An SDN-inspired Model for Faster Network Experimentation

2014 IFIP Networking Conference, 2014

Network emulations are widely used for testing novel network protocols and routing algorithms in realistic scenarios. Up to now, there is no emulation tool that is able to emulate large software-defined data center networks that consist of several thousand nodes.

2014 IEEE International Conference on Communications (ICC), 2014

Multipath TCP is an experimental transport protocol with remarkable recent past and non-negligible future potential. It has been standardized recently, however the evaluation studies focus only on a limited set of isolated use-cases and a comprehensive analysis or a feasible path of Internet-wide adoption is still missing. This is mostly because in the current networking practice it is unusual to configure multiple paths between the endpoints of a connection. Therefore, conducting and precisely controlling multipath experiments over the real "internet" is a challenging task for some experimenters and impossible for others. In this paper, we invoke SDN technology to make this control possible and exploit large-scale internet testbeds to conduct end-to-end MPTCP experiments. More specifically, we establish a special purpose control and measurement framework on top of two distinct internet testbeds. First, using the OpenFlow support of GÉANT, we build a testbed enabling measurements with real traffic. Second, we design and establish a publicly available large-scale multipath capable measurement framework on top of PlanetLab Europe and show the challenges of such a system. Furthermore, we present measurements results with MPTCP in both testbeds to get insight into its behavior in such not well explored environment.

Computer networks have evolved due to new trends in the society's needs since their emergence as a way of providing remote access and sharing of computational resources. The architecture inflexibility of the computer networks presents a challenge for researchers, since their experiments can hardly be evaluated in real networks. Thus, in general, tests of new technologies are conducted on network simulators, which imply in a streamline of the reality. The paradigm of Software Defined Networks (SDN) and OpenFlow architecture, offer a way for the implementation of a programmable network architecture, able to be implemented gradually in production networks, which offers the possibility of separating the control mechanisms of the many traffic flows served, so that a scientific experiment can be performed in a real network (adapted for SDN) without interfering with its operation. This paper contextualizes the existing problems in current computer networks, and presents the SDN network as one of the main proposals for the viability of the Internet of the Future. In this context, it is discussed the OpenFlow architecture, which allows the creation of applications for Software Defined Networks. Finally it is presented the network simulator SDN, the Mininet, which implements the OpenFlow interface in a network simulation scenario containing a controller POX with two components, one OpenFlow switch and three nodes. The main objective was to evaluate the communication and bandwidth between nodes.

2007 9th International Conference on Telecommunications, 2007

Evaluating the performance of distributed applications, such as mobile telephone or video conferencing devices, has become increasingly complex as the diversity of scenarios and data rates in networks increase. In this paper, we present the design and implementation of RplTrc, a network emulation tool. RplTrc uses traces to drive interceptions of the Linux protocol stack to delay, drop or duplicate packets. The length of traces is not limited. Thus, unlike other well-known emulators such as NIST Net or the emulator module in the Linux 2.6 kernel, RplTrc is capable to account for important performance characteristics inherent to real networks such as long-range dependence and selfsimilarity of cross-traffic. Tests show that RplTrc installed on a commodity PC is capable to emulate a 100MBit/s or a lightlyloaded 1GBit/s network and has a precision in reproducing packet delays in the order of 100 µs. RplTrc thus enables extensive performance evaluation of distributed applications in lab environments.

The 30 th International Conference on Software, Telecommunications and Computer Networks (SoftCOM 2022)

The Software Defined Network (SDN) concept has changed the traditional network architecture to achieve higher flexibility, the programmability of the network, the abstraction based on a logically centralized control plane, and the data (forwarding) plane decoupling. The SDN control centralization raises scalability issues. This work focuses on the scalability concerns of the Ryu SDN controller in various scenarios with distinct network topologies like tree and datacenter. Throughout this paper, we assess the consequences of distributing subscriber data requests over the controller performance into different network topologies also using a different number of subscribers. The performance of the Ryu is tested by observing the throughput of the controller. The assessment was accomplished by employing an Iperf traffic generator and Mininet. The paper studies two matters that impact the Ryu throughput: the exponential increase of the number of nodes in tree topology and the adoption of the same condition over a datacenter topology. The experiments involve the two types of topologies mentioned and a topologies size range from 200 to 2000 nodes and assessing the controller performance using the throughput as the performance criterion. The results of this research will be used in a research project oriented to 5G slicing, where the control plane is based on SDN.

