畢業(yè)論文 文獻(xiàn)翻譯.doc

英文資料翻譯系 別 物聯(lián)網(wǎng)技術(shù)學(xué)院 專 業(yè) 物聯(lián)網(wǎng)應(yīng)用技術(shù) 班 級(jí) 物聯(lián)網(wǎng)1102 學(xué)生姓名 程路 學(xué) 號(hào) 100110999 指導(dǎo)教師 錢學(xué)明 2014年 4 月Challenges for Database Management in the Internet of ThingsThis article discusses the challenges for Database Management in the Internet of Things. We provide scenarios to illustrate the new world that will be produced by the Internet of Things, where physical objects are fully integrated into the information highway. We discuss the different types of data that will be part of the Internet of Things. These include identification, positional, environmental, historical, and descriptive data. We consider the challenges brought by the need to manage vast quantities of data across heterogeneous systems. In particular, we consider the areas of querying, indexing, process modeling, transaction handling, and integration of heterogeneous systems. We refer to the earlier work that might provide solutions for these challenges. Finally we discuss a road map for the Internet of Things and respective technical priorities.The Internet of Things (IoT) is a self-styled term to describe objects that are able to communicate via the Internet. Objects range from sensor inputs to actua- tors that control physical objects with new interactions requiring advances in machine and human interfaces. It is widely forecast that these objects will number in the trillions over the next five years of Internet development. Haller et al. have provided the following definition.“A world where physical objects are seamlessly integrated into the information network, and where they, the physical objects, can become active participants in business processes. Services are available to interact with these smart objects over the Internet, query their state and any information associated with them, taking into account security and privacy issues.”Historically, the IoT referred mainly to Radio-frequency identification (RFID) tagged objects that used the Inter- net to communicate. Its origins lie in the manufacturing area, for example, the Auto-id project. The Cambridge Auto-id laboratory produced a number of white articles, journals, and conference articles on the project. RFID is not the only means of connection to the IoT. Wireless sensor networks will provide continuous streams of data on various environmental characteristics, which may be fed into the IoT. Other more sophisticated bridges to the IoT include identification of objects via sensing devices, for example, object recognition via digital imaging. Evermore capable display personal.Taken to its extreme, any everyday object might become part of the IoT and be made intelligent: each book we read, every device in our homes, our pets, every food product, every item of clothing, and even ourselves. Of course one can imagine good and bad scenarios in this vision. It might be convenient to arrive home after a period of absence to a reception of our most comfortable home environment in terms of heating, lighting, and digital entertainment. This environment is set up and maintained economically and energy-efficiently through sensors that capture every environmental detail, making decisions based on their input, knowledge of commod- ity tariff factors, and our preferences. Let us consider another case. We cannot find the book we are reading so we press a button on the screen embedded into our living room wall and in an instance the location of the book is revealed. Or even more futuristically the book detects our arrival home, knows we want to read it, and therefore bleeps to indicate its presence.The IoT offers a wealth of possibilities, but perhaps there is a downside. Our possessions, movements, and activities might be detected by rogue receivers operated by people with intentions to put such information to illegal use. Even overzealous legitimate data collection can seem an inva- sion of privacy. Security systems are not fail-proof. Would we wish surveillance agents to know everything about us, for instance our reading habits, the music we listen to, the food we eat, with whom we associate, and where we are? Such matters need to be carefully considered.3.1 Size, Scale and Indexing The size and scale of the data in the IoT will be vast. Data will need to be managed via responsible local ownership. Local owners will decide which data and services to make available to the global network. Thus, the IoT may operate on more than one level: private and public. Users may join groups for access to certain privately owned data or may, on the other hand, access data publicly available over the public Internet. There may be differ- ences in quality of data depending on ownership and level of care. Gradually trust and reputation systems will provide information to users on the quality of the data. In the global space there will be a need for a central authority, for managing addresses and identifiers, as there is with the current Internet. Indexing will be a major challenge. Finding a particular item in a world where all physical objects have an IP address will not be easy, unless we can devise suitable indexing methods. Work in the library catalog management might provide some point- era on how to do this, but the IoT will encompass many different types of objects. Creating a catalog of everything in the world, readable across countries and languages, is a daunting task. Some objects will be publicly accessible; some will need various levels of access control, and some maybe private to the owner. At first the IoT is likely to develop through local systems that can be indexed coher- ently within a bounded domain. As local systems merge with global systems, new indexing methods will need to be developed. Categories of things will need to be defined together with subcategories. Specialized search sites may provide access to certain categories. For instance, if one wants to find