Updated taxonomy for the network and service management research field

Noname manuscript No. (will be inserted by the editor) Updated Taxonomy for the Network and Service Management Research Field Jeroen van der Hooft ·Maxim Claeys · Niels Bouten ·Tim Wauters · J¨urgen Sch¨onw¨alder ·Aiko Pras · Burkhard Stiller ·Marinos Charalambides · R´emi Badonnel ·Joan Serrat · Carlos Raniery Paula dos Santos · Filip De Turck Received: date / Accepted: date Abstract Network and service management is an established research field within the general area of computer networks. A few years ago, an initial taxonomy, organizing a comprehensive list of terms and topics, was established through interviews with experts from both industry and academia. This taxonomy has since been used to better partition standardization efforts, identify classes of managed objects and improve the assignment of reviewers to papers submitted in the field. Because the field of network and service management is rapidly evolving, a biyearly update of the taxonomy was proposed. In this paper, a large-scale questionnaire is presented which was answered by experts in the field, evaluating the relevance of each individual topic for the next five years. Missing topics, which are likely to become relevant over the next few years, are identified as well. Furthermore, an analysis is performed of the records of papers submitted to major conferences in the area. Based on the obtained results, an updated version of the taxonomy is proposed. Jeroen van der Hooft ·Niels Bouten ·Tim Wauters ·Filip De Turck Ghent University - imec, Ghent, Belgium E-mail: [email protected] Maxim Claeys University of Antwerp - imec, Antwerp, Belgium J¨urgen Sch¨onw¨alder Jacobs University Bremen, Bremen, Germany Aiko Pras University of Twente, Enschede, The Netherlands Burkhard Stiller University of Z¨urich, Z¨urich, Switzerland Marinos Charalambides University College London, London, United Kingdom R´emi Badonnel University of Lorraine, Nancy, France Joan Serrat Polytechnic University of Catalonia, Barcelona, Spain Carlos Raniery Paula dos Santos Federal University of Santa Maria, Santa Maria, Brazil 2 Jeroen van der Hooft et al. Keywords Network and Service Management ·Research Topics ·Taxonomy · Analysis ·Questionnaire ·Conference Proceedings 1 Introduction Network and service management has established itself as a research field in the general area of computer networks. Originally, the Association for Computing Machinery (ACM) created a multi-level classification system which includes the keywords Network Services,Network Management and Network Monitoring [1]. This classification, however, does not provide any details to further structure the field of network and service management. For this reason, a new taxonomy has been proposed by a joint effort of the Flamingo FP7 Project and the Committee of Network Operations and Management (CNOM) of the Communications Society (COMSOC) of the Institute of Electrical and Electronics Engineers (IEEE) and the Working Group 6.6 of the International Federation of Information Processing (IFIP) [2]. A total of 7 categories and 57 topics are identified, covering the different types of network, service and business management, and the technologies and methods applied in the field. The resulting taxonomy has since been used to better partition standardization efforts, identify classes of managed objects and improve the assignment of reviewers to papers submitted in the field. Because the field is rapidly evolving, regular updates to this taxonomy are required: small updates (i.e., adding topics) should be incorporated at least every two years, while thorough updates (i.e., adding, removing and potentially merging topics) should be made every five years [2]. In this work, an updated version of the network and service management taxonomy is proposed, based on an analysis of established and emerging topics in the field. The remainder of this paper is organized as follows. In Sect. 2, the original taxonomy of the network and service management field is presented. In Sect. 3, a detailed description of a questionnaire on the taxonomy is discussed. This questionnaire was answered by 154 experts from industry and academia, allowing to evaluate the interest in the topics of the taxonomy. A few minor changes to the taxonomy are presented as well, which were made based on discussions within the Flamingo consortium. In Sect. 4, the original taxonomy is matched to papers submitted to the major conferences in the field (i.e., NOMS, IM and CNSM). Based on the obtained results from the questionnaire and the analysis of paper submissions, an updated version of the taxonomy is proposed in Sect. 5. Finally, Sect. 6 concludes this paper. 