4
Role of Clustering in Discovery Services for the Semantic Internet of Things

In the evolving field of the Internet of Things (IoT), one of the prime requirements is to empower heterogeneous devices with the capability of searching, identifying and communicating with each other. Standardization body ITU-T too has recognized the importance of cross-technology communication among diverse devices and recommended the inclusion of the directory service function in IMT-2020 architecture. In order to cater to the complex requirements, such as huge number of devices, heterogeneity, various protocols, constrained resources and dynamicity in the IoT domain, discovery services have been proposed in the literature. The chapter presents a comprehensive review of this literature and also explores how clustering algorithms can address some of the prevalent issues like scalability, while designing and developing discovery services, particularly, semantic-based discovery services.

4.1. Introduction

In recent years, the number of connected devices and objects has increased enormously (Al-Fuqaha et al. 2015) and this trend is expected to continue in the future. It is forecasted that more than 1.25 billion devices will be connected to the Internet by 2030 (Marletta 2019), primarily because the importance of IoT has been realized in diverse fields ranging from healthcare, logistics, smart transportation, smart education and connected cars to smart grids (Al-Sarawi et al. 2020). Almost all aspects of life are expected to transform with the use of the IoT. The prime objective of the IoT is to connect all things in the physical world to the Internet. The terms “things”, “objects” and “devices” are used interchangeably in IoT literature. Connectivity of all things to the Internet gives rise to many challenges, such as heterogeneity, scalability and dynamicity (Atzori et al. 2010). Among these problems, heterogeneity of devices and protocols restricts interoperability. The lack of adoption of a global standard further adds to the problem.

One of the possible approaches to solve the problem of heterogeneity and interoperability in the IoT domain is by building discovery services. The prime task of discovery services is to locate a device offering a specified service in the network. The very idea of discovery services is not new and has its origin in the web. However, traditional web-based discovery service solutions cannot be applied to the IoT network because of its very nature of connecting diverse objects in dynamic environments, as well as the presence of a wide range of protocols and technologies (Sharma 2019). The simplest approach for implementing a discovery service is to create a single centralized repository or directory of all the devices and the services offered by them. Any device discovery or service discovery request is then processed at the repository. This not a feasible solution for IoT as it is envisioned to connect billions of devices. Moreover, the dynamic nature of the IoT devices is also a hindrance. Hence, distributed discovery services are expected to provide a better solution.

The directory-based discovery service function has also been included in the IMT-2020 architecture by standardization body ITU-T for heterogeneous devices management (ITU-T 2019). A usecase from ITU-T (2019) is shown in Figure 4.1, where a car (Car A) reaching a traffic crossing prepares a message consisting of its position, speed, direction, etc., and sends it to another car (Car B). This is to alert Car B about the arrival of Car A at the crossing so that it can apply the required safety measures and any occurrence of an accident can be avoided. Car B then validates the message from Car A using the directory-based discovery service function. Such directory services are deployed near intersections to reduce delays. The directory service stores the records of all approaching vehicles. Car B queries the directory service, which in turn authenticates the identity of Car A and sends a response back to Car B. Upon successful authentication of Car A, Car B takes safety action such as stopping at the intersection. This usecase also highlights some key requirements of IoT directory-based discovery services. The two cars can be of any make or manufacturer and hence are heterogeneous objects. The two cars may support different communication technologies and protocols. The usecase also represents a dynamic environment, as vehicles approaching the intersection appear and disappear frequently, hence the directory needs to be updated continuously.

There are various types of discovery services presented in the literature. Some authors have classified them into two broad categories, namely, directory-based and directory-less (Marino et al. 2019), while others have grouped them into six, namely, distributed and P2P discovery services, centralized architecture for resource discovery, Constrained Application Protocol (CoAP) based service discovery, semantic-based discovery, search engine for resource discovery and utilization of ONS (Object Naming Service) and DNS (Domain Name System) (Datta et al. 2015). In a recent work, Zorgati et al. (2019) have categorized discovery services in the IoT into two, namely, protocol-based and semantic-based. Irrespective of the categorization of discovery services, their role in achieving cross-technology communication and thus interoperability cannot be overlooked.

Schematic illustration of usecase highlighting the importance of directory-based discovery services for connected vehicles in I M T-2020 architecture.

