Using Materialized View as a Service of Scallop4SC for Smart City Application Services
Advances in Intelligent Systems and Computing
Smart city provides various value-added services by collecting largescale data from houses and infrastructures within a city. However, it takes a long time and man-hour and needs knowledge about big data processing for individual applications to use and process the large-scale raw data directly. To reduce the response time, we use the concept of materialized view of database, and materialized view to be as a service. And we propose materialized view to be as as service (MVaaS). In our
... n, a developer of an application can efficiently and dynamically use large-scale data from smart city by describing simple data specification without considering distributed processes and materialized views. In this paper, we design an architecture of MVaaS using MapReduce on Hadoop and HBase KVS. And we demonstrate the effectiveness of MVaaS through three case studies. If these services uses raw data, it needs enormous time of calculation and is not realistic. Keywords: large-scale: house log: materialized view: high-speed and efficient data access: MapReduce: KVS: HBase; 1 Smart City and Scallop4SC Smart City and Services The principle of the smart city is to gather data of the city first, and then to provide appropriate services based on the data. Thus, a variety of data are collected from sensors, devices, cars and people across the city. A smart city provides various value-added services, named smart city services, according to the situation by big data within a city. Promising service fields include energy saving , traffic optimization , local economic trend analysis , entertainment , community-based health care , disaster control  and agricultural support . The size and variety of gathered data become huge in general. Velocity (i.e., freshness) of the data is also important to reflect real-time or latest situations and contexts. Thus, the data for the smart city services is truly big data. Due to the limitation of storage, the conventional applications were storing only necessary data with optimized granularity. Therefore, the gathered data was applicationspecific, and could not be shared with other applications.