2 Original Network and Service Management Taxonomy In 2015, a new taxonomy for the network and service management field was proposed [2]. Different topics are assigned to one of seven categories, creating a twolevel taxonomy. The first three categories deal with what is being managed, more specifically: 1. Network Management: ‘What kind of network is being managed?’ 2. Service Management: ‘What kind of service is being managed?’ 3. Business Management: ‘How does management relate to business aspects?’ Updated Taxonomy for the Network and Service Management Research Field 3 Table 1 Original network and service management taxonomy [2]. 1. Network Management •Ad-Hoc Networks •Wireless and Mobile Networks •IP Networks •Local Area Networks •Optical Networks •Sensor Networks •Overlay Networks •Virtual Networks •Software-Defined and Programmable Networks •Data Center Networks •Smart Energy Grids 2. Service Management •Multimedia Services •Data Services •Hosting (Virtual Machines) •Grid Services •Cloud Services •Resource Provisioning and Management •QoE-Centric Management •Service Discovery, Migration, and Orchestration 3. Business Management •Legal and Ethical Issues •Process Management 4. Functional Areas •Fault Management •Configuration Management •Accounting Management •Performance Management •Security Management •SLA Management •Event Management 5. Management Approaches •Centralized Management •Distributed Management •Autonomic and Self Management •Policy-Based Management •Federated Network Management •Pro-Active Management •Energy-Aware Network Management 6. Technologies •Protocols •Middleware •Mobile Agents •Peer-to-Peer •Grids •Data, Information, and Semantic Modeling •Cloud Computing •Internet of Things •Human-Machine Interaction •Operations and Business Support Systems 7. Methods •Control Theories •Optimization Theories •Economic Theories •Machine Learning and Genetic Algorithms •Logics •Probabilistic, Stochastic Processes, Queuing Theory •Simulation •Experimental Approach •Prototype Design •Monitoring and Measurements •Data Mining and (Big) Data Analytics The fourth category deals with the functional areas: 4. Functional Areas: ‘Which functional areas are covered?’ The last three categories address how something is managed: 5. Management Approaches: ‘How is network and service management achieved?’ 6. Technologies: ‘Which technologies are used in the management process?’ 7. Methods: ‘What methods are used to address the management problem?’ The original taxonomy is shown in Table 1. In the next sections, an analysis is performed of relevant and emerging topics in the field, and an updated version of the taxonomy is proposed. 4 Jeroen van der Hooft et al. Table 2 Five groups of participants from industry. Industry Group Participants Network provider 12 Network equipment provider 16 Cloud system provider 3 Software application provider 7 Other 11 Table 3 Geographic distribution of the participants. Continent Industry Academia Total Total [%] Asia 4 4 8 5.19 Europe 32 81 113 73.38 North America 13 9 22 14.29 South America 0 8 8 5.19 Table 4 Distribution of the level of experience in the management of networks and services of the participants. Exp. [years] Industry Academia Total Total [%] 0 - 5 6 29 35 22.73 5 - 10 10 28 38 24.68 10 - 15 10 19 29 18.83 15 - 20 8 12 20 12.99 20+ 15 17 32 20.78 3 Analysis of Taxonomy Topics Based on Questionnaire To identify required changes to the existing taxonomy, international experts from industry and academia were invited to participate in a questionnaire. Each participant was asked to provide his or her demographical information, including the level of experience, both in network and service management specifically, the background (industry or academia) and geographical location of the current employment (continent). In total, 154 people in the survey, out of which 49 from industry and 105 from academic institutes. As illustrated in Table 2, the participants from industry can be classified in five groups, covering network, network equipment, cloud system and software application providers. Table 3 and