Figure 4.1. Usecase highlighting the importance of directory-based discovery services for connected vehicles in IMT-2020 architecture

(source: adapted from ITU-T (2019))

In this context, efforts have been made to apply Semantic Web technologies (SWT) to the IoT, leading to the emergence of the Semantic IoT (SIoT) (Rhayem and Gargouri 2020). It is expected that SWT can solve the problem of heterogeneity and interoperability. While designing new discovery services of any type, another issue relates to the handling of an enormous number of devices, services or resources in the IoT. The importance of clustering in the IoT was first recognized by Sharma et al. (2011). Later on, Kapoor and Sharma (2019) have shown how clustering can reduce the search time in directory-based discovery services for the IoT. Chirila et al. (2016) also propose a recommendation system for device discovery, where they have utilized clustering methods to reduce search space.

This chapter presents a detailed discussion on various types of discovery services in the IoT and the role of clustering therein. The chapter is organized into six sections. Section 4.2 provides an overview of various basic categories of discovery services in the IoT. An alternate method to implement discovery services is through ontologies and semantics, which is discussed in section 4.3. The role of clustering methods in improving the search of discovery services is outlined in section 4.4. The IoT specific requirements from clustering methods and a comparative analysis of some of the popular clustering methods in fulfilling these requirements is presented in section 4.5. Finally, section 4.6 contains the conclusion and future directions on improving the efficiency of discovery services in the IoT.

4.2. Discovery services in IoT

In this section, various classes of discovery services in the IoT are described. Typically, discovery services can be centralized or distributed. Because of the dynamicity and scalability requirements of the IoT, distributed discovery service architectures are being investigated more. A point worth mentioning is that the Internet Engineering Task Force (IETF) has standardized Service Location Protocol (SLP) for locating a service in both directory-based and directory-less discovery architectures (Veizades et al. 1999).

4.2.1. Directory-based architectures

In this architecture, information about devices and services is kept in a repository, called a directory. The devices and service providers register to the directory whereas service consumers interrogate it. Even though centralized directory-based discovery service architectures are not very suitable for the IoT, for completeness, a brief discussion on them is provided in the following subsection.

4.2.1.1. Centralized directory

A single directory of all devices and services is created in centralized directory architecture. Owing to the simplicity and ease of implementation, there is interest from researchers in exploring centralized directory architecture in the context of the IoT.

Jini (2016) allows devices supporting Java to discover each other through a centralized discovery service architecture. The service provider devices in Jini architecture publish services to a lookup server, which can be interrogated by the service consumer devices. The Jini discovery services architecture can be applied in a local network. However, there are global directory-based discovery service architectures as well, such as ONS (ONS 2013). It is a DNS-based network that allows us to search and obtain information about Electronic Product Code (EPC) enabled devices. The standard DNS query can be initiated by encoding EPC into a proper Fully Qualified Domain Name (FQDM). An implementation of ONS service directory is by GS1. Another notable work is Digcovery by Jara et al. (2014) for smart cities, allowing devices supporting heterogeneous technologies to discover and communicate using its centralized global discovery service architecture. It provides a REST API interface for its lookup service. A search engine like discovery framework has been proposed in Datta and Bonnet (2015) using a CoAP resource discovery mechanism. It is worth highlighting here that CoAP allows devices to find resources via a centralized resource repository (Shelby et al. 2014). A CoAP resource directory server provides a common entry by a specific URI “well-known/core”. It replies with the details of hosted resources in the CoRE Link Format (Shelby 2010). In Datta and Bonnet (2015), a new URI “well-known/servers” has been introduced to retrieve the list of CoAP servers that are reachable, so that a hierarchy of linked CoAP servers can be created and the global discovery of resources and services can be enabled.

Since the centralized directory discovery service architectures fail to provide required scalability for IoT and are also prone to failure, distributed directory-based discovery service architecture is discussed next.

