An Open-Source Cloud Architecture for Big Stream IoT Applications - Interoperability and Open-Source Solutions for the Internet of Things - page 84

- to provide instantiations of cloud-based and utility-based sensing services
enabling the concept of “Sensing-as-a-Service,” via an adaptive middleware
framework for deploying and providing services in cloud environments.
Another example of this approach is given by the FIware project [ 21 ], an
open cloud-based infrastructure for cost-effective creation and delivery of Inter-
net applications and services. FI-WARE API specifications are public, royalty-
free, and OCCI (Open Cloud Computing Interface)-compliant [ 22 ], driven by the
development of an open source reference implementation which allows develop-
ers, service providers, enterprises, and other organizations to develop innovative
products based on FI-WARE technologies. The FI-WARE solution is based on
the Openstack project [ 23 ], a global collaboration of developers producing an
ubiquitous open-source cloud computing platform for public and private clouds.
The project aims to deliver solutions for all types of clouds by being simple
to implement, massively scalable, and feature-rich. The majority of research
projects listed above addressing their work on Cloud and IoT architectures opt
to use open source components for the implementation of their systems.
Other projects related to real-time and stream management are Apache
Storm [ 24 ] and Apache S4 [ 25 ]. Storm is a free and open source distributed
real-time computation system to reliably process unbounded streams of data.
The system can be integrated with different queueing and database technologies
and provides mechanisms to define topologies in which nodes consume streams
of data and process those streams in arbitrarily complex ways. S4 is a general-
purpose, near real-time, distributed, decentralized, scalable, event-driven, and
modular platform that allows programmers to implement applications for pro-
cessing streams of data. Multiple applications nodes can be deployed and inter-
connected on S4 clusters to create more sophisticated systems. Although there
are several similarities between these systems and the proposed architecture,
such as modularity, scalability, latency minimization and the graph topology,
there are same notable differences. The most relevant use cases for Storm and
S4 are stream processing and continuous computations related to data stored in
databases (e.g., message processing for database update). The proposed archi-
tecture, on the other hand, is specifically designed to work in dynamic IoT
scenarios comprising heterogeneous data sources and making no assumption on
the repositories (if needed) where data can be retrieved or stored. Another major
difference is related to the nature of the topology of the processing units. While
Storm stream management is based on an operator-defined and static topology
of the graph, the proposed architecture is extremely dynamic, as the number of
nodes and edges in the Graph Framework can change according to the workload
and listener requirements.
6 Conclusions
In this paper, we presented a novel Cloud architecture for the management of
Big Stream applications in IoT scenarios. After describing the selected scenario
requirements in terms of decreasing the latency between a point in time when