skip to main content

Explore Research Products in the NSF-PAR

Title: A hardware-in-the-loop emulation testbed for high fidelity and reproducible network experiments
The transformation of innovative research ideas to production systems is highly dependent on the capability of performing realistic and reproducible network experiments. In this work, we present a network testbed consisting of container-based network emulation and physical devices to advocate high fidelity and reproducible networking experiments. The testbed integrates network emulators (Mininet), a distributed control environment (ONOS), and physical switches (Pica8). The testbed (1) offers functional fidelity through unmodified code execution in emulated networks, (2) supports large-scale network experiments using lightweight OS-level virtualization techniques and capable of running across distributed physical machines, (3) provides the topology flexibility, and (4) enhances the repeatability and reproducibility of network experiments. We validate the testbed fidelity through extensive experiments under different network conditions (e.g., varying topology and traffic pattern). We also use the testbed to reproduce key results from published network experiments, such as Hedera, a scalable and adaptive network traffic flow scheduling system.  more » « less
Award ID(s):
1730488
NSF-PAR ID:
10063958
Author(s) / Creator(s):
; ; ; ;
Date Published:
Journal Name:
2017 Winter Simulation Conference (WSC)
Page Range / eLocation ID:
408 to 418
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; ; ; ( , Journal of Optical Communications and Networking)

    With the ever-increasing size of training models and datasets, network communication has emerged as a major bottleneck in distributed deep learning training. To address this challenge, we propose an optical distributed deep learning (ODDL) architecture. ODDL utilizes a fast yet scalable all-optical network architecture to accelerate distributed training. One of the key features of the architecture is its flow-based transmit scheduling with fast reconfiguration. This allows ODDL to allocate dedicated optical paths for each traffic stream dynamically, resulting in low network latency and high network utilization. Additionally, ODDL provides physically isolated and tailored network resources for training tasks by reconfiguring the optical switch using LCoS-WSS technology. The ODDL topology also uses tunable transceivers to adapt to time-varying traffic patterns. To achieve accurate and fine-grained scheduling of optical circuits, we propose an efficient distributed control scheme that incurs minimal delay overhead. Our evaluation on real-world traces showcases ODDL’s remarkable performance. When implemented with 1024 nodes and 100 Gbps bandwidth, ODDL accelerates VGG19 training by 1.6× and 1.7× compared to conventional fat-tree electrical networks and photonic SiP-Ring architectures, respectively. We further build a four-node testbed, and our experiments show that ODDL can achieve comparable training time compared to that of anidealelectrical switching network.

     
    more » « less
  2. ; ; ; ( , ACM Transactions on Internet Technology)

    Distributed denial of service (DDoS) attacks have been prevalent on the Internet for decades. Albeit various defenses, they keep growing in size, frequency, and duration. The new network paradigm, Software-defined networking (SDN), is also vulnerable to DDoS attacks. SDN uses logically centralized control, bringing the advantages in maintaining a global network view and simplifying programmability. When attacks happen, the control path between the switches and their associated controllers may become congested due to their limited capacity. However, the data plane visibility of SDN provides new opportunities to defend against DDoS attacks in the cloud computing environment. To this end, we conduct measurements to evaluate the throughput of the software control agents on some of the hardware switches when they are under attacks. Then, we design a new mechanism, calledScotch, to enable the network to scale up its capability and handle the DDoS attack traffic. In our design, the congestion works as an indicator to trigger the mitigation mechanism.Scotchelastically scales up the control plane capacity by using an Open vSwitch-based overlay.Scotchtakes advantage of both the high control plane capacity of a large number of vSwitches and the high data plane capacity of commodity physical switches to increase the SDN network scalability and resiliency under abnormal (e.g., DDoS attacks) traffic surges. We have implemented a prototype and experimentally evaluatedScotch. Our experiments in the small-scale lab environment and large-scale GENI testbed demonstrate thatScotchcan elastically scale up the control channel bandwidth upon attacks.

     
    more » « less
  3. ; ; ( , 2020 29th International Conference on Computer Communications and Networks (ICCCN))
    null (Ed.)
    Network traffic modeling plays an important role in the generation of realistic network traffic in test environments. Especially in cases where researchers carry out experiments with real production-like traffic, as seen in specific home, enterprise, campus, LAN, or WAN networks. We present our ongoing work on a new framework that enables the methodical creation of application-agnostic traffic models from given network traces of a known network topology. The framework then uses these models to generate realistic traffic on a given network topology. We share a preliminary evaluation of the framework based on repeatable experiments where we model a typical web application traffic and then regenerate the traffic based on the model in a test network on our VTS (Virtual Topology Services) testbed. 
    more » « less
  4. ; ; ( , ACM SIGSIM-PADS, June 2023)
    P4’s data-plane programmability allows for highly customizable and programmable packet processing, enabling rapid innovation in network applications, such as virtualization, security, load balancing, and traffic engineering. Researchers extensively use Mininet, a popular network emulator, integrated with BMv2, for fast and flexible prototyping of these P4-based applications, but due to its lower performance in terms of throughput and latency compared to a production-grade software switch like Open vSwitch, it is crucial to have an accurate and scalable emulation testbed. In this paper, we develop a lightweight virtual time system and integrate it into Mininet with BMv2 to enhance fidelity and scalability. By scaling the time of interactions between containers and the underlying physical machine by a time dilation factor (TDF), we can trade time with system resources, making the emulated P4 network appear to be faster from the viewpoint of the switch/host processes in the container. Our experimental results show that the testbed can accurately emulate much larger networks with high loads, scaled by a factor of TDF with extremely low system overhead. 
    more » « less
  5. ; ; ; ( , IEEE systems journal)
    The ongoing electrification of the transportation fleet will increase the load on the electric power grid. Since both the transportation network and the power grid already experience periods of significant stress, joint analyses of both infrastructures will most likely be necessary to ensure acceptable operation in the future. To enable such analyses, this paper presents an open- source testbed that jointly simulates high-fidelity models of both the electric distribution system and the transportation network. The testbed utilizes two open-source simulators, OpenDSS to simulate the electric distribution system and the microscopic traffic simulator SUMO to simulate the traffic dynamics. Electric vehicle charging links the electric distribution system and the transportation network models at vehicle locations determined using publicly available parcel data. Leveraging high-fidelity synthetic electric distribution system data from the SMART-DS project and transportation system data from OpenStreetMap, this testbed models the city of Greensboro, NC down to the household level. Moreover, the methodology and the supporting scripts released with the testbed allow adaption to other areas where high-fidelity geolocated OpenDSS datasets are available. After describing the components and usage of the testbed, we exemplify applications enabled by the testbed via two scenarios modeling the extreme stresses encountered during evacuations. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email