4.2.1.2. Distributed directories

A novel distributed discovery service architecture has been proposed by Sharma (2019), which is presented in Figure 4.2. The distributed directories (called DSN in this work) are constructed based on logical attributes characterizing devices. The service provided by a device is also its one such attribute. Thus, this architecture is capable of handling heterogeneity. The DSNs are responsible for establishing communication with the devices under control following protocols supported by them. The DSNs can be connected to each other over the Internet or any other wireline or wireless network. The devices are required to know the identity of the nearest DSN. A service consumer device sends a query consisting of the required service to the nearest DSN, which in turn makes an intelligent decision by sending the query to the most probable DSN if the queried service is not available with it. A probabilistic flood search algorithm is implemented in all DSNs to resolve the queries intelligently. The details of the architecture and algorithm can be found in Sharma (2019).

An altogether different mechanism of distributed directory-based discovery service is proposed in Kozat and Tassiulas (2002), where a network of directory nodes is created using a Minimum Dominating Set (MDS) algorithm. Service provider devices advertise services to backbone nodes. The backbone node that receives a query forwards it to others in case the query cannot be satisfied locally.

Liu et al. (2016a) discuss many distributed discovery architectures, also mentioning the limitation of distributed architecture as it leads to a high volume of network traffic when compared to centralized architectures.

4.2.1.3. Distributed P2P directories

Distributed Hash Table (DHT)-based directories are also studied in the literature. DHT is a distributed hash table data structure (Balakrishnan et al. 2003) with its origination in computer networks, and is widely used in P2P networks for the dissemination of information. In DHT-based architectures, gateways act as directory nodes, tracking devices entering and leaving the network.

Paganelli and Parlanti (2012) have presented a DHT-based P2P discovery service architecture. In DHT, the key space is partitioned among nodes thus requiring knowledge of the key a priori. If there is a change in the key set then the whole hash table needs to be reconstructed. This is a major limitation in dynamic IoT environments as whenever a new device is introduced in the network, the hash table needs to be reconstructed. A distributed resource discovery (DRD) architecture is proposed in Liu et al. (2013), where peers are responsible for both service or resource registration and discovery. There are three components in a peer – the resource registration component is responsible for registering devices resources by storing their IP address, resource path, resource type, content type, etc.; the resource discovery component performs the task of look-up based on the description in the request; and a proxy layer handles CoAP and HTTP messages.

Schematic illustration of attributed-based distributed directory discovery service architecture.

Figure 4.2. Attributed-based distributed directory discovery service architecture

(source: adapted from Sharma (2019)).

4.2.2. Directory-less architectures

The directory-less discovery service architectures are much simpler than directory-based architectures as they do not require dedicated directory nodes. The service provider devices simply advertise their services, while the service consumer devices broadcast the requests. Universal Plug and Play (UPnP) is a well-known industry standard of directory-less architecture (Balakrishnan et al. 1989). Another example is Bluetooth SDP by BluetoothSIG (1988), in which a service is represented by a set of attribute-value pairs. A query consisting of attributes is broadcasted. The devices with matching descriptions respond.

The main challenge in directory-less architectures is that a lot of bandwidth and energy of the devices get wasted in finding the appropriate frequency of the advertisements. Several solutions are proposed in the literature to reduce this wastage (Ratsimor et al. 2002; Gao et al. 2004; Chakraborty et al. 2006; Lee et al. 2006; Nguyen and Aggarwal 2018).

4.3. Semantic-based architectures

An alternate method to implement directories is using ontology and semantics. This has become an integral part of the SIoT.

Zhou and Ma (2012) present a proof of concept of vehicular sensors ontology and an algorithm to find appropriate web services by computing matching degree using semantic similarity and relativity. Alam and Noll (2010) propose a semantic-based framework using service advertisements by IoT devices to accelerate the service registration. The advertisement contains service metadata such as name, identifier, endpoint, location and semantic annotation link. A Semantic Web based service discovery middleware is proposed in Liu et al. (2016b), which uses sensor data. A semantic-based discovery service, called DiscoWoT, is proposed in Mayer and Guinard (2011) for the Web of Things (WoT). DiscoWoT applies a mapping scheme internally to semantically discover services. It uses JSON to represent services semantically. This ensures interoperability between devices. Gomes et al. (2019) developed an ontology-based multi-repository system, QoDisco, that can handle queries spanned over multiple attributes and range. Many optimizations are also discussed to reduce the time of semantic search in this work.

The literature on semantic-based discovery service architectures in the IoT can be broadly divided into two classes – search engine based and ONS DNS based.