Table 4 show the geographic distribution and the distribution of the level of experience in the field of the participants, respectively. Before sending out the questionnaire, the original taxonomy was slightly altered based on discussions during CNOM/IFIP WG 6.6 and Flamingo project meetings. Changes cover both moving topics to another category, in order to better fit the semantics of the categorization described above, and adding new emerging topics, in order to account for new research directions. The following topics were moved to another category: -Internet of Things is moved from category 6. Technologies to category 1. Network Management; -QoE-Centric Management is moved from category 2. Service Management to category 5. Management Approaches; Updated Taxonomy for the Network and Service Management Research Field 5 Fig. 1 Screenshot of the taxonomy questionnaire for the category Network Management. -Resource Provisioning and Management is moved from category 2. Service Management to category 6. Technologies; -Service Discovery, Migration and Orchestration is moved from category 2. Service Management to category 6. Technologies. The following emerging topics were added: -Internet of Things Services was added to category 2. Service Management; -Security Services was added to category 2. Service Management; -Regulatory Perspective was added to category 3. Business Management; -Economic Aspects was added to category 3. Business Management; -Software-Defined Networking was added to category 6. Technologies; -Network Function Virtualization was added to category 6. Technologies. A screenshot of the taxonomy questionnaire for the category Network Management is presented in Figure 1. For each of the seven categories of topics covered by the taxonomy, the participant was asked to indicate which topics should be addressed by the network and service management research community during the next five years (2016-2020), on a scale from 0 (highly irrelevant topic) to 5 (highly relevant topic). Given the broad range of topics presented in the taxonomy, the participant was given the option to not express an opinion. For each category, the participant was also asked to propose important topics which are not yet listed in the current taxonomy. The results of this survey are presented and discussed below. 6 Jeroen van der Hooft et al. Table 5 Topics considered important by academia. Ten of those topics are considered important by industry and two topics are not (†). Results denote the average rating given to each of the topics in the taxonomy (on a scale from 0 to 5) and its standard deviation, the percentage of abstentions (A) and the percentage of times the maximum rating was assigned (M). Topic Industry Academia Rating A [%] M [%] Rating A [%] M [%] 1. Network Management •Wireless and Mobile 4.24 ±1.23 0.00 61.22 4.25 ±1.04 0.00 53.33 Networks •Virtual Networks 4.24 ±1.05 2.04 51.02 4.31 ±0.90 2.86 52.38 •Software-Defined and 4.44 ±1.23 2.04 71.43 4.60 ±0.77 1.90 70.48 Programmable Networks •Internet of Things (†) 4.04 ±1.30 2.04 48.98 4.48 ±0.96 2.86 67.62 2. Service Management •Cloud Services 4.30 ±0.99 4.08 57.14 4.32 ±0.85 0.95 51.43 •IoT Services (†) 3.94 ±1.37 4.08 42.86 4.41 ±0.97 0.95 60.95 •Security Services 4.32 ±1.04 4.08 55.10 4.34 ±0.92 3.81 52.38 4. Functional Areas •Security Management 4.29 ±1.13 2.04 57.14 4.36 ±1.04 1.90 60.00 5. Management Approaches •Autonomic and Self 4.25 ±1.30 2.04 65.31 4.20 ±1.04 0.00 53.33 Management 6. Technologies •Software-Defined Networking 4.19 ±1.17 4.08 53.06 4.45 ±0.75 0.00 57.14 •Network Function 4.46 ±1.03 6.12 63.27 4.37 ±0.85 0.00 58.10 Virtualization 7. Methods •Data Mining and (Big) Data 4.32 ±1.04 4.08 53.06 4.34 ±0.92 3.81 61.90 Analytics 3.1 Analysis of Important Topics The results of the survey allow to select the most important topics, both for industry and academia. An important topic is defined as a topic that received the maximum rating from at least 50% of the participants. Table 5 lists the ten topics which are considered important to both industry and academia, and the two topics which are considered important to academia only. For most of the identified topics, such as Virtual Networks and Software- Defined and Programmable Networks, the average rating for industry and academia are similar. However, for others there is a significant bias towards one of the groups. Wireless and Mobile Networks,Cloud