Elsevier

Computer Networks

Volume 63, 22 April 2014, Pages 205-220
Computer Networks

Tools to foster a global federation of testbeds

https://doi.org/10.1016/j.bjp.2013.12.038Get rights and content

Abstract

A global federation of experimental facilities in computer networking is being built on the basis of a thin waist, the Slice-based Federation Architecture (SFA), for managing testbed resources in a secure and efficient way. Its success will depend on the existence of tools that allow testbeds to expose their local resources and users to browse and select the resources most appropriate for their experiments. This paper presents two such tools. First, SFAWrap, which makes it relatively easy for a testbed owner to provide an SFA interface for their testbed. Second, MySlice, a tool that allows experimenters to browse and reserve testbed resources via SFA, and that is extensible through a system of plug-ins. Together, these tools should lower the barriers to entry for testbed owners who wish to join the global federation.

Introduction

The internet is an old network, based upon protocols that were developed forty years ago. Its design choices were not made with current needs, such as mobility, wireless communications, real-time interactions, audio and video transmission, contemporary security concerns, etc., in mind. Applications such as the web, peer-to-peer communications, video streaming services, and many others were not anticipated. As we look to the future, with an explosion in the number of communicating agents using the network, and an increasing heterogeneity in the way the network is used, there will be a growing incoherence between the underlying technology and the uses to which it is put.

The shortcomings of the current internet architecture have prompted research initiatives to develop the so-called Future Internet, either through a continuation of incremental deployments or via a clean-slate approach based upon a complete rethinking of the internet’s architecture. Testbeds have emerged from the need to explore this range of possibilities experimentally. They allow for experimentation on a multitude of existing and emerging technologies (sensor networks, wireless access, distributed computing clusters, switches, optical networking, etc.).

A great challenge in the development of these testbeds is to make them both highly capable and easily accessible to experimenters. The capabilities that we seek include great flexibility (for instance, to try radically new approaches at all layers of the protocol stack), global scale, an access to real end-users, high performance, and the ability to test cutting edge technologies. Such capable testbeds are complicated and expensive to put in place, and any given type of testbed might exist at only a few institutions, requiring experimenters to cross administrative boundaries in order to access them.

Because of scarcity, there is a need to share testbed resources among many different experimenters, granting each one what is termed a slice of the whole. Slicing can be achieved through sharing, either through virtualization, where this is possible; through other forms of simultaneous resource division (spatial, frequential); or through temporal division of resources, granting them to different users at different times.

Efforts are underway in Europe (the FIRE initiative), the United States (GENI [1]), and elsewhere in the world to promote the widespread availability of Future Internet testbeds through federation. In a federated system, experimenters at one institution can be authenticated and authorized to gain access to testbed resources hosted by other institutions. In addition, if physical interconnections exist between testbeds, a highly capable federation has the potential to allow experiments that span multiple testbeds at once.

The de facto standard for testbed federation today is the Slice-based Federation Architecture, or SFA. A testbed owner who wishes to enter their testbed into the global federation needs to provide that testbed with an SFA interface. However, this is not necessarily an easy task. Several variants of SFA exist, both in the form of working code and written specifications [2], [3], [4]. Some aspects of SFA, such as its authentication mechanisms, require specialized knowledge to implement. Other aspects, such as parsing capabilities for resource descriptions, can be laborious to put in place.

The contribution of this paper is the introduction of two tools, SFAWrap and MySlice, that lower the barriers to entry for testbed owners who wish to join the global federation. SFAWrap enables an owner to easily provide an SFA interface for their testbed, and MySlice is a tool for experimenters that accepts plug-ins that understand the semantics of individual testbeds. Both tools follow an open community development model and are released under free and opensource licenses.

Fig. 1 shows both tools in relationship to the experimenter (on the left), the testbed and its resources (on the right), SFA (in the middle), and other federated testbeds (below). Section references in the figure describe the organization of this paper. Section 2 provides background on SFA. Section 3 describes SFAWrap. And Section 4 describes MySlice. Section 5, not shown on the figure, draws conclusions and points to future work.

The technologies and tools introduced in this article have a broad range of applications. We are specifically using them as the building blocks of the OneLab experimental facility [5], a federation of major testbeds for experimentation in computer networking.

Section snippets

The Slice-based Federation Architecture (SFA)

To set the context for understanding the two major contributions of this paper, the SFAWrap and MySlice tools, this section describes the Slice-based Federation Architecture (SFA) that they support.