4.3.1. Search engine-based

The search engine utilizes a lot of semantic matching algorithms and the same idea can be used in designing discovery services in the IoT. In Ding et al. (2012), a hybrid three layer search engine is proposed that supports spatio-temporal, value and keyword searches. This search engine enables data generated by IoT devices to be searched for. Hence, it has a sensor and device monitoring layer that connects to physical devices and collects data, a storage layer that is responsible for storing data and an index layer that allows searching on the three criteria mentioned above. It is shown in this work that keyword search gives the best performance when searching real-time sensor data. This work does not provide device or service searches. However, it shows the possibility of using search engines for device and service discovery.

4.3.2. ONS DNS-based

ONS stands for object name service and it takes advantage of DNS to find information or locate a service, usually for EPC devices.

A distributed information system using ONS is presented in Minbo et al. (2013). The ONS-based lookup service provides a mapping between a product code and IoT system resource address, whereas DNS stores the related data. The system is shown to have application in agricultural products.

The focus of most of the proposed discovery service architectures is on improving the performance by reducing the search time, handling heterogeneity and interoperability and returning quality results. The issue of scalability is not much addressed in the literature of discovery services. The next section discusses some studies that suggest using clustering to develop scalable architectures.

4.4. Discovery services and clustering

The very idea of applying clustering methods in the IoT was proposed by Sharma et al. (2011). In the context of discovery services, the idea of forming clusters of services based on physical and semantic closeness is used in Klein et al. (2003). Each cluster has a Service Access Point (SAP) that handles registration and queries for that cluster. The advantage of organizing services in clusters is that a query can be quickly narrowed down and answered by search space reduction. In Schiele et al. (2004), clusters are created based on the mobility pattern of the devices. A cluster head is chosen periodically and handles queries. Chirila et al. (2016) proposed a broker-based service discovery and recommendation system for IoT devices. A new service clustering algorithm is used in the recommendation system. The clustering method basis is a new similarity metric.

Noting that the IoT is a dynamic environment and service discovery is not a trivial task there, Fredj et al. (2014) emphasize the need for re-clustering. They also propose an approach based on the hierarchy of semantic gateways to fasten the discovery of IoT Semantic Web services. An incremental clustering algorithm is used to group similar services. An optimized clustering-based framework for discovering services is presented in Bharti and Jinal (2021) using Web Ontology Language (OWL) for efficient semantic matching.

The following section illustrates some IoT specific requirements for clustering methods and provides an overview of the HiCHO clustering method.

Schematic illustration of requirements for designing clustering methods in the I o T and comparative analysis of some existing algorithms.

Figure 4.3. Requirements for designing clustering methods in the IoT and comparative analysis of some existing algorithms

(source: adapted from Kapoor et al. (2014)).

4.5. Clustering methods in IoT

Heterogeneity, dynamicity and scalability are the key requirements in the IoT domain. Considering them, an attribute-based clustering method for IoT devices was first proposed in Sharma et al. (2011). Using the logical attributes of devices makes the method independent of any technology. As noted by Sharma et al. (2011), the set of logical attributes may change over time and the classification mechanism must allow the modification of attributes.

The following definitions are given in Sharma et al. (2011):

Both attributes and values are strings and form an attribute-value or av-pair together.

Using these definitions, a hierarchical, incremental, online clustering algorithm HiCHO was developed in Sharma et al. (2011) with two levels – Level 0 clustering based on attributes and Level 1 clustering based on attribute-values. The performance of this algorithm has also been analyzed in detail, and both the feasibility and accuracy are ascertained.

There are some requirements specific to clustering in the IoT domain, which are discussed in Kapoor et al. (2014). The same are shown in Figure 4.3. Some existing clustering methods are also compared and analyzed against these requirements. According to the discussion in Sharma et al. (2011) and Kapoor et al. (2014), these requirements are as follows:

– Online: in order to support the dynamicity of the IoT environment, it is essential that the clustering method is online. Particularly, as the clustering has to be performed using logical attributes to address the issue of heterogeneity and the value of these attributes may change with time, also, in the worst case even the attribute may change.

