c 2017 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any cur rent or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
[doi: 10.1109/ZINC.2017.7968650]
Multi Channel Sensor Measurements in Fog Computing Architecture Ilija Radovanovic
Ivan Popovic
School of Electrical Engineering, University of Belgrade Innovation center of School of Electrical Engineering Belgrade, Serbia
[email protected]
School of Electrical Engineering, University of Belgrade Belgrade, Serbia
Dejan Drajic School of Electrical Engineering, University of Belgrade Belgrade, Serbia
resource allocation, scheduling, load balancing techniques and algorithms is available for overcoming the problem, the limited cloud performance, service response and availability is achieved. In order to better cope with dynamically scaling data streams fog computing model was introduced [2]. However, cloud computing concept, since the obvious advantagies in the domain of core IoT services, can be seen as a supplementary computing model to the fog computing.
Abstract — This paper presents the design concept of multi channel based sensing application at the physical edge of IoT network. The approach is based on reference Fog computing model, where sensor devices are connected to the edge fog node enabling the implementation of distributed real-time sensing application. The proposed concept implies that the sensor data processing is moved to the connected fog node, reducing hardware and power requirements of the sensor device. The reference fog node architecture, as well as application layer services for data communication, aggregation and processing are also given.
Fog computing is a novel trend in computing that aims to process data near data source. Fog computing pushes applications, services, data, computing power, and decision making away from the centralized cloud nodes at the IoT core, to the logical extremes of a IoT network. Fog computing extends the traditional cloud computing concept to the physical edge of the IoT network, enabling the creation of refined, time-aware applications and services [1].
Keywords—Fog computing; IoT; Sensor devices; Distributed application; Real-time processes
I. INTRODUCTION The focus of IoT research and development projects is on producing concrete results applicable in everyday life and in processes related to the production of goods and service providing. These efforts enable creation of smart environments, for the benefit of society and individuals, integrating the variety of end-devices and applications [2]. In such environment every object, or simply thing, is accessible through the utilized communication technology, providing or consuming information passed in or out of networked infrastructure of IoT. Thus, in such environment, the available real, digital and/or virtual-derived information is utilized in order to create smart environment providing more efficient usage of energy, storage, transportation, healthcare, etc. The objects themselves can make certain decisions and processing based on data communication with other network-accessible objects. Additionally they can be a part of wider context, as components of more complex services and systems. [5]
The basic elements of the fog computing reference architecture are presented in section II. The design concept of multi channel sensing application according to the fog computing architecture is presented in section III. The final conclusion of the paper is given in section IV. II. FOG COMPUTING ARCHITECTURE The proximity of the fog computing architecture to the end user and end-devices is one of the main characteristics that differentiates fog from cloud computing model [1]. Since the fog infrastructure resides at the physical edge of the IoT network, where cloud is located closer to the IoT core, there is no need for all the data to be transferred toward cloud infrastructure. Therefore, data aggregation is the one of the basic fog computing properties as moving closer to IoT core services. This lead to the decreased data communication throughput requirements resulting in increased processing performance. On the other hand, data processing at the edge of the IoT network enable scalability in the deployment of complex services in the form of incrementally built intelligence as approaching cloud edge.
Cloud computing concept, as a major part of IoT, enable internet-based ubiquitous computing model and on-demand access to shared pool of processing, storage resources according to different available service models. In order to deal with huge amount of data originating from the diversity of sensors and smart things, cloud concept introduces several challenges regarding dynamically scaled processing, data communication and storage requirements. Although different
In the case of multi channel sensor measurements and data processing as a part of distributed real time application, them
1
c 2017 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any cur rent or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
[doi: 10.1109/ZINC.2017.7968650] Additionally the fog computing model offers higher reliability, scalability and flexibility in implementation compared to the cloud computing concept. At the same time it is brings the advantages of cloud computing closer to the data source. Each element in fog reference architecture may represent a hierarchy of fog clusters fulfilling the same functional responsibilities. Depending on the scenario, multiple fog and cloud elements may collapse into a single physical deployment. Fog nodes may securely discover and communicate with each other for exchanging context-specific intelligence [6]. This communication is performed through the Node-to-Node secure communication pathways over the different packet based communication protocols, like COAP, MQTT or SOAP. In the context of the fog to cloud computing space, fog node offer the Node-to-Cloud communication pathways, as a proxy of cloud servers towards their associated frontend devices. While aggregating and representing these frontend devices to the cloud servers, all mentioned pathways at the fog node side shell cooperate to preserve the interoperability among the frontend devices and the cloud core services [6].
