A Middleware Infrastructure for Utility-Based Provisioning of Iot Cloud Systems

A Middleware Infrastructure for Utility-Based Provisioning of Iot Cloud Systems

A Middleware Infrastructure for Utility-based Provisioning of IoT Cloud Systems Stefan Nastic, Hong-Linh Truong, and Schahram Dustdar Distributed Systems Group, TU Wien, Austria Email: [email protected] Abstract—Recently, we have witnessed numerous benefits of for: i) representing the IoT Cloud infrastructure resources, exploiting Cloud Computing models and technologies in the ii) managing their configuration and deployment models, as context of the Internet of Things and Edge Computing. However, well as iii) composing low-level resource components into utility-based provisioning paradigm, one of the most important properties of Cloud Computing, is yet to be realized in emerging usable infrastructures, capable to support novel application IoT Cloud systems. In this paper, we introduce a novel mid- requirements. dleware, which provides comprehensive support for multi-level In this paper, we continue our line of research towards provisioning of IoT Cloud systems. The main features of our utility-based provisioning of IoT Cloud systems by intro- middleware include: i) A generic, light-weight resource abstrac- tion mechanism, which enables application-specific customization ducing a novel provisioning middleware for IoT Cloud. Our of Edge devices; ii) Support for automated provisioning of Edge middleware builds on the previously introduced concepts and resources and application components in a logically centralized frameworks [18], [19], extending them with comprehensive manner, via dynamically managed APIs; and iii) Flexible provi- support for scalable multi-level provisioning of IoT Cloud sioning models that enable self-service, on-demand consumption systems. The main features of our middleware include: i) A of the Edge resources. We evaluate our middleware using real-life applications in the domain of building management systems. generic, light-weight resource abstraction mechanism, based on software-defined gateways, which enables application- I. INTRODUCTION specific customization of Edge devices; ii) Support for auto- mated provisioning of Edge resources and application compo- Recently, Cloud Computing and the Internet of Things nents in a logically centralized manner, via dynamically man- (IoT) have been converging ever stronger, sparking creation aged APIs; and iii) Flexible provisioning models that enable of very large-scale, geographically distributed systems, called self-service, on-demand consumption of the Edge resources. IoT Cloud systems [1]–[3]. IoT Cloud systems intensively The remainder of the paper is organized as follows: Sec- exploit Cloud Computing models and technologies, predom- tion II presents a motivating scenario, background and main inantly by utilizing large and remote data centers, but also research challenges. In Section III, we introduce our middle- nearby Cloudlets [4], [5] to enhance resource-constrained Edge ware and discuss its architecture in detail. Section IV outlines devices (e.g., in terms of computation offloading [6]–[8] and the main runtime mechanism for multi-level provisioning. data staging [9]) or to provide an execution environment for Section V describes experimental results and outlines current cloud-centric IoT applications [10], [11]. prototype implementation. In Section VI, we discuss the One of the main advantages of Cloud Computing is reflected related work. Finally, Section VII concludes the paper and in its support for self-service, on-demand resource consump- gives an outlook of our future research. tion, where users can dynamically allocate appropriate amount of infrastructure resources (e.g., computing or storage) re- II. MOTIVATION &BACKGROUND quired by an application [12], [13]. To date, we have witnessed numerous benefits of this utility-based provisioning model in A. Scenario terms of more flexible and cheaper IT operations [14]. There- Let us consider a case of Building Management System fore, it would be natural to expect that this flagship property (BMS), which provides applications to manage various build- of Cloud Computing would be inherited by IoT Cloud, as ing facilities, such as HVAC systems, elevators and emergency well. Unfortunately, this is still not the case, because current alarms. SAPP is a BMS application developed in collaboration approaches dealing with IoT Cloud provisioning mostly focus with our industry partners1. It is written in Sedona [20] and on providing virtualization solutions for the Edge devices, it is responsible to monitor environmental conditions within such as IoT gateways [15]–[17]. Although device virtualization buildings and to regulate the HVAC facilities accordingly. is one of the preconditions for utility-based provisioning, Figure 1 shows a high-level system architecture of SAPP such approaches are usually intended to support a specific application. The solid arrows show typical propagation of task, e.g., data integration or data-linking, and largely rely sensory data within the SAPP. The dashed elements represent on rigid provisioning models. This inherently prevents con- IoT Cloud infrastructure resources and the dashed arrows suming