– Incremental: the majority of the IoT devices in the future will be mobile. The future devices are not known; hence, the clustering should be incremental in nature. This is in contrast to traditional data mining where the complete dataset is known.

– Attributes and their values based: as the clustering is to be used in discovery services, the devices should be classified in groups such that the search time gets minimized. The devices or services can be looked up by specifying complete or partial CS. If the clustering is performed in multiple levels, that is hierarchy, first by using attributes and then by using av-pairs, then the search domain gets reduced quickly at the first level itself. This leads to a faster discovery of service or device.

– Capable of working with numeric as well as categorical data: attributes are categorical, however, values can be numeric or categorical. The clustering method then should be capable of handling both types of data.

The current literature on clustering does not fulfill the above requirements collectively. For instance, the work of Friedman (2004) is based on attributes, but is an offline algorithm and only deals with numeric data. The real-time clustering algorithm (real-time OPTICS) by Shao et al. (2010) and data stream clustering by Wang et al. (2008) require training before they can be deployed. Real-time OPTICS is density-based clustering, that is, clusters are built around core objects using core-distances, and thus only work with numeric data. Stream data clustering (Wang et al. 2008) uses entropy measures to create micro-clusters that are later updated in the online process when data arrives as a stream. It should be noted that entropy is a measure of uncertainty. A cluster with less entropy is more dense compared to one with more entropy. The goal of entropy-based clustering algorithms is to divide the dataset into groups, such that the entropy of entire system (i.e. collection of all the clusters) gets minimized. There are also many clustering algorithms for categorical data (Li et al. 2004), but they are mostly not incremental and are not online.

These requirements can be treated as guidelines to design new clustering algorithms in the IoT.

4.6. Conclusion

The chapter presents an extensive discussion of various discovery service architectures in the IoT. The existing literature on discovery services can be broadly divided into three types – directory-based, directory-less and semantic-based. Semantics play a pivotal role in the IoT leading to the SIoT, and help in building more intelligent systems. It has been realized that the introduction of the clustering of services, devices, as well as other resources while designing discovery services can reduce search space and thereby lookup time. A set of guidelines while designing new clustering algorithms for the IoT are also discussed.

The future research on discovery services must also include new clustering algorithms considering the discussed guidelines to improve their performance.

