Detailed Program

Speakers

Mark Handley (UC London) - Problems of the current Internet, CHANGE project introduction
Laurent Mathy (Lancaster University) - Network Virtualization
Srini Seetharaman (Standford University/Deutsche Telekom R&D Labs), Paul Weissmann - Introduction to OpenFlow
Paul Weissmann (Deutsche Telekom), Srini Seetharaman - How mature is OpenFlow to be introduced in production networks
Siamak Azodolmolky - Extended OpenFlow protocol for circuit switching domain and optical network virtualization
Hideyuki Shimonishi, Yasuhito Takamiya, Kazushi Sugyou (NEC Japan) - OpenFlow platform Trema, hands-on session
Christian Esteve Rothenberg (CPqD - Telecomm. R&D Center, Brazil) - RouteFlow: Virtualized IP routing services in OpenFlow networks
Richard Mortier (University of Cambridge/University of Nottingham) - Building a functional router with Mirage and OpenFlow
Costin Raiciu (Universitea Politehnica Bucuresti) - Centralized or Distributed? An analysis of solutions that efficiently utilize datacenter neworks
Andrew Moore (University of Cambridge) - NetFPGA
Felipe Huici (NEC) - Flowstream and ClickOS: Flexible Flow Processing
Luigi Rizzo (Università di Pisa) - Software packet processing on commodity OS: the netmap framework

Monday

Speaker: Wolfgang Mühlbauer (ETH Zurich)
Title: Organization of the Summer School
Date: Monday, November 7, 12:30-12:35
Slides

Speaker: Mark Handley (UC London)
Title: Problems of the current Internet, CHANGE project introduction
Date: Monday, November 7, 12:35-15:00
Slides

Speaker: Laurent Mathy (Lancaster University)
Title: Network Virtualization
Date: Monday, November 7, 15:30-18:00
Abstract: This tutorial present a brief introduction to network virtualization, with the emphasis on defining a framework to compare and contrast various approaches to virtualization. Subjects discussed include VLAN, VPN, virtual routers, virtual programmable routers and virtual network architectures.
Slides

Tuesday

Speaker: Srini Seetharaman, Paul Weissmann
Title: Introduction to OpenFlow
Date: Tuesday, November 8, 09:00-11:30
Abstract: OpenFlow is being widely embraced, both by the industry and academia, to create platforms for innovation. This talk will provide you hands-on experience with building, programming and troubleshooting the platform. Following an introduction to OpenFlow and the Software-defined Networking (SDN) architecture, each participant will learn how to create a flow-based Ethernet switch within a slice in a hands-on approach. Along the way, you'll learn the OpenFlow software suite: you'll view flow tables with dpctl, dissect packets with Wireshark, write a simple NOX-based controller, simulate a multi-switch/multi-host network with Mininet (emulation environment) on your laptop, and slice it using the FlowVisor. The talk will conclude with a short demonstration of using Expedient/Opt-in Manager for slice management.
Slides

Speaker: Paul Weissmann, Srini Seetharaman
Title: How mature is OpenFlow to be introduced in production networks
Date: Tuesday, Tuesday, November 8, 13:30-15:00
Abstract: This talk will discuss the importance and relevance of OpenFlow for production networks.
Slides

Speaker: Siamak Azodolmolky
Title: Extended OpenFlow protocol for circuit switching domain and optical network virtualization
Date: Tuesday, Tuesday, November 8, 15:30-17:30
Abstract: One of the main promises of Software Defined Networking (SDN) paradigm is to enable the network operators to control and manage their networks in an efficient way and in line with their customer's needs. OpenFlow protocol is a novel mean to materialize this vision, however in its current version the circuit switched (specially optical networking) domain is not comprehensively considered and covered. In this tutorial the extension of OpenFlow protocol to support circuit switched domain will be presented. In addition to the state-of-the-art, interworking/integration of OpenFlow with GMPLS control plane will be over viewed. The use of Extended OpenFlow protocol for Optical network visualization will be also presented in this session.
Slides

Wednesday

Speaker: Hideyuki Shimonishi, Yasuhito Takamiya, Kazushi Sugyou, NEC Japan
Title: OpenFlow platform Trema, hands-on session
Date: Wednesday, November 9, 9:00-11:00
Abstract: Trema is an Open Source modular framework for developing OpenFlow controllers in Ruby/C. The scope of Trema is to help researchers and developers to easily develop their own OpenFlow controllers, and not aiming at providing a specific OpenFlow controller implementation. Trema provides various basic libraries on which users can build their own OpenFlow controllers, as well as integrated network emulator and a lot of sample controllers written in C and Ruby. In this hands-on-session, the architecture and implementation of Trema are introduced, followed by the demonstration of installation, execution on an emulation network, as well as, sample programming of a tiny Trema application.
Slides