IoT Cloud infrastructure resources as generic compute or storage utilities and requires rethinking existing support 1http://pcccl.infosys.tuwien.ac.at/ Monitoring data processing Control sequences Stakeholders IoT Cloud Applications Developers/ Repositories Engineers Middleware Platform Operations Software‐defined Device and Managers IoT units Identity orchestration management access management Configuration Domain Experts Communication control models Virtual Virtual Elastic pools of sensors actuators protocols VMs and containers Infrrastructure Virtualization Layer and Business Software Stakeholders artifacts Physical Infrastructure (Edge devices, network elements and data centers) IoT Cloud Environment Monitoring data processing Control sequences Stakeholders IoT Cloud Applications Domain Expert Developers Designers Middleware Platform Stakeholders Virtual Communication VMs and Sensors Protocols Containers Operations Infrastructure Virtualization Layer Developers Managers Sensors/ Data Gateways Actuators Centers Physical Infrastructure Stakeholders Business Operations Stakeholders Managers BMS Application considered during application provisioning. Therefore, con- control sequence suming IoT Cloud infrastructure as a utility and provisioning publish even a simple application such as our SAPP is a challenging get Local Data events Core Application data Processing Business Logic Field Bus Network task. send reading B. Background deployed on deployed on notify Previously we have introduced a conceptual model for Sensors Sensors SensorsIoT Sensors software-defined IoT Cloud systems [18]. The core con- ActuatorSensors SensorSensors Sensors CloudSensors VM Gateway User cept of the provisioning model is software-defined IoT unit. The software-defined IoT units describe IoT Cloud resources Fig. 1. Overview of SAPP architecture. (e.g., virtual sensors), their runtime environments (e.g., IoT gateways) and capabilities (e.g., communication protocols or data point controllers). Such units are used to encapsulate illustrate deployment of main application components. SAPP infrastructure resources and to abstract their provisioning in is highly distributed and executes atop both the Edge devices software. To this end, they expose well-defined APIs and (e.g., IoT gateways) and the Cloud infrastructure. Generally, can be composed at different levels, creating virtual runtime the Edge part of application’s business logic is responsible infrastructures for IoT Cloud applications. The main purpose for obtaining sensory readings from building facilities and of such software-defined IoT Cloud infrastructures is to enable performing initial data processing. The processed data is then utility-based provisioning of IoT Cloud resources by providing transmitted over the network (Wi-Fi or 3G) to the Cloud a uniform view on the entire resource pool, as well as by for further processing. Finally, SAPP’s Cloud services invoke allowing IoT Cloud applications to customize and consume actions, such as to notify a user or to perform remote actuation, those resources dynamically and on-demand. e.g., regulate water pressure. In [19], we have presented a cloud-based provisioning The BMS is a large-scale, geo-distributed system respon- framework for provisioning IoT Cloud applications. The sible for managing hundreds of buildings. It executes atop framework provides a scalable provisioning controller, which complex IoT Cloud infrastructure that includes: (i) various is responsible for managing, deploying and installing appli- Edge devices, such as sensors, actuators and gateways, which cation components in Edge devices. Further, the introduced are installed throughout the buildings, (ii) network elements, framework provides mechanisms to manage the Edge devices, and (iii) cloud services, e.g., for complex event processing, e.g., to register new devices and detect disconnected devices. NoSQL data storage, and streaming data analysis. Since BMS One of the distinguishing features of IoT Cloud systems is service provider is usually not the owner of physical infrastruc- infrastructure virtualization layer and there is a number of ture, e.g., Edge devices installed in buildings, the IoT Cloud existing approaches that deal with Edge devices virtualiza- infrastructure is consumed by BMS as a utility and it needs tion, exposing such devices to the upper layers on different to be dynamically provisioned to satisfy BMS requirements. levels

View Full Text

Details

  • File Type
    pdf
  • Upload Time
    -
  • Content Languages
    English
  • Upload User
    Anonymous/Not logged-in
  • File Pages
    13 Page
  • File Size
    -

Download

Channel Download Status
Express Download Enable

Copyright

We respect the copyrights and intellectual property rights of all users. All uploaded documents are either original works of the uploader or authorized works of the rightful owners.

  • Not to be reproduced or distributed without explicit permission.
  • Not used for commercial purposes outside of approved use cases.
  • Not used to infringe on the rights of the original creators.
  • If you believe any content infringes your copyright, please contact us immediately.

Support

For help with questions, suggestions, or problems, please contact us