SIaaS - Sensing Instrument as a Service Using Cloud Computing to Turn Physical Instrument into Ubiquitous Service

2012 10th IEEE International Symposium on Parallel and Distributed Processing with Applications

SIaaS - Sensing Instrument as a Service Using cloud computing to turn physical instrument into ubiquitous service Roberto Di Lauro, Francesca Lucarelli and Raffaele Montella Department of Applied Science, University of Naples Parthenope, Italy Email: {roberto.dilauro, francesca.lucarelli, raffaele.montella}@uniparthenope.it Abstract—Sensing Instrument as a Service (SIaaS) is a new cloud service that allows users to use data acquisition instruments shared through cloud infrastructure. It offers a common interface to managing physical sensing instrument and it permits to take all advantages of using cloud computing technology in storing and processing acquired data. With SIaaS different research groups could share physical instrument, in a controlled way, event if they are geographically distributed and user can access to them using an internet connected device without the need to install any sort of program. With the interaction of different own framework we have created a new cloud service for transforming physical resources in a ubiquitous service. Index Terms—Cloud computing, Sensing instrument, GVirtus, JaClouX, Abstract Execute Framework.

high performance computing clouds [3] that provides needed modules to create connection between physical resource and the high level interface. To take full advantage from SIaaS we use JaClouX (JCX) [4]. JCX is an open source, cloud independent, Java API to design, develop and simulate high performance cloud computing application. We have developed a new Driver component to interact with SIaaS cloud service and is able to use its feature in cloud application. The rest of the work is so organized: the section 2 is about Abstract Instrument Framework and GVirtus as main modules of SIaaS; in section 3 we describe SIaas architecture; section 4 is about JaClouX as a component to realize application that use SIaaS; finally conclusions and the highlights on future developments are discussed in section 5.

I. I NTRODUCTION Cloud Computing is an internet-based model in which virtualized and standard resource are provided as a service over the Internet. It provides a minimal management effort or service provider interaction and users interact with a virtual and dynamically scalable set of resources that can manage depending on their needs. Cloud providers differ for the service provided and for the cloud architecture. The main consolidated service models are : SaaS - Software As A Service ; PaaS - Platform As A Service; IaaS - Infrastructure As A Service. We propose a new service model called Sensing Instrument as a Service (SIaaS). It is based on the idea of virtualize sensing instruments and sharing them as a common resource to final users in a controlled way. The main difference with other kind of cloud services is that in our model cloud provider is not the owner of shared resources but it offers the possibility to share and manage a client’s sensing instrument using cloud infrastructure through a virtualization system. We have combined different framework to reach this effort and for the first time we’ve created a system that allow users to share instrument that are not “fundamental computing resources”. The problem of the virtualization of these kind of resources is solved using Abstract Instrument Framework (AIF). It is a Java framework that provides a standard interface for several scientific instruments for data acquisition and allows to handle different kind of instrument through a high level device driver [1]. The connection layer between physical device and the high level interface provided to the user through a virtual machine equipped with AIF is done by leveraging on GVirtus features. GVirtus [2] is our transparent virtualization component for 978-0-7695-4701-5/12 $26.00 © 2012 IEEE DOI 10.1109/ISPA.2012.135

II. A BSRACT I NSTRUMENT F RAMEWORK AND GV IRTUS AS MAIN COMPONENT OF SI AA S Abstract Instrument Framework (AIF) is a software system for the secure sharing and aggregation of data acquisition instruments, geographically distributed, for several scientific and engineering applications. The first goal of AIF was to develop a framework for the virtualization of data acquisition instruments in order to create a common interface focused on interoperable standards components. Our solution is very innovative because it allows the developer to take advantage of a concrete support, in terms of adaptability and reusability, in the instrument virtualization. AIF virtualizes the scientific instruments through the implementation of logical interfaces that provide a high level of interaction common to all instruments. The AIF has been developed using Java programming language and using a modular programming style that allows to extend its functionality in easily way through the implementation of plug-in in order to allow the management of any acquisition instrument. AIF is an intermediate component between real instruments and the software component responsible for the processing and storage data. In according to the philosophy of development any instrument which constitutes the virtual laboratory idealized, is composed by a series of acquisition data sensor and a series of actuators (handlers) for giving commands to the instrument. AIF ha two main component : a frontend and a backend. The frontend is deployed on a virtual machine that act as proxy machine. The frontend is deployed on a virtual machine that acts as 861

instrument through cloud backend on which is deployed AIF backend. This backend, through GVirtus, has access to physical instrument managed by a sink that is a real machine equipped with mandatory driver interact with physical resource. Requests of different clients are handled by the same cloud backend that works as a scheduler granted a specified quality of service and managing propriety. Each instrument could be shared in a controlled way with a specific group of user and all data acquired by instrument are stored trading on cloud storage services. Backend is able to scale resources for managing requests using the elastic computing power offered by cloud provider. IV. JAC LOU X AS A FREMWORK TO PROFIT BY SI AA S

Fig. 1.