Speaker: Christian Esteve Rothenberg
Title: RouteFlow: Virtualized IP routing services in OpenFlow networks
Date: Wednesday, November 9, 13:30-15:00
Abstract: The RouteFlow project proposes a commodity routing architecture that combines the line-rate performance of OpenFlow-enabled hardware with the flexibility of open-source routing stacks (remotely) running on general purpose computers. The outcome is a promising approach in the design space of routing solutions with far-reaching implications towards virtual routers and IP networks as a service. This talk will present the RouteFlow design along its different modes of operation and its main architectural components. We will discuss the prototype implementation and our experiences so far in building an open-source community around the RouteFlow components. The talk will include a demo use case and will conclude by discussing topics of our research agenda.
Slides

Thursday

Speaker: Andrew Moore
Title: NetFPGA
Date: Thursday, November 10, 09:00-11:30
Abstract: An open platform called the NetFPGA has been developed by a group of developers world-wide, lead by Stanford University. The NetFPGA platform enables researchers and instructors to build high-speed, hardware-accelerated networking systems. The platform can be used in the classroom to teach students how to build Ethernet switches and Internet Prototcol (IP) routers using hardware rather than software. The platform can be used by researchers to prototype advanced services for next-generation networks. The NetFPGA enables new types of packet routing circuits to be implemented and detailed measurements of network traffic to be obtained. During the presentation, we will see the NetFPGA used to determine the amount of memory needed to buffer TCP/IP data streaming through the Gigabit/second router and to measure and plot the occupancy of buffers. Circuits will be downloaded into reconfigurable hardware and tested with live, streaming Internet video traffic. Demonstrations of the new 10Gbps NetFPGA hardware will also be made. Attendees will learn about the NetFPGA platform and what it can do. No knowledge of Verilog/VHDL required.
Slides: pdf, ppt

Speaker: Richard Mortier
Title: Building a Functional Router with Mirage and OpenFlow
Date: Thursday, November 10, 13:30-15:30
Abstract: Mirage is an exokernel for constructing secure, high-performance network applications across a variety of cloud computing and mobile platforms. Code can be developed on a normal OS such as Linux or MacOS X, and then compiled into a fully-standalone, specialised microkernel that runs under the Xen hypervisor. Mirage is based around the OCaml language, with syntax extensions and libraries which provide networking, storage and concurrency support that are easy to use during development, and map directly into operating system constructs when being compiled for production deployment. The framework is fully event-driven, with no support for preemptive threading. In this tutorial we will introduce the Mirage framework using a range of examples, exploring construction of network services in detail, with reference to OpenFlow controller and datapath implementations.
Slides (html)

Speaker: Costin Raiciu
Title: Centralized or Distributed? An Analysis of Solutions that Efficiently Utilize Datacenter Networks
Date: Thursday, November 10, 16:00-18:00
Abstract: Datacenter network topologies have become multipath, with several paths available between any two servers. The industry standard is to randomly choose a path for any given connection, either via Equal Cost Multipath or the new TRILL IETF standard. However, random choices create collisions where mutiple flows are placed on the same link, wasting capacity elsewhere in the network. The net effect is reduced utilization: in a particularly bad case, 60% of the network capacity is just wasted. In this talk I will analyze two recent solutions to this problem. Hedera [NSDI 2011] uses a centralized approach to compute optimal flow placement and enforces it by leveraging OpenFlow. MPTCP [Sigcomm2011] splits every connection on multiple paths, finding capacity automatically.
Slides

Friday

Speaker: Felipe Huici
Title: Flow Processing and the Rise of Commodity Network Hardware
Date: Friday, November 11, 09:30-10:30
Abstract: Over the past few years commodity hardware such as x86-based servers have become viable systems for performing flexible yet high-performing flow processing. In this talk we present Flowstream, a flow processing platform based on a set of such servers inter-connected via programmable switches like Openflow and managed by a common platform controller. In addition, we present mechanisms that allow flexible, high-performance and isolated processing on different flows on a single server, as well as light-weight mechanisms for stateful flow migration between servers and platforms.
Slides

Speaker: Luigi Rizzo
Title: Software packet processing on commodity OS: the net map framework
Date: Friday, November 11, 10:30-12:30
Abstract: This talk will present an overview of mechanisms available for software packet processing on commodity OS, discussing their features and performance. We will then introduce the netmap framwork, showing how it can significantly improve the performance of a variety of applications. This will be followed by a discussion of research areas which may benefit from a critical revisitation of software architectures and techniques and strategies used to implement packet processing system, network protocols, and network applications.
Slides (html)

Questions / Problems

Let me (schatzmann@tik.ee.ethz.ch) know if you have any questions or problems accessing the content.