both, local data aggregation and data processing, approaches are applicable and supported through the fog computing architecture. The illustration of the fog computing infrastructure and its basic architectural components is presented on the Fig 1.
III. MULTI CHANNEL SENSOR MEASURMENT The introduced concept is focused on the several aspects regarding the integration of sensing devices in distributed measurement application based on reference fog architecture. These aspects include end-to-end data communication, data aggregation and additional multi channel information processing. The individual functionalities are given in the form of configurable group of micro services that reside on fog-node application service layer. Real-time application is given in the form of collection of tightly coupled micro services at fog node side and network accessible data producing services at sensor device side. The loosely coupled integration of sensor-side services is achieved through dedicated active system components utilizing available end-toend communication. The reference fog node architecture for multi-channel measurement application is given on Fig. 2.
Fig. 1. Fog computing infrastructure and basic components
As presented on Fig. 1. front-end devices are connected to the fog infrastructure through the node-to-device communication pathways. Since the legacy devices are supported, various point-to-point and network communication technologies, from USB, UART, RS-485, Modbus, and Ethernet are supported.
At the bottom line of the fog node application service layer is the fog connector service as an interface to a collection of front-end devices. On the device side, the local service management layer (LSM) providing the interface to the corresponding fog node. Device-to-node pathways are given as a collection of end-to-end data communications, where the context of the individual communication is managed through the service agent as a dedicated system level component. This component provides the top-level abstraction of end-to-end communication while maintaining the time-aware communication context. The service agent operation is positioned at the top of the data link layer supporting legacy point-to-point communication technologies. In the case of communication over the packet based IP network, it operates over transport layer regardless reliable or best effort network data transport model is used. From the application point of view, service agents are given in the form of configurable micro services as a part of fog connector operation. The unique service agent forms the particular Device-to-Node pathway, where its configuration includes the communication parameters. The operation of service agent is either time-
The common fog infrastructure is given in the tired hierarchical form. Nodes closer to the IoT network edge are typically focused on sensor data acquisition/collection, data normalization, command and control of sensors and actuators. Their basic functionality in simplified form is realted to the data gathering and aggregation, while providing the basic information processing. As a result the basic level of applicable knowledge is introduced in fog infrastructure. The higher tire fog nodes, closer to the IoT core, are focused on data filtering, compression, and transformation. They may also provide some edge analytics required for critical real time or near real time information processing. As moving away from the true network edge, higher level machine and system learning, eg. analytic capabilities, can be noticed [6]. In general the fog tired organization is introduced in order to deal more efficiently with the huge amount of data that needs to be processed, in the same way providing better operational performance and system intelligence.
2
c 2017 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any cur rent or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
[doi: 10.1109/ZINC.2017.7968650] measured physical property in the exact time instance. In the case of multi channel sensor data processing at the fog node side, data aggregation service is mandatory in order to provide the unique aggregated information sample. The information sample can be interpreted as a collection of multiple sensor data measurement at the given time instance. As shown on Fig. 2. aggregation service resides at application service layer and is implemented as a unique configurable aggregation buffer. The buffer configuration data define the aggregation input channels and output channel event notification condition, as well as callback function at the specified output channel. The event notification in the form of callback function execution upon the specified condition is met provide the triggering mechanism for the processing of aggregated information.
triggered or event-based; this way supporting both, requestresponse and publish-subscribe message exchange pattern with end-device. Similarly in the case of multi channel sensor measurements, each of the sensors is available through end-toend communication, so that the individual communication context is being managed as a part of the fog connector services in the form of the collection of the service agents, as presented on Fig. 2.