a particular car part, one may go to a search site that specializes in that type of product, and from there be guided to a specific IP address.3.2 Query LanguagesCurrent popular query languages in database systems rely on structured data. Structured Query Language (SQL) is the most prominent example. Over the last few years, however, there have been proposals for query languages for semi-structured data, which is more typical of the data held on the Internet the quantities of data are so vast that it would be unrealistic to expect any sort of uniform structure, except perhaps that of the loosest variety, to be imposed on the IoT. Extensible Markup Language (XML) offers a means of representing less structured as well as structured data, together with some level of self description. It is a well- accepted technology that supports interoperability at a technical rather than a semantic level. XQuery has been developed by the World Wide Web Consortium(W3C XQuery a language for querying XML, can combine docu- ments, web pages, and links to relational databases.Query languages for semi-structured data usually adopt an underlying unidirectional graph. Required objects are specified by providing a path expression in a language that is usually quite intuitive. There are, however, inherent problems with hierarchical data models, such as, difficulty in representing many-to-many relationships. In spite of this, the hierarchi- cal data model has been embraced by the web community as a useful, intuitive, and practical structure. The IoT will have brieflyvisitasitetopickupsomedataorinformation, expert users that know exactly what data they need and where to find it, and users that lie some where in between. In fact in different contexts the same person can be any one of these different types of users. It therefore seems necessary that different types of data access facilities be available. Casual users will need to access the IoT via a user-friendly graphical user interface (GUI), with detailed explanation available on any object, and more flexible, powerful, and efficient access interfaces will be needed for expert users. Services can be used to provide both types of access. Work in query languages for semi- structured data will be relevant for these developments.3.3 Process Modeling and Transactions It is likely that most processes will be developed and supplied as services on the IoT. Service Oriented Architects- ture (SOA) is becoming an important means of supporting interoperabilityinweb-basedsystems.Thecentralideaisthatindependentoutfitsofferservicesinauniform manner, which other users can then take up. Thus imple- mentation details are hidden from the users of the services. Application processes will typically be made up of a num- ber of lower level transactions. Transactions in turn will be made up of lower level operations or services. Therefore, the question of transaction processing in the IoT arises. In the traditional database systems the matter of concurrent transaction processing has been handled through the maintenance of ACID properties through times tamping, locking, and a two-phase commit. ACID properties are atomicity, consistency, isolation, and dura- ability. A transaction must complete in its entirety or not at all, a transaction must leave the database in a consistent state, transactions should not show other transactions, and intermediate results and changes made by a transac- tion must be permanent. In distributed database systems a two-phase commit is used to preserve consistency. All participating sites must confirm their readiness to commit before the commit command is issued by the coordinating site and written to the database log. It has been recognized that the ACID properties do not fit web transaction processing well. This is because the individual web services are essentially autonomous and must independently preserve consistency. This require- meant might conflict with a consistency requirement of a users global transaction. For instance a user sees booking a holiday as a transaction consisting of two operations: booking a flight from one operator and booking a hotel room from another operator. As far as the user is con- cerned the transaction should complete in its entirety or fail. However, the operators of the two booking systems may be independent and one booking may be successful while the other is not. As the underlying systems in this example are independent it is not feasible to impose a two-phase commit procedure. New methods and models have been produced for web-based transaction handling These include the use of compensating transactions and transaction systems that relax various ACID prop- erties . It has been found that the maintenance of ACID properties is not required by all applications. Sometimes, increased throughput of transactions may be a greater priority than preserving ACID properties.3.4 Heterogeneity and Integration Section 2 has outlined the many different types of data that will need to be handled in the IoT. The IoT will furthermore consist of billions of independent nodes, which will have their own systems for holding the data. Interoperability will not be achieved without a standard approach at some level of abstraction. In the context of databases the areas of heterogeneity and integration have been researched since the 1980s, once it was considered useful to achieve interoperability across heterogeneous systems 40-44. Considering that one might have a personnel system stored at one company in a relational database system, and in another company a similar system might be held in a network database system or even a different relational database system, questions arise as to how to integrate such data. Various solutions have been offered . Some promising solutions suggest the use of a canonical data model, for instance a functional or binary data model . However, it seems that often the solutions offered do not warrant