2008

Performance analysis techniques are fundamental to aid the process of large-scale protocol design and network operations. There has been a tremendous explosion in the variety of tools and platforms available (eg: ns-2, SSFNet, Click Router toolkit, Emulab, Planetlab). However, we still look at the sample results obtained from such tools with skepticism because they are isolated (potentially random) and may not be representative of the real-world. The first issue (random isolated results) can be addressed by large-scale experiment design techniques that extract maximum information and confidence from a minimum number of carefully designed experiments. Such techniques can be used to find "good" results fast to guide either incremental protocol design or operational parameter tuning. The second issue (representativeness) is more sticky and relates to formulating benchmarks. In this paper, we explore the former case, i.e. large-scale experiment design and black-box optimization (i.e. large-dimensional parameter state space search). We propose a new platform ROSS.Net that combines large-scale network simulation with large-scale experiment design and XML interfaces to data sets (eg: Rocketfuel, CAIDA) and models (traffic, topology etc). This is a step towards the broader problem of understanding meta-simulation methodology, and speculate how we could integrate these tools with testbeds like Emulab and Planetlab. Examples of large-scale simulations (routing, TCP, multicast) and experiment design are presented.

Proceedings of the sixth international workshop on Data intensive distributed computing - DIDC '14, 2014

Software-defined networking combined with distributed and parallel applications has the potential to deliver optimized application performance at runtime. In order to investigate this enhancement and design future implementation, a datacenter with a programmable topology integrated with application state is needed. Towards this goal, we introduce the Flow Optimized Route Configuration Engine (FORCE). The FORCE is an emulated datacenter testbed with a programmable interconnection controlled by an SDN controller. We also utilize Hadoop as a case study of distributed and parallel applications along with a simulated Hadoop shuffle traffic generator. The testbed provides initial experimental evidence of support to our hypothesis for future SDN research. Our experiments on the testbed show a difference in application runtime a factor of over 2.5 times on shuffle traffic for Hadoop MapReduce jobs and the potential for significant speedup in warehouse scale data centers.

CSET, 2018

New paradigms and architectures, such as Software Defined Networking (SDN), have added an unprecedented increase in the rate of research and development conducted in the field of computer networks. With this increase , there is a rising need for platforms that can enable researchers and operators to experiment with various scenarios involving performance testing, SDN, security research, topology designs, etc. However, the available emulators fail to address fundamental needs of the experiments requiring diverse and scalable set of topolo-gies. In this work, we propose a novel approach to embed arbitrary topologies on a substrate network of pro-grammable ToR switches using our network virtualiza-tion technique, called Bare-metal Network Virtualization (BNV). BNV is entirely software configurable and has been implemented on open source software and unmodified OpenFlow-enabled switches. The system has been deployed and currently running in a production testbed in National Cybersecurity Laboratory (NCL) for a year. Our evaluations show that BNV can support various data-center topologies with less number of switches which can facilitate building a high fidelity, repeatable and isolated experimentation platform for data-center, SDN and security research in computer networks.

… Conference on Testbeds …, 2005

Technical Note FTN- …, 2004

Network emulation is valuable largely because of its ability to study applications running on real hosts and "somewhat real" networks. However, conservatively allocating a physical host or network link for each corresponding virtual entity is costly and limits scale. We present a system that can faithfully emulate, on relatively low-end PCs, virtual topologies over an order of magnitude larger than the physical hardware, when running typical classes of distributed applications that have modest resource requirements. The new Emulab virtualizes hosts, routers, and networks, while retaining near-total application transparency, good performance fidelity, responsiveness suitable for interactive use, high system throughput, and efficient use of resources. Our key design techniques are to use the minimum degree of virtualization that provides transparency to applications, to exploit the hierarchy found in real computer networks, to perform optimistic automated resource allocation, and to use feedback to adaptively allocate resources. The entire system is highly automated, making it easy to use even when scaling to thousands of virtual nodes. This paper describes the system's motivation, design, and preliminary evaluation. 1 Introduction Network experimentation environments that emulate some aspects of the environment-network testbeds-play an important role in the design and validation of distributed systems and networking protocols. In contrast to simulated environments, testbeds like Emulab [27] and PlanetLab [18] provide more realistic testing grounds for developing and experimenting with software. Emulated environments implement virtual network configurations atop real hardware: this means that experimenters can use real operating systems and other software, run their applications unmodified, and obtain actual (not simulated) performance measures.