Services,Autonomic and Self Management and Network Function Virtualization are very important for industry, while Software-Defined Networking and Data Mining and (Big) Data Analytics are more important for academia. This indicates that, although general directions are the same, some differences between industry and academia still exist in the field of network and service management. Updated Taxonomy for the Network and Service Management Research Field 7 Table 6 Topics considered less important by academia. Nine of those topics are considered less important by industry and one topic is not (†). Results denote the average rating given to each of the topics in the taxonomy (on a scale from 0 to 5) and its standard deviation, the percentage of abstentions (A) and the percentage of times a rating of 1 or 2 was assigned (M). Topic Industry Academia Score A [%] M [%] Score A [%] M [%] 1. Network Management •Ad-Hoc Networks 2.81 ±1.56 12.24 32.65 2.90 ±1.39 4.76 33.33 •Local Access Networks 2.91 ±1.21 4.08 38.78 2.46 ±1.20 1.90 50.48 2. Service Management •Data Services 3.13 ±1.47 2.04 42.86 2.86 ±1.24 0.95 38.10 •Grid Services 2.78 ±1.28 8.16 40.82 2.93 ±1.31 5.71 34.29 3. Business Management •Ethical Issues 2.98 ±1.41 8.16 30.61 2.95 ±1.34 4.76 34.29 4. Functional Areas •Accounting Management 2.87 ±1.33 6.12 36.73 2.93 ±1.37 4.76 36.19 5. Management Approaches •Centralized Management (†) 3.63 ±1.37 4.08 20.41 3.03 ±1.40 0.95 57.14 6. Technologies •Mobile Agents 2.82 ±1.50 10.20 32.65 2.73 ±1.48 2.86 40.00 •Peer-to-Peer 2.41 ±1.50 10.20 40.82 2.59 ±1.34 2.86 42.86 •Grids 2.15 ±1.48 16.33 46.94 2.32 ±1.29 8.57 46.67 3.2 Analysis of Less Important Topics As in the case for the important topics, the results allow to select the less important topics for industry and academia. A less important topic is defined as a topic that received a rating of 1 or 2 from at least 30% of the participants. Table 6 lists the nine topics which are considered less important to both industry and academia, and one topic which is considered less important to academia only. Again, the average ratings for industry and academia are similar for most of the topics, which includes Ad-Hoc Networks,Peer-to-Peer and Grids. The most notable exception is Centralized Management, which is considered not important by 57.1% of participants in academia, but only by 20.4% in industry. The obtained ratings indicate a decreasing interest in these topics, and can be taken into account when removing topics from the taxonomy in future revisions. 3.3 Analysis of Proposed Topics Each of the questionnaire participants was given the opportunity to suggest new topics for every category that, in their opinion, should be included in the taxonomy. This was done by adding a text field in which the participants could enter a number of topics as they pleased. Table 7 gives an overview of the number of new topics per category that were identified from the questionnaire, and also shows the number of unique topics that were identified through an analysis of the proposed topics. Most topics were proposed in the first two categories, which is a consequence of 8 Jeroen van der Hooft et al. Table 7 Number of topics proposed per category. Category Topics (total) Topics (unique) 1. Network Management 71 35 2. Service Management 38 29 3. Business Management 13 9 4. Functional Areas 27 22 5. Management Approaches 17 14 6. Technologies 25 18 7. Methods 11 10 the general nature of these categories and of the fact that participants were not provided with the full taxonomy at once, and thus often proposed topics which were already covered in other categories. Table 8 gives an overview of all topics which were suggested by at least two participants. Most popular topics in the management categories include Information- Centric Networks,Network Function Virtualization,Home Networks,Content Distribution Networks,Service Integration and Management and Privacy Management. Most of these topics have indeed gotten more attention in recent years, and are valid candidates for the revised taxonomy. Focusing on the category Technologies, the topics Machine Learning and Big Data Analytics have been suggested multiple times. Both topics were in fact already present in the taxonomy, but in the category Methods. This does confirm, however, that these topics have significant importance in the field of network and service management, and should certainly remain in the revised taxonomy. 