Middleware For RFID Systems: An Overview: Jameela Al-Jaroodi, Junaid Aziz, and Nader Mohamed
Middleware For RFID Systems: An Overview: Jameela Al-Jaroodi, Junaid Aziz, and Nader Mohamed
Middleware For RFID Systems: An Overview: Jameela Al-Jaroodi, Junaid Aziz, and Nader Mohamed
1. Introduction
A large number of organizations aim to reduce operational costs and increase productivity. However, this requires adapting the business models, enhancing the infrastructure and advancing the technology solutions used. RFID technology offers several advantages to businesses to streamline their work and optimize operations. RFID (Radio Frequency Identification) is generally used to trace tagged objects and is considered as a generation ahead of barcode systems. RFID tags allow for bulk identification of objects and could carry far more information about the tagged object than a barcode could, which allows for better organization and operations based on stored information. Because of the advantage of its reading flexibility, RFID is adopted in a wide range of applications [1,2,3,4]. Industries like Retail outlets, Pharmaceuticals and Manufacturers specially, have shown a great interest in the application of RFID technology into their day to day business operations. The data collection, identification and processing requires some sophisticated software solutions to accompany the RFID system. In addition, as businesses grow and diversify, they require more complex software solutions to support their operations. As a result middleware has become an important component in many types of systems and RFID systems are no exception. Based on the report of Frost
Data Management
This survey provides a general overview of some of the most relevant RFID middleware systems and highlights the modern solutions and goals that still need to be achieved. The paper is structured as follows. Section 2 provides an overview of the current RFID middleware solutions. Section 3 provides the analysis of these solutions and highlights the open research challenges. Section 4 discusses some related work and Section 5 summarizes the contributions of the paper.
2. RFID Middleware
According to Cheng-Ming [8], developing an RFID-enabled application with real time data processing and process decision-making capability is a complicate effort. It requires meeting the following challenges, where such application must be: Flexible to accommodate future changes Flexible to accommodate on-going business rules updates Flexible for process intelligent evolutions Capable of integrating with other new technology Architecture efficient to be able to process high volumes of data intelligently and in real-time Interoperable with variety of devices and systems Highly reliable for mission critical usages Scalable according to the business demands Secure thus protecting the data and supporting access control Compliant with on-going standards Considering the above challenges, the design of an RFID middleware should be capable of addressing issues such as reliability, scalability, load balancing, security and data management. In light of these issues, the remainder of this section surveys most relevant middleware platforms proposed by different researchers.
smoothing the unreliable data captured by RFID readers and a temporal data model proposed in [11] to facilitate application event processing. MARM architecture was implemented using PASSI methodology [12], which is one of the AOSE methodologies for designing and developing MAS (Multi Agent Systems). The architecture is divided mainly in three layers which are device management layer, data management layer and interface layer. The client applications at the interface level specify the data they wish to receive and the kind of data they want to be reported by communicating with the data management agents such as ECReport-Request manager and ECReportSpec Manager respectively. Data management agents in turn communicate with the appropriate reader agents in the device layer, collect the raw data, process it, generate the requested reports and send them all back to the client application. Two ontology class diagrams have been modeled to achieve the objective of successful and timely coordination among agents, one of them is Domain Ontology Description (DOD) and the other is Communication Ontology Description (COD). DOD describes the ontology of middleware design by representing predicates of the domain and some agent actions. COD describes the ontology about the knowledge of agents and the association between the communicating agents. MARM is designed to provide not only the functionalities such as data capturing, smoothing and event processing but also the processing of application level events within the middleware.
