the primary reactive abstraction available, but they are riddled with complex semantic issues for developers ( 76 ) . Furthermore, there are concerns regarding trigger scalability ( 36 ) , which are only made worse in highly dynamic and distributed environments. Distributed in-memory application architectures have proved to be a viable solution to address the aforementioned performance challenges ( 2 ; 12 ) . These architectures often leverage the actor model as a programming abstraction for middle tier implementation ( 39 ; 44 ) , while employing distributed DBMS for selective persistence of actor state ( 11 ) . Application logic and data are encapsulated into actors, each of which naturally represents the digital twin of a real-world object ( 75 ; 13 ) . The decomposition of application logic and data into actors, as well as the distributed and concurrent actor model execution enable efficient use of computational resources to boost scalability. In light of the success of the actor model in building high-throughput and low-latency applications, we envision it as a compelling implementation abstraction for the application tier in reactive moving object applications – each reactive object, as the digital twin of a real-world physical reactive moving entity, can be modeled as an actor. However, present actor runtimes do not have any built-in support for representing movement of objects in space, for querying their locations, or, most importantly, for specifying reactive behaviors to be triggered in response to the movement of other objects. Developers must face a choice between either: (a) scalably implementing these challenging reactive and spatial functionalities in the application tier; or (b) weaving together a complex web of systems, e.g. , spatial streaming platforms, spatial DBMS, and distributed mid-tier frameworks, with little support to navigate the trade-offs in cross-system data consistency semantics. In this article, we investigate how to design and build an actor-based data platform with high-throughput, low-latency, and reactive spatial data management functionalities. Our solution supports the development of application tiers for data-intensive reactive moving object applications. Our platform alleviates the burden on application developers by allowing them to concentrate on their application logic instead of being distracted by complex spatial data management tasks. In line with the terminology in ( 13 ) , we call such a platform a Moving Actor-Oriented Database (M-AODB). There are several major challenges to achieve this goal. First, it is necessary to provide a proper abstraction for representing reactive moving objects. In this sense, the classic actor model needs to be enriched with features required by reactive moving object applications, including geo-referenced attributes, spatial indexing and querying, spatial-driven reactive behaviors, etc. Second, in the actor model, actors update and query data in a concurrent and asynchronous manner. In our context, concurrency semantics have to be defined, not only to provide proper data consistency guarantees to developers, but also to achieve trade-offs between data latency and