4 Analysis of Taxonomy Topics Based on Conference Proceedings The answers of the questionnaire illustrate the necessity of updating the original taxonomy to account for current and future directions of the network and service management research field. Similar to previous work, the records of submitted and accepted papers in the field are studied as well, in order to deduce popular topics and current trends. Table 9 shows the percentage of submitted (including rejected) and accepted (inside parenthesis) papers in all editions of NOMS, IM, and CNSM from 2010 to 2015, according to the original taxonomy’s specific, individual topic areas, as described in Table 1. The percentages are obtained by dividing the number of submitted/accepted papers that address a topic by the total number of submitted/accepted papers of each edition. Numbers which exceed a threshold of 10%, are presented in italics. Several interesting trends can be observed from these results. The topics Wireless and Mobile Networks and Security, Fault and Performance Management are well ranked along 2010 to 2016, and are expected to continue doing so. This is in line with the obtained answers to the questionnaire, which indicate that these topics are likely to have a high impact in the research field. Other topics, such as Virtual Networks,Software-Defined Networking (SDN) and Monitoring and Measurements, show a strictly increasing trend. This indicates an increased relevance to and interest in the topic, which certainly holds for SDN-based solutions. Other topics, such as IP Networks,Process Management,Performance Management and Updated Taxonomy for the Network and Service Management Research Field 9 Table 8 New topics proposed by at least two participants. Topic Participants 1. Network Management •Information-Centric Networks 6 •Network Function Virtualization 6 •Home Networks 5 •Security and Privacy 5 •Content Distribution Networks 4 •(Wireless) Mesh Networks 3 •Opportunistic/Social Networks 3 •Vehicular Networks 3 •Body Area Networks 2 •Cellular Networks / 5G Networks 2 •Cloud Networks 2 •Cognitive (Radio) Networks 2 •Enterprise Networks 2 •Hybrid Networks 2 •Factory Networks (Industry 4.0) 2 •Inter-Domain Networks 2 •Non-IP Networks 2 2. Service Management •Service Integration and Management 4 •Industry 4.0 Services 3 •Big Data Management 2 •Information Networking Services 2 •Personalized and Context-Aware Services 2 •Service Assurance 2 3. Business Management •Privacy Management 5 4. Functional Areas •Data Mining and Management of Network Data 3 •Monitoring and Measurement Management 2 •Optimization Management 2 •Privacy and Security Management 2 5. Management Approaches •Hierarchical Management 3 •Cognitive Management 2 6. Technologies •Machine Learning / Data Mining 5 •(Distributed) (Big) Data Analytics 4 7. Methods •Artifical Intelligence 2 Peer-to-Peer, are facing a drop in recent years. This indicates a decreased interest in the community, and means that these topics might be valid candidates for merger or removal of the taxonomy in the near future. It is also worth noting that the topic Wireless Networks is most popular in even years, which could be caused by different interests between NOMS and IM. 16 Jeroen van der Hooft et al. References 1. ACM: The 2012 ACM Computing Classification System. ACM Publications. http://www.acm.org/publications/class-2012 (2012). Accessed 25 August 2017 2. dos Santos, C.R.P., Famaey, J., Sch¨onw¨alder, J., Granville, L.Z., Pras, A., De Turck, F.: Taxonomy for the Network and Service Management Research Field. Journal of Network and Systems Management 24(3), 764–787 (2016) Jeroen van der Hooft obtained his M.Sc. degree in Computer Science Engineering from Ghent University, Belgium, in July 2014. In August of that year, he joined the Department of Information Technology at Ghent University-imec, where he is currently active as a Ph.D. student. His main research interest is the end-to-end Quality of Experience optimization in adaptive video streaming. Maxim Claeys obtained his M.Sc. and Ph.D. degree in Computer Science Engineering from Ghent University, Belgium, in 2012 and 2017, respectively. His research focused on the end-to-end Quality of Experience optimization, ranging from the design of autonomous clients to intelligent in-network decision taking. He is currently active as a post-doctoral fellow at the IDLab group at the University