2.2 Accada
Accada which has been renamed Fosstrak [13] is an open source RFID platform. Its implementation is based on EPCglobal Network Specifications. Fosstrak is developed after an extensive research work done at ETH Zurich. It provides an RFID infrastructure to the RFID Community benefiting the researchers and application developers with the implementation of lowlevel message bindings with EPC Network components. This fosters the use of EPC Network and availability of modifiable software components according to their requirements rather developing them from scratch. Fosstrak platform consists of three modules: reader, filtering and collection middleware, and EPC Information Services (EPCIS). Reader module performs filtering, synchronous/asynchronous data dissemination as specified in EPCglobal reader protocol and reader management. It can be used in three different modes. In the first mode it wraps a proprietary RFID reader protocol and runs on a separate server. In the second mode, Fosstrak reader is used in simulation mode in which it will make use of
155
its built in simulated RFID readers belonging to different vendors. This can be handy especially in the case no physical RFID reader is available. In the third mode, Fosstrak reader implementation is deployed on the RFID reader itself to provide data filtering and dissemination capabilities. Once the readers capture the relevant tag data, they notify the Fosstrak filtering and collection middleware, which combines the data arriving from different readers in a report that is sent to the subscribed applications according to a predetermined schedule. The interface between filtering and collection middleware and a host application is based on EPCglobal Application Level Events (ALE) specification [14]. Since it also provides the aggregation functionality, it can omit the redundant read events from different readers originating from the same location. EPC Information Services (EPCIS) [15] is a component responsible for receiving the RFID data from the filtering and collection middleware, translating them into business events and making them available to the application. It comprises of three different modules: EPCIS repository, EPCIS capture module which receives the RFID data and EPCIS query module which provides the facility to retrieve events from the repository. The current Fosstrak implementation does not provide persistence for RFID data in the filtering and collection middleware module and it provides limited support for proprietary reader protocols.
be used by variety of different enterprise applications as well as warehouse management and supply chain management systems. The data presentation layer is an application layer which uses the data supplied from the xml framework layer for visualization and decision making. WinRFID is designed using the .NET Framework which facilitates adding functionality to the applications by means of runtime plug-ins. To reap this benefit, WinRFID has an adaptable rule engine therefore end users can easily add their own rules by way of plug-ins.
2.3 WinRFID
WinRFID [16] is an RFID middleware, designed and developed at The University of California, United States. Its architecture is based on five layers: RFID hardware layer, protocol layer, Data processing layer, XML framework, and Data presentation layer. The RFID hardware layer handles the deployment, configuration and communication related issues with readers, tags and the I/O module of its readers. There are three components in this layer: reader, tag and I/O. The protocol layer abstracts the most common readertag protocols such as ISO 15693, ICode, EPC Class 0 and EPC Class 1. It runs a protocol engine whose responsibility is to parse and process the tagged data in the raw format as per any of the standard protocols mentioned earlier. The data processing layer implements the rules to process the raw data coming from the protocol layer by ways of removing the duplicate reads and verifying the tag reads. Any abnormality at this layer is handled through different forms of alerts such as emails, messages and user defined triggers. The clean and formatted tag data then gets forwarded to the xml framework layer to be presented in xml format to pave the way for this data to
2.5 RF2ID
RF2ID (Reliable Framework for Radio Frequency Identification) [18] is designed and implemented with the aspiration of providing a middleware with traits such as reliability, load balancing, high throughput, scalability and data organization. The architecture of the middleware is built on the concept of virtual reader (VR) and virtual path (VP). Each VR is associated with a set of physical RFID readers (PR) in its geographical region. The main responsibilities of each VR are the following: filtering the data coming from the connected PRs and from other VRs, timestamping the data, virtual path management and query management. Virtual Paths (VPs) are logical channels dynamically created by VRs in order to provide functionalities like high scale reliability of the data, efficient query responses and load balancing among the readers.
156
2.7 REFiLL
REFiLL [21] is a light weight filtering and collection middleware platform based on EPCglobal network architecture. It is designed as a programmable environment, which provides flexibility to the developers of RFID applications in the presence of heterogeneous hardware platforms and anecdotal requirements. REFiLL is positioned between the reader virtualization layer and the Application Level Events (ALE). REFiLL engine comprises of filters and output managers. Filters are used to receive input from the readers or other filters. Each filter has the following attributes: input ports, out put ports and event processors. Filters are defined and interconnected with the help of editable XML-based markup files which are then compiled into java code to provide the functionality of collection and filtration of RFID events. Output managers provide the functionality of receiving the processed data from the filters and abstractions to the output systems such as RDBMS, Files, Sockets and Web Services.