the efforts needed to achieve them. Now with the abundance of data and different systems on the web the problems of heterogeneity and interoperability arise anew. XML has played its part in offering a solution to some degree. It has offered a technical, practical, and efficient means to pass data from one system to another. However, XML does not solve the semantic problem. /For instance, does the data item “student” in one system mean exactly the same as data item “student” indifferent- ent system? This question cannot be answered accurately without domain knowledge. Efforts to capture domain knowledge revolve around the concepts of ontology and these mantic web. Onto loges define concepts and the relationships that exist between them. Current work has roots in the Artificial Intelligence (AI) and knowledge representation work of previous decades . OWL (Web Ontology Language) is a family of languages for rep- resenting ontology on the web(W3C2004).The idea is that communities will agree on common technologies and represent these using an OWL system, which will in turn provide the necessary support for semantic interoperable- ity. In the future, agents will play a role in using semantic information to support the improved use of the IoT. Rellermeyeret al. consider the lack of a scalable model to develop and deploy applications atop a het- erogenous collection of ubiquitous devices as one of the biggest challenges in making the IoT a reality. They propose a model based on an extension of the ideas already in use for modular software development. It is likely that the SOA will play an important role in provid- ing a fabric into which the heterogeneous applications of the IoT can be weaved.3.5 Time Series Aggregation Time series aggregation is an interesting area, which has been noted as raising challenges in various applica- tion domains . It has been recognized that inap- propriate time aggregations can give rise to spurious causality. The problem revolves around the ability to select the optimal sampling period for continuous data. Trade-offs include processing time and storage space against accuracy and realistic representation. In the database field, interesting work has been carried out in the stream data capture. It has been recognized that new models were needed for data streams . A number of articles on the topics of database systems, data streams, stream mining, classification, and sum- marization have been produced. It has been recog- nized that traditional query languages such as SQL are not suitable for querying time series data. Other work has considered how missing data, which might occur through exceptions such as power breaks, can be estimated. These developments will be important for intelligent data streaming capture systems in the IoT.In the IoT, the optimal time sampling period will depend very much on the nature of the data and the application area. Suitable querying facilities will need to be defined. These are questions that will need to be addressed by data owners who will offer sampling services on their continu- ous data. Work that has been ongoing in data streaming will contribute to the streaming services of the IoT. An interesting new idea is data-centric middleware for context-aware pervasive computing, where contextual data drives both application behavior and service adap- tation inside the middleware system where sensors are treated as data stream publishers. Adaptable sche- mas could become part of the solution. 3.6 Archiving In recent times there has been much interest in archiving the Internet. A non-profit organization, Internet Archive, was founded in 1996, to build an Internet library, which included an archive of all web pages. The aim of the organization was to offer “permanent access for research- ers, historians, and scholars to historical collections that exist in digital format and to stop digital publications disappearing” 71. The web archive can be accessed via the Wayback machine. As the Internet is so vast, archiving is often done automatically with crawlers, which take copies of web pages at defined intervals. Internet Archive is supported by Alexa Information, the National Science Foundation, the Library of Congress, and other institutions. Database archiving is a long-established technique that involves taking copies of entire databases at specified intervals and keeping them in a secure store. In a web context archived databases can be converted to XML and basic querying can be permitted. Processes can also be archived. This is achieved by using the capture software that takes copies of each request and response entering a service, be that a web service, a database, or any other software system. Such techniques will be applicable to the IoT. Interesting areas arise when we consider the nature of some of the data and the magnitude of the IoT. The solution is likely to be local management of archived data with good indexing and discovery facilities. In the archived data, the main operation will be retrieval, update will only be necessary in exceptional cases. This simplifies the problem space. Interesting areas will revolve around efficient storage, querying, and perfor- mance. Work in data warehousing and data mining may offer some directions. 3.7 Data Protection The IoT is likely to hold much more personal information than is held on the present Internet. Furthermore, access to such information technically is borderless from the point of view of national boundaries. Many countries such as the UK have Data Protection laws. The UK Data Protection Act requires all organizations which handle personal information to comply with a number of important principles regarding privacy and disclosure. Unless exempted, all UK data controllers of personal information have to register with the Information Com- missioners Office. Countries in the EC operate along principles similar to the UK law, the main principle of which is that dat