of Antwerp, Belgium. Niels Bouten obtained his M.Sc. and Ph.D. degree in Computer Science Engineering from Ghent University, Belgium, in 2011 and 2016, respectively. His research focused on end-to-end Quality of Experience optimization, providing specific solutions for HTTP adaptive streaming and how these can be applied in a virtualized network environment. He is currently active as product manager at Televic Rail, Belgium. Tim Wauters obtained his M.Sc. and Ph.D degree in Electro-technical Engineering from Ghent University, Belgium, in 2001 and 2007, respectively. He is currently active as a post-doctoral fellow at the Department of Information Technology at Ghent University-imec. His research interests include network and service architectures, and management solutions for multimedia delivery services. J¨urgen Sch¨onw¨alder obtained his M.Sc. and Ph.D. degree in Computer Science from the Technical University Braunschweig, Germany, in 1990 and 1996, respectively. Since 2003, he is professor of Computer Science at Jacobs University Bremen. His research interests include network management, network measurement, network security and distributed systems. Aiko Pras is professor in the area of network operations and management at the University of Twente, the Netherlands. He obtained a Ph.D. degree for his thesis titled ”Network Management Architectures”. His research interests include network management technologies, network monitoring, measurements and security. He is steering committee member of several conferences, including IM/NOMS and CNSM, and series/associate editor of IJNM and TNSM. Burkhard Stiller obtained his Diploma (M.Sc.) in Computer Science and Dr. rer.-nat. degree from the University of Karlsruhe, Germany, in 1990 and 1994, respectively. He is currently a full professor at the University of Zrich UZH, Switzerland, where he leads the Communication Systems Group CSG. His research interests include charging and accounting, overlay networks and blockchains, economic and network/service management, and regulatory aspects of communications. Marinos Charalambides obtained his M.Sc. degree in Communications Networks and Software from the University of Surrey, the United Kingdom, in 2002. In 2008, he obtained his Ph.D degree from the same university. He is currently active as a research associate at University College London, the United Kingdom, where he has been involved in several research projects and courses. His research interests are in the area of network and service management. R´emi Badonnel obtained his M.Sc. degree in Computer Engineering from ESIAL, Nancy, France, in 2003. In 2008, he obtained his Ph.D. degree in Computer Science from Henri-Poincar´e University, Nancy, France. He is currently active as an associate professor of Computer Science at the TELECOM Nancy Engineering School, University of Lorraine, France. His research interests are focused on network and service management, smart and autonomic environments, and the Internet of Things. Joan Serrat obtained his M.Sc. and Ph.D. degree in Telecommunications Engineering from the Universitat Politcnica de Catalunya (UPC), Spain, where he has been active as a full professor since 2008. He is currently leading the UPC research group MAPS in the Management of the Future Internet (FLAMINGO), a Network of Excellence Project supported by the European Commission. His research interests are in network and service management and enabling technologies. Updated Taxonomy for the Network and Service Management Research Field 17 Carlos Raniery Paula dos Santos obtained his M.Sc. degree in Computer Science from the Federal University of Rio Grande do Sul, Brazil, in 2008. In 2013, he obtained his Ph.D. degree in Computer Science from the same university. He is currently active as an adjunct professor at the Federal University of Santa Maria, Brazil, where he has been involved in several supervision and education activities on undergraduate and graduate courses. Filip De Turck obtained his M.Sc. and Ph.D. degree in Electronic Engineering from Ghent University, Belgium, in 1997 and 2002, respectively. He is currently active as a full professor at Ghent University-imec, where he leads the network and service management research group at the Department of Information Technology. His research interests include telecommunication network and service management, and design of efficient virtualized network systems.