most important challenges faced by RFID middleware and ways to cope with them. Most of the middleware solutions discussed in Section 2 are heavily relying on finding ways to handle the processing and management of raw data coming from RFID readers effectively. This is achieved after applying the following steps: removing the duplicate reads, combining the data coming from different readers and efficient data dissemination. Since a typical RFID middleware handles large amounts of data coming from numerous RFID readers and shares it with the enterprise applications, it is necessary to have a reliable system in place which ensures an uninterrupted flow of correct and required RFID data. This is achieved by using different event management techniques. Based on our findings we are providing the comparison of the middleware solutions under discussion in Table1 related to their scalability, reliability, load balancing, security and data management. A typical RFID network can have hundreds of RFID readers attached to it while in some cases this number can increase to thousands. Data coming from all of these readers will gather at a particular point and thus would create a large volume of data. Handling such amounts of raw data can pose a serious challenge to the performance of the middleware. Therefore in order to ease this process a distributed network technique is clearly required in which each node is responsible of gathering the raw data from its designated data source according to its preset load limits. Handling such a large amount of data also poses security threats such as unsafe data processing and privacy threats such as using the RFID data for malicious activities. Threats can also be aroused by humans or the environment in management and technical areas as discussed in [22]. These arguments clearly justify the need to have a middleware system with a protocol level security and message level security. The example of such system is proposed in [23] which deals with message level security such as client authentication, access control mechanism, and security of the messages transported between the enterprise applications and the data providers.
4. Related Work
RFID and its applications have been used for a while now in industry and have proven to facilitate several business processes and simplify and speed up many operations that used to be time difficult and time consuming. For example, in warehousing and inventory control, RFID systems have simplified object registration and tracking throughout its journey within and outside the organization.
157
RFID Reliability Middleware MARM SMURF (Statistical Smoothing for Unreliable RFID data) Accada Implementation of EPC Reader Protocol WinRFD Hybrid Middleware RF2ID LIT Middleware REFiLL Processing Rules Peer-to-Peer Group Communication Model Virtual Readers Implementation of EPC Reader Protocol
Table1. Comparison of RFID Middleware Scalability Load Data Management Security Balancing Multi Agent Systems Not Addressed Temporal data model for Not Addressed RFID data Reader implementation in Subscription of three modes: on a separate Readers server, simulation mode, on RFID Reader Distributed Middleware Not Addressed Modules Multi-ring Peer-to-Peer Network Virtual Paths between Virtual Readers and Physical Readers Common Reader Management Interface Peer-to-Peer systems Path Management EPCglobal Application Level Events Specification Data persistence and Customizable business rule engine Large-scale decentralized event notification infrastructure Name Server and Path Server Subscription feedback mechanism Authentication and Access Restriction Node subscription level only Not Addressed Not Addressed Not Addressed
REFiLL Engine with Lightweight implementation of EPC- programmable middleware RP (EPC Reader Protocol) framework
State-based EPCglobal Application Execution Model Level Events Specification Not Addressed REFiLL Filters
It has become easy to locate, register and follow any object while it is being transferred or used. However, in the research community, the activities are not as intense. Several research teams started investigating the issues and challenges associated with the design, development and deployment of RFID systems. Yet quite few have approached middleware as an effective component to provide a spectrum of abstractions and features for RFID systems. Middleware can be used to support all levels of functionality in an RFID system starting from supporting data collection and management at one end and to facilitating application development on the other. A recent survey [24] was conducted comparing different research work done in academia and industry. This survey mainly focuses on comparing different RFID Middleware systems on the basis of data processing and data integration. It has considered a small scope of mostly commercial projects and did not cover many of the academic efforts in the area thus leaving some of the middleware systems such as Fosstrak and REFiLL behind. Our work is solely done by studying and comparing most relevant RFID Middleware systems proposed by different researchers on the basis of their scalability, reliability, load balancing, security and data management. There are some other RFID Middleware platforms available in the market, which are developed
by different vendors such as SUN, BEA Systems and IBM. Since there are either very limited or no technical publications detailing these products therefore their comparison has not been included in this work.
5. Conclusion
Although RFID is not a new technology, research and development have just picked up the pace in the recent years due to its wide support and application domain especially in supply chain management. Researchers are still trying to find the solutions which can help the RFID Industry to deal with data management issues and performance related issues in a better and more efficient way. In this work we have tried to discuss different RFID middlewares proposed by different researchers. Based on our discussion we have concluded that a scalable, robust, efficient and most importantly secure RFID Middleware is required to meet the challenges faced by RFID Industry.
Acknowledgments
This work is supported by UAEU Research Grant # 01-04-9-11/08.