4.7. References

  1. Al-Fuqaha, A., Guizani, M., Mohammadi, M., Aledhari, M., Ayyash, M. (2015). Internet of things: A survey on enabling technologies, protocols, and applications. IEEE Communication Surveys & Tutorials, 17(4), 2347–2376.
  2. Al-Sarawi, S., Anbar, M., Abdullah, R., Al Hawari, A.B. (2020). Internet of Things market analysis forecasts, 2020–2030. 4th World Conference on Smart Trends in Systems, Security and Sustainability (WorldS4), London.
  3. Alam, S. and Noll, J. (2010). A semantic enhanced service proxy framework for internet of things. IEEE/ACM International Conference on Green Computing and Communications & International Conference on Cyber, Physical and Social Computing, Hangzhou.
  4. Atzori, L., Antonio, I., Morabito, G. (2010). The Internet of Things: A survey. Computer Networks, 54(15), 2787–2805.
  5. Apache River (2016). DJ – JiniTM Discovery & Join Specification [Online]. Available at: https://river.apache.org/release-doc/current/specs/html/discovery-spec.html [Accessed 28th November 2021].
  6. Balakrishnan, H., Kaashoek, M.F., Karger, D., Morris, R., Stoica, I. (1989). Universal plug and play: Background [Online]. Available at: http://www.upnp-hacks.org/upnp.html [Accessed 1st November 2021].
  7. Balakrishnan, H., Kaashoek, M.F., Karger, D., Morris, R., Stoica, I. (2003). Looking up data in P2P systems. Communications of ACM, 46(2), 43–48.
  8. Bharti, M. and Jinal, H. (2021). Optimized clustering–based discovery framework on Internet of Things. The Journal of Supercomputing, 77, 1739–1778.
  9. Bluetooth (2020). Bluetooth Core Specifications Version 5.2 Feature Overview [Online]. Available: https://www.bluetooth.com/wp-content/uploads/2020/01/Bluetooth_5.2_Feature_Overview.pdf [Accessed 15th November 2021].
  10. Chakraborty, D., Joshi, A., Yesha, Y., Finin, T. (2006). Toward distributed service discovery in pervasive computing environments. IEEE Transactions on Mobile Computing, 5(2), 97–112.
  11. Chirila, S., Lemnaru, C., Dinsoreanu, M. (2016). Semantic-based IoT device discovery and recommendation mechanism. IEEE 12th International Conference on Intelligent Computer Communication and Processing (ICCP), Cluj-Napoca.
  12. Datta, S.K. and Bonnet, C. (2015). Search engine based resource discovery framework for Internet of Things. IEEE Global Conference on Consumer Electronics, Osaka.
  13. Datta, S.K., Costa, R., Bonnet, C. (2015). Resource discovery in Internet of Things: Current trends and future standardization aspects. IEEE 2nd World Forum on Internet of Things (WF-IoT), Milan.
  14. Ding, Z., Xu, G., Guo, L., Yang, Q. (2012). A hybrid search engine framework for the internet of things based on spatial-temporal value-based and keyword-based conditions. IEEE International Conference on Green Computing and Communications (GreenCom), Besançon.
  15. Fredj, S.B., Boussard, M., Kofman, D., Noirie, L. (2014). Efficient semantic-based IoT service discovery mechanism for dynamic environments. IEEE 25th Annual International Symposium on Personal, Indoor, and Mobile Radio Communication (PIMRC), Washington, DC.
  16. Friedman, J.H. (2004). Clustering objects on subsets of attributes. Journal of Royal Statistical Society, 66(4), 815–849.
  17. Gao, Z., Yang, X., Ma, T., Cai, S. (2004). RICFFP: An efficient service discovery protocol for manets. Embedded and Ubiquitous Computing EUC 2004, 3207, Springer, Berlin, Heidelberg.
  18. Gomes, P., Cavalcante, E., Batista, T., Taconet, C., Conan, D., Chabridon, S., Delicato, F.C., Pires, P.F. (2019). A semantic-based discovery service for the Internet of Things. Journal of Internet Services and Applications, 10 [Online]. Available at: https://doi.org/10.1186/s13174-019-0109-8.
  19. ITU-T (2019). ITU-TY.3074 framework for directory service management of large numbers of heterogeneously-named objects in IMT-2020.
  20. Jara, A.J., Lopez, P., Fernandez, D., Castillo, J.F., Zamora, M.A., Skarmeta, A.F. (2014). Mobile digcovery: Discovering and interacting with the world through the Internet of Things. Pervasive and Ubiquitous Computing, 18, 323–338.
  21. Kapoor, S. and Sharma, S. (2019). Dynamic directory of objects using based on logical attributes. US Patent, US010423608B2.
  22. Kapoor, S., Sharma, S., Srinivasan, B. (2014). Clustering devices in an Internet of Things. US Patent, USOO8671099B2.
  23. Klein, M., Konig-Ries, B., Obreiter, P. (2003). Service rings – A semantic overlay for service discovery in ad hoc networks. International Workshop on Database and Expert Systems Applications, Prague.
  24. Kozat, U.C. and Tassiulas, L. (2002). Service discovery in mobile ad hoc networks: An overall perspective on architectural choices and network layer support issues. Adhoc Network, 2, 2–44.