Motivations

The aim of both SFAWrap, discussed in this section, and MySlice, discussed in Section 4, is to lower the barriers to entry into the global testbed federation for testbed owners. These owners know their equipment, they know how to describe it, the ways in which it can be reserved and provisioned, the types of experiments that their testbeds can support. They should be allowed to focus their development efforts on the aspects of the control and experimental planes that directly relate to these

Motivations

As does SFAWrap, MySlice aims to lower the barriers to entry to the global testbed federation for testbed owners. Whereas SFAWrap handles the server side, MySlice handles the client side. Federated environments pose a challenge to testbeds, since tools that work across a federation do not necessarily adapt easily to heterogeneity. MySlice’s architecture allows testbed owners to design plug-ins that expose the unique features of a facility to experimenters, thereby making it easier to bring that

Conclusion and future work

We are at a point in time when SFA has emerged as a de facto global standard for the federation of Future Internet testbeds. However, for a testbed owner, providing one’s testbed with an SFA interface can be a daunting task. This is due to the complexity of some features (authentication, parsing) and the still fluid nature of the SFA specification, which will require regular code updates for any implementation that needs to remain current.

Our objective is to facilitate the participation of new

Acknowledgments

The research leading to these results has received partial funding from a number of sources. The design and development of these technologies is funded by the European Commission’s Seventh Framework Programme (FP7/2007–2013) under Grant Agreement No. 287581, OpenLab. The application of our tools to different testbed technologies is funded by the European Commission’s Seventh Framework Programme (FP7/2007–2013) under Grant Agreement No. 257867, NOVI (for the FEDERICA testbed), and by the French

Jordan Augé received a degree in engineering from the Ecole Nationale Supérieure d’Informatique pour l’Industrie et l’Entreprise (ENSIIE) in Evry in 2004. He has been a PhD student at Orange Labs and Telecom ParisTech on the modeling and evaluation of a flow-level traffic engineering proposal that breaks with traditional approaches for QoS; then has spent a year in the Computer Laboratory of the University of Cambridge, working on high speed passive measurements. He is currently a research

References (28)

  • M. Berman, J.S. Chase, L. Landweber, A. Nakao, M. Ott, D. Raychaudhuri, R. Ricci, I. Seskar, Geni: a federated testbed...
  • L. Peterson, R. Ricci, A. Falk, J. Chase, Slice-based federation architecture (sfa), Working draft, Version 2.0,...
  • GENI AM API, 2013....
  • OpenSFA, 2013....
  • S. Fdida, T. Friedman, T. Parmentelat, OneLab: an open federated facility for experimentally driven future internet...
  • R. Morgan et al.

    Federated security: the Shibboleth approach

    EDUCAUSE Quart.

    (2004)
  • SensLAB (project website), 2012....
  • F-Lab: Federating Computing Resources (project website), 2012....
  • M. Campanella, F. Farina, The federica infrastructure and experience, Computer Networks – special issue on Future...
  • NOVI (project website), 2012....
  • A.-C. Anadiotis et al.

    A new slicing scheme for efficient use of wireless testbeds

  • Teagle (project website), 2012....
  • OpenLab (project website), 2012....
  • M. Suñé, L. Bergesio, H. Woesner, T. Rothe, A. Köpsel, D. Colle, B. Puype, D. Simeonidou, R. Nejabati, M. Channegowda,...
  • Cited by (10)

    • SLICES, a scientific instrument for the networking community

      2022, Computer Communications
      Citation Excerpt :

      In Europe, the ambition was to federate testbeds with very heterogeneous resources. Both initiatives were nicely articulated and produced SFA [7], the Slice-based Federation Architecture, proposing a practical solution to federate the facilities managed by independent authorities. The third phase already started with initiatives such as NSF PAWR in the US [8], CENI in China and ICT 17/19/52 in Europe [9–11].

    • A survey on key technologies of network innovation testbeds

      2017, Jisuanji Yanjiu yu Fazhan/Computer Research and Development
    • Software-defined exchange for the virtualized WiFi network towards future Mobile Cloud services

      2016, 2016 IEEE International Conference on Communications Workshops, ICC 2016
    View all citing articles on Scopus

    Jordan Augé received a degree in engineering from the Ecole Nationale Supérieure d’Informatique pour l’Industrie et l’Entreprise (ENSIIE) in Evry in 2004. He has been a PhD student at Orange Labs and Telecom ParisTech on the modeling and evaluation of a flow-level traffic engineering proposal that breaks with traditional approaches for QoS; then has spent a year in the Computer Laboratory of the University of Cambridge, working on high speed passive measurements. He is currently a research engineer at UPMC Sorbonne Universités, affiliated with LIP6 and LINCS laboratories in Paris, France. His current research interests are focused on the conception of efficient distributed algorithms for capturing the Internet IP graph dynamics.