158
References
[1] Intel to track human activity using RFID to help caregivers: http://www.infoworld.com/d/developer-world/inteldeveloping-video-search-technology-359?page=0,1 (2008). [2] Worlds Third Largest Retailer Completes Warehouse RFID Implementation.: http://www.informationweek.com/story/showArticle.jh tml?articleID=57702741 (2005). [3] Tesco Pushes on with Full-scale RFID Rollout.: http://www.computing.co.uk/news/1160636 (2005). [4] R. B. Ferguson, Logan Airport to Demonstrate Baggage, Passenger RFID Tracking, eWeek, 2006. [5] EPCglobal, a not-for-profit standards organization: www.epcglobalinc.org, Feb. 2009. [6] Frost & Sullivan,"World RFID Middleware Markets," Frost & Sullivan Research Report, 2005. [7] Oh, G.O., D.Y. Kim, S.I. Kim, S.Y. Rhew, A Quality Evaluation Technique of RFID Middleware in Ubiquitous Computing, in proc. International Conference on Hybrid Information Technology (ICHIT'06), pp.730-735, Nov. 2006 [8] Cheng-Ming, An Integrated Software Platform for RFID-Enabled Application Development, in proc. IEEE International Conference on Sensor Networks, Ubiquitous, and Trustworthy Computing (SUTC 2006), Taiwan, June 2006. [9] Massawe, L.V., F. Aghdasi, J.D.M. Kinyua, The Development of a Multi-agent based Middleware for RFID Asset Management System using the PASSI Methodology, 11th Annual Research Seminar of Faculty of Engineering, Central University of Technology, South Africa, 2008. [10] Jeffery, S.R., G. Alonso, and M.J. Franklin. Adaptive Cleaning for RFID data streams, in proc. VLDB Conference, 2006. [11] F. Wang and P. Liu, Temporal Management of RFID Data, in proc. VLDB Conference, Trondheim, Norway, 2005. [12] Chella, A., M. Cossentino, and L. Sabatucci, Tools and patterns in designing multi-agent systems with PASSI, in WSEAS Transactions on Communications, vol 3, no 1, pp.352-358, Jan. 2004. [13] Fosstrak, An open source RFID platform, www.fosstrak.org, Last accessed 2008. [14] EPCglobal, The Application Level Events (ALE) Specification, Version 1.1, www.epcglobalinc.org, Feb 2008.
[15] EPCglobal, EPC Information Services (EPCIS) Specification, Version 1.0.1, www.ecpglobalinc.org, Sep. 2007. [16] B. S. Prabhu et al, WINRFID A Middleware for the enablement of Radio Frequency Identification (RFID) based Applications, white paper, UCLA Wireless Internet for the Mobile Internet Consortium, Jan. 2008. http://www.informationweek.com/whitepaper/Wireless/wp 901137?articleID=901137 [17] Cervantes F. L., Lee Y., Yang H., Lee J., A Hybrid Middleware for RFID-based Parking Management System using Group Communication in Overlay Networks, in proc. International Conference on Intelligent Pervasive Computing, 2007. [18] Nova Ahmed, Rajnish Kumar, Robert Steven French, and Umakishore Ramachandaran, RF2ID: A Reliable Middleware Framework for RFID Deployment, in proc. IEEE International Parallel and Distributed Processing Symposium (IPDPS), 2007. [19] Kabir A., Hong B., Ryu W., Ahn S., LIT Middleware: Design and Implementation of RFID Middleware Based on the EPC Network Architecture, in proc. First LogDynamics International Conference (LDIC 2007), Bremen, Germany, 2007. [20] Carney D, etintemel U, Rasin A, Zdonik S, Cherniack M, Stonebraker M, Operator Scheduling in a Data Stream Manager, in proc. 29th VLDB Conference, Berlin, Germany, pp 838849, Aug. 2003. [21] A.P. Anagnostopoulos, et al., REFiLL: A lightweight programmable middleware platform for cost effective RFID application development, in Journal of Pervasive [22] Kim, J., Yang, C., Jeon, J., 2007, A Research on Issues Related to RFID Security and Privacy, In Integration and Innovation Orient to E-Society Volume 2, for the Seventh IFIP International Conference on eBusiness, e-Services, and e-Society (I3E2007), Oct. 2007, Wuhan, China, pp:412-420. [23] Song J., Kim T., Lee S., Kim H., Security Enhanced RFID Middleware Systems, in World Academy of Science, Engineering and Technology, vol 10, Dec. 2005, ISSN 1307-6884. [24] Quan Z. Sheng, Xue Li, Sherali Zeadally, "Enabling Next-Generation RFID Applications: Solutions and Challenges," Computer, vol. 41, no. 9, pp:21-28, Sept. 2008, doi:10.1109/MC.2008.386.
159