• Monday, 27 June 2016, 09:00-12:30, Poseidon B

Speakers

• Miquel Tarzan (Fundació i2CAT, Spain)
• Dimitri Staessens (iMinds, Belgium)
• Peyman Teymoori (University of Oslo, Norway)
• Vladimir Vesely (Brno University of Technology, Czech Republic)
• Leonardo Bergesio (Fundació i2CAT, Spain)
• Vincenzo Maffione (Nextworks, Italy)

Motivation and Context

RINA is the only complete network architecture proposed to date that departs fundamentally from the assumptions and design decisions underpinning the current Internet protocol suite. RINA is based on a strong theory that not only describes computer networking but also the fundamentals of distributed computing. The simplicity and elegance of RINA’s structure provides a general solution for many of the problems that today can only be mitigated under certain circumstances using point solutions, such as multi-­homing, mobility, security or quality of service. Therefore, RINA has the potential to become a real revolution in computer networking; allowing the field to move from an artisan approach – in which knowledge is based on experience and tradition – to a more scientific one -­ in which network architects have models to reason about their designs before actually building them.

EUCNC 2016 provides a very good timing and venue for a RINA tutorial. RINA is already gaining momentum in Europe, where the European Commission has funded a number of research projects such as IRATI, IRINA, PRISTINE or ARCFIRE (starting January 2016). US-­based RINA activities have also been funded by the National Science Foundation, including experimental activities on GENI testbeds. Institutions in Mexico and Brazil are also involved in RINA research and education. RINA standardization activities are being discussed under the auspices of the ISO SC6 WG7(“Future Networks”), and may be of interest to ETSI under the Next Generation Protocols ISG, whose goal is to discuss sustainable alternatives to the current TCP/IP networking protocol suite.

Moreover, a number of open-­source tools such as an OMNeT++ RINA Simulator or the IRATI implementation for Linux/OS are becoming more mature and fit for the purpose of education and experimentation. Tutorials and detailed examples explaining how to reproduce RINA results reported in scientific publications are also becoming available.

Structure and Content

The goal of the tutorial is to provide the audience with an introduction to the concepts, motivation and state of the art of the Recursive InterNetwork Architecture (RINA)1, a network architecture based on the InterProcess Communication paradigm as an alternative to the incumbent TCP/IP protocol suite. RINA is a minimalistic, programmable network architecture that provides the elements necessary to support distributed computer networking at scales ranging from LANs over exascale datecenters to WANs. As such, it is an ideal candidate to effectively support the requirements for future networking scenarios such as 5G, IoT or private custom clouds, while providing the flexibility of SDN and NFV in a natural way. The tutorial will motivate and introduce RINA, discuss congestion control and security in RINA networks comparing them to the current state of the art, and present two tools that can be used to experiment with RINA: an OMNeT++ simulator and a programmable C/C++ implementation for the Linux OS.

The following paragraphs provide a short overview of the contents of each talk.

RINA 101: Motivation and introduction. This talk will first motivate RINA by analyzing some of the main shortcomings of the current Internet architecture. Next the main concepts behind the RINA architecture will be introduced: networking as distributed Inter Process Communication (IPC), a single type of programmable layer that recurses as required or how separation of mechanism from policy bounds the number of protocols required in the architecture. The talk will show how several trends in networking such as SDN or
network virtualization are naturally supported by RINA, meeting requirements for 5G networks, IoT and customized private clouds.

Recursive congestion management. This lecture will analyze how the recursive structure exhibited by RINA can be exploited to improve the management of the load offered to a network, allowing layers to react sooner and respond more effectively to temporary congestion than in today’s Internet. In RINA, congestion management is a per-­layer function. Therefore, the network designer can size each layer to bind the time to detect and react to congestion. Moreover, each layer can deploy the congestion management scheme that is better suited to its operational environment, maximizing the efficiency of the reaction to congestion. Lower layers can apply selective backpressure to upper layers, slowing down the flows that are congesting one or more of the layer’s resources. This talk will present these concepts and illustrate the benefits of recursive congestion management via simple scenarios simulated with RINAsim.

Securing RINA networks. It is no news that Internet security is a big issue today, forcing service providers, companies, and individual users to invest in a myriad of tools and systems to increase their safety level to reasonable values. Some argue that not considering security requirements in the Internet design is the main cause of its weak security, but this talk will explore another angle: how good design contributes to the security of an architecture. The talk will guide the audience through the different design principles that make RINA inherently more secure than the current Internet – and at a much lower cost: securing layers instead of protocols, recursion allows for isolation and layers of smaller scope, separation of mechanism and policy, decoupling of port allocation from synchronization and the use of a complete naming and addressing scheme. The talk will end analyzing how the usual mechanisms of authentication, access control, and confidentiality can be used to protect a layer.

RINA for converged operator networks in 5G scenarios. One of the main challenges to achieve the 5G (and beyond) vision is to design converged operator networks that can effectively support heterogeneous access technologies and address a differentiated set of application requirements (delay, loss, capacity) using the same infrastructure. The clean, repeating building blocks of RINA and the programmability of its functions position this architecture as an ideal framework to attack the 7 / 8 problem. This talk will introduce how one would design a converged operator network making the most of RINA, and compare it to current designs highlighting the key benefits of RINA converged operator networks.

RINASim: OMNeT++ RINA simulator. Simulators are essential tools to educate people in new technologies and to research/experiment with technologies to better understand their behavior and assess their benefits. This session will introduce RINASim, a RINA simulator for the OMNeT++ framework. The talk will briefly explain the main design concepts and jump to an interactive session, allowing interested members in the audience to run simple, pre-­packaged scenarios that will allow them to understand better the concepts behind RINA. The tutorial instructors will prepare an easy‐to‐install RINAsim package before the tutorial, with all the required installation and setup instructions (this way tutorial attendees can prepare before the actual session).

IRATI: A programmable RINA implementation for Linux/OS. The last lecture will be devoted to presenting and explaining IRATI, an open‐source, programmable RINA implementation for the Linux/OS. IRATI was initially started by the FP7 IRATI project and is currently under active development by the FP7 PRISTINE project. The talk will highlight the prototype’s design goals, explain the software architecture, provide an overview of the Software Development Kit that allows developers to exploit RINA programmability in practice and conclude showing examples of experiments performed using the IRATI RINA implementation.