    Thierry Parmentelat is a senior research engineer at INRIA, his areas of interest include experimental testbeds, networking, and programming languages. Prior to joining INRIA, he has held various positions within the software industry, has been CTO at Centile, a VOIP company, and formerly he has designed and implemented critical systems for industrial sectors, like nuclear energy and aerospace. He graduated from École Polytechnique in 1984, and received his PhD in 1991.

    Nicolas Turro received the MS degree in computer science from the Ecole Superieure en Sciences Informatiques, at Sophia-Antipolis in 1994, and the PhD degree from Université de Nice in 1999. He is currently a research engineer at INRIA Grenoble, in charge of the federation aspects of the Senslab platform.

    Sandrine Avakian is an INSA Lyon engineer, specialized in software development on Wireless Sensor Nodes. She has been working at Inria since 2008 where she was first involved in the Espad project. Then she joined the F-Lab ANR and the FIT equipex development teams where she worked on testbeds federation using SFA for the Senslab and FIT-Iotlab platform.

    Loïc Baron received a degree in computer science from ITIN (Cergy), followed by a Master at Paris 10 University (Nanterre) and finally an Engineer title at CNAM (Paris). He has been working in ICT companies as a consultant in software development, testing and integration. He has worked for different clients such as France Telecom, Expedia, Caisse d’Epargne, and is experienced in various development technologies. He joined CNRS (National Center for Scientific Research) and UPMC (Paris 6 University) in 2012 to contribute to Future Internet projects and develop software components such as SFA and MySlice.

    Mohamed Amine Larabi holds an MSc in computer systems from Ecole nationale Supérieure d’Informatique (ESI), Algiers, Algeria, in 2009. After some work experiences in the RFID industry, satellite telecommunications industry and mobile telecommunications industry, he is currently an expert engineer in the “DIANA” team (f.k.a. PLANETE) at INRIA Sophia Antipolis, working on the federation of heterogeneous network testbeds and cross-testbed resource provisioning.

    Mohammed Yasin Rahman is a Software Engineer at Pierre & Marie Curie University, UPMC, affiliated with LIP6 and LINCS laboratories, Paris, France. He received a Bachelor degree (4 years) in Computer Science from Islamic University of Technology, IUT, Bangladesh in 2009. He started his career as a full time lecturer in Computer Science and Engineering department, IUT, Bangladesh. In 2010 he was awarded Erasmus Mundus scholarship (funded by EC) to conduct Master in Data Mining and Knowledge Management (2010–2012) at UPMC, Paris, France and University of Eastern Piedmont, Italy. After completing master in 2012 he joined as a Software Engineer at UPMC.

    Timur Friedman received the BA degree in philosophy from Harvard University in 1987, the MS degree in management from Stevens Institute of Technology in 1991, and the MS and PhD degrees in computer science from the University of Massachusetts, Amherst, in 1995 and 2001, respectively. He is currently a Maître de Conférences (Assistant Professor) with the Department of Engineering at UPMC Sorbonne Universités, Paris, France, and a Researcher at the LINCS and LIP6 laboratories. His research interests include network measurement systems and networking testbeds.

    Serge Fdida is a Professor with the University Pierre & Marie Curie (UPMC) since 1995. He has been leading many research projects in High Performance Networking in France and Europe, notably pioneering the European activity on federated Internet testbeds and the associated FP6&7 projects ONELAB, ONELAB2 and OPENLAB. Currently, he is also leading the Equipex FIT, a large-scale testbed on the Future Internet of Things. He has also developed a strong experience related to innovation and industry transfer, – he was the Director of EURONETLAB, a joint laboratory established in 2001, together with Thales, – the co-founder of the Qosmos company, – one of the active contributor to the creation of the Cap Digital cluster. He has published many scientific papers and books, and is a Distinguished ACM Member. He has also a long experience of research management. From 2000 to December 2005, he was appointed scientific adviser with CNRS-STIC. In addition, he has been the Vice-President of the RNRT. More recently, has was appointed Vice-President in charge of International Affairs at UPMC from 9/2011 to 12/2011 and is now VP in charge of Europe.

    View full text