In the case of multi channel sensor measurement, the data from the sensor devices is transferred to the data aggregation service which couples relevant data by time of their creation. That information is provided to auto correction service, through event notification mechanism. Correction service accesses and processes information available in data aggregation buffer. The meaning of the correction service in the context of multi channel sensor measurement is usually the compensation of measured physical property from the influence of other environmental properties. At the top of the fog node application service layer are getaway services that enable fog node vertical communication over the Node-to-Cloud pathways enabling information flow towards the IoT core cloud services. IV. CONCLUSION The introduced concept addresses the implementation of the device-to-node communication, data aggregation and processing services at the edge of fog computing infrastructure. It is applicable in different distributed sensing application with time-aware processing requirements. In the context of IoT network this approach results in localization of application deployment, reducing the network communication requirements and end-device processing capabilities. The utilized fog computing infrastructure enables the further integration of locally generated information and knowledge at the IoT network domain supplementing the benefits and functionalities of already deployed cloud computing core services. The presented architectural approach is appropriate in hard real-time or near real-time constrained application that reside at the boundary of the IoT network infrastructure. Utilization of the presented concept for multi channel sensor data measurement increases operational performance, improves end-device energy efficiency, reduces application response time, in the same time providing the framework for building scalable and reliable real-time distributed sensing applications.
Fig. 2. Fog node architecture for multi channel sensing applications
The operation of the sensing devices is independent in terms of acquisition and sampling, while the communication rate, transmission methods are in the competence of the connected fog node. By moving the data path management and the features regarding event processing from the end device node to the edge fog node the hardware and power requirements at the end device side are reduced, while the software is simplified. Beside the fog connection service the different application services are needed in order to convert raw data to more useful information. The application services are able to provide the data aggregation and auto-correction features, prediction services, as well as gateway services to other fog and cloud nodes. Together with application support services and backplane services, application services provide the framework for the implementation of distributed application as a collection of micro services [6].
Acknowledgment The authors gratefully acknowledge financial support from the Ministry of Education and Science, Government of the Republic of Serbia through the Project No. 32043: “Development and modeling of energy efficient, adaptive,
In the case of sensor measurement the obtained information is time dependant and corresponds to the
3
c 2017 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any cur rent or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
[doi: 10.1109/ZINC.2017.7968650] [3]
multi-processor and multi sensor low power electronic systems”.
[4] [5]
References [1]
[2]
G. Eason, B. Noble, and I.N. Sneddon, “On certain integrals of Lipschitz-Hankel type involving products of Bessel functions,” Phil. Trans. Roy. Soc. London, vol. A247, pp. 529-551, April 1955. (references) J. Clerk Maxwell, A Treatise on Electricity and Magnetism, 3rd ed., vol. 2. Oxford: Clarendon, 1892, pp.68-73.
[6]
[7]
4
I.S. Jacobs and C.P. Bean, “Fine particles, thin films and exchange anisotropy,” in Magnetism, vol. III, G.T. Rado and H. Suhl, Eds. New York: Academic, 1963, pp. 271-350. K. Elissa, “Title of paper if known,” unpublished. R. Nicole, “Title of paper with only first word capitalized,” J. Name Stand. Abbrev., in press. Y. Yorozu, M. Hirano, K. Oka, and Y. Tagawa, “Electron spectroscopy studies on magneto-optical media and plastic substrate interface,” IEEE Transl. J. Magn. Japan, vol. 2, pp. 740-741, August 1987 [Digests 9th Annual Conf. Magnetics Japan, p. 301, 1982]. M. Young, The Technical Writer’s Handbook. Mill Valley, CA: University Science, 1989.