JaClouX (JCX) is an open source, cloud independent, Java API for high performance cloud computing application design, development, simulation and evaluation [4]. JCX permits to create cloud-aware applications that could be deployed on several cloud providers. This is made possible thanks to its plug in architecture that allows to manage different cloud provider’s services just implementing a new Driver component. So JCX is fully independent of the underlying cloud software and it can interacts with components as Images, Instances, Locations and Sizes through a common high level interface. JCX’s Driver component provides an abstraction of Amazonlike services , such as Simple Storage Services (S3) and Elastic Block Services (EBS) using respectively the Simple Storage Component (SSC) and the Elastic Block Component (EBC). The Driver component can provide the missing cloud features by software emulation. A special Driver component is creating to managing AIF virtual machines and to create application that uses services offered by SIaaS. DryRain (DR) [4] is the simulator component of JCX. It permits to evaluate costs and performances of a cloud application. Its engine is based on the simulated event approach. The cost profiles are manually loaded or automatically acquired using real cloud interfaces. JCX become a really useful framework to create scalable application using SIaaS features and to simulate their execution before deploying them on a real cloud provider.

SIaaS architecture

proxy machine . The user has access to this virtual machine and through AIF Frontend can interact with the instruments in a transparent way.The AIF backend is deployed on a virtual machine called cloud backend, and acts as a scheduler, resource manager and user permission management. It offers also common interface to acquired data by transforming them in a standard format. The connection between frontend and backend is done with an ad-hoc plugin developed in GVirtus. GVirtuS is born as a tool to fill the gap between in-house hosted computing clusters, equipped with GPGPUs devices, and pay-for-use high performance virtual clusters deployed via public or private computing clouds. GVirtuS allows an instanced virtual machine to access GPGPUs in a transparent way, with an overhead slightly greater than a real machine GPGPU setup [2]. The basic virtualization idea is based on a split driver approach [3] through the utilization of the frontend and backend component mentioned above. In AIF the device exported to the guest domains through the frontend is the virtualized acquisition instrument and backend constitutes the driver modules that receives requests from the frontend library and manage the instrument resource. We have developed a plugins for GVIrtus that intercept the request coming from virtual machine and turns them to the backend that make the real request to the instrument (or proxy machine that manage the instrument). To realize SIaaS we split backend components in two separate ones : a real backend called “sink” deploys on a real machine directly connected to the instrument, and a cloud backend deployed on a specialized virtual machine on cloud that manage virtualized instrument resourced and user’s requests. Let’s see in more detail SIaaS architecture in following section.

V. C ONCLUSION AND FUTURE WORK We have developed a new service model for cloud computing to transform physical resources as sensing instrument in shared ones using virtualization and cloud infrastructure. R EFERENCES [1] “Abstract Instrument Framework”, http://osl.uniparthenope.it/projects/ abstractinstrument/ [2] R.Montella, G. Coviello, G. Giunta, G. Laccetti, F. Isaila and J. G. Blas, “A GPU Accelerated High Performance Cloud Computing Infrastructure for Grid Computing Based Virtual Environmental Laboratory” [3] R. Montella, G. Coviello, G. Giunta, G. Laccetti, F. Isaila and J. G. Blas, “A GPGPU Transparent Virtualization Component for High Performance Computing Clouds” [4] JaClouX - Java bindings for Cloud eXtension, http://osl.uniparthenope.it/ projects/jacloux [5] Abstract Execute Framework - A Java framework for executing external program from code, http://osl.uniparthenope.it/projects/abstractexec

III. SI AA S A RCHITECTURE AIF is the core of SIaaS architecture (1). It is inserted as component of a specific virtual machine images that users can use by instantiating the frontend virtual machine on cloud to interact in transparent way with virtualized acquisition

862