  25. Lee, C., Helal, S., Lee, W. (2006). Gossip-based service discovery in mobile ad hoc networks. IEICE Transactions on Communications, E89-B, 2621–2624.
  26. Li, T., Ma, S., Ogihara, M. (2004). Entropy-based criterion in categorical clustering. Proceedings of the 12th International Conference on Machine Learning, Association for Computing Machinery, New York.
  27. Liu, M., Leppanen, T., Harjula, E., Ou, Z., Ylianttila, M., Ojala, T. (2013). Distributed resource discovery in the machine-to-machine applications. International Conference on Mobile Ad-Hoc and Sensor Systems, Hangzhou.
  28. Liu, L., Antonopoulos, N., Zheng, M., Zhan, Y., Ding, Z. (2016a). A socio ecological model for advanced service discovery in machine-to-machine communication networks. ACM Transactions on Embedded Computing Systems, 15(2), 1–28.
  29. Liu, M., Yang, X., Hu, H., Mohammed, A. (2016b). Semantic agent-based service middleware and simulation for smart cities. Sensor, 16(12) [Online]. Available at: https://doi.org/10.3390/s16122200.
  30. Marino, F., Moiso, C., Petraccaa, M. (2019). Automatic contract negotiation, service discovery and mutual authentication solutions: A survey on the enabling technologies of the forthcoming IoT ecosystem. Computer Networks, 148, 176–195.
  31. Marletta, M. (2019). Life after graduation: Iot: Forecasts, challenges and opportunities. IEEE Instrumentation & Measurement Magazine, 22(6), 76–77.
  32. Mayer, S. and Guinard, D. (2011). An extensible discovery service for smart things. WoT ’11: Proceedings of the2nd International Workshop on Web of Things, Association for Computing Machinery, New York.
  33. Minbo, L., Zhu, Z., Guangyu, C. (2013). Information service system of agriculture IoT. Automatika-Journal for Control Measurement Electronics Computing and Communications, 54(4), 415–426.
  34. Nguyen, P.T. and Aggarwal, A. (2018). Enhanced DNS-based service discovery in an Internet of Things (IoT) environment. US Patent, 9906605.
  35. Object Name Service (2013). Object Name Service (ONS) [Online]. Available at: https://www.gs1.org/standards/epcis/epcis-ons/2-0-1 [Accessed 2nd December 2021].
  36. Paganelli, F. and Parlanti, D. (2012). A DHT-based discovery service for the Internet of Things. Journal of Computer Networks and Communications [Online]. Available at: https://doi.org/10.1155/2012/107041.
  37. Ratsimor, O., Chakraborty, D., Joshi, A., Finin, T. (2002). Allia: Alliance-based service discovery for ad-hoc environments. International Workshop on Mobile Commerce, Atlanta.
  38. Rhayem, A., Mhiri, M.B.R., Gargouri, F. (2020). Semantic Web technologies for the Internet of Things: Systematic literature review. Internet of Things, 14 [Online]. Available at: https://doi.org/10.1016/j.iot.2020.100206.
  39. Schiele, G., Becker, C., Rothermel, K. (2004). Energy-efficient cluster-based service discovery for ubiquitous computing. EW 11: Proceedings of the 11th Workshop on ACM SIGOPS European Workshop, Association for Computing Machinery, New York.
  40. Shao, F., Cao, Y., Gu, J., Wang, Y. (2010). A new real-time clustering algorithm. Journal of Information and Computational Science, 7(10), 2110–2121.
  41. Sharma, S. (2019). Attribute based discovery architecture for devices in Internet of things (IoT). International Conference for Convergence of Technology (I2CT), Mumbai.
  42. Sharma, S., Kapoor, S., Srinivasan, B.R., Narula, M.S. (2011). HiCHO: Attributes based classification of ubiquitous devices. MobiQuitous 2011, 104, 113–125.
  43. Shelby, Z. (2010). Core link format. IEFT [Online]. Available at: https://datatracker.ietf.org/doc/html/draft-shelby-core-link-format [Accessed 13th November 2021].
  44. Shelby, Z., Hartke, K., Bormann, C. (2014). Constrained application protocol (CoAP). Memo, RFC 7252 Internet Engineering Task Force.
  45. Veizades, J., Guttman, E., Perkins, C., KaplanGuttman, S. (1999). Service location protocol [Online]. Available at: https://www.ietf.org/rfc/rfc2608.txt [Accessed 27th November 2021].
  46. Wang, S., Fan, Y., Zhang, C., Xu, H., Hao, X., Hu, Y. (2008). Entropy based clustering of data streams with mixed numericand categorical values. IEEE/ACIS International Conference on Computer and Information Science, Portland.
  47. Zhou, M. and Ma, Y. (2012). A web service discovery computational method for IoT system. IEEE 2nd International Conference on Cloud Computing and Intelligence Systems, Hangzhou.
  48. Zorgati, H., Djemaa, R.B., Amor, I.K.B. (2019). Service discovery techniques in internet of things: A survey. IEEE International Conference on Systems, Man and Cybernetics (SMC), Bari.

Note

  1. Chapter written by Shachi SHARMA.
..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset
18.189.141.125