Demo/Poster Night Participants 

Table 1: SyncScan: A tool for Scanning, Decoding, and Leveraging 5G SIB Data for Advanced Applications

In this demo, we present SyncScan, a tool engineered to identify active 5G TDD cells across a range of frequencies and decode their system information for use in advanced applications. We will demonstrate how SyncScan is employed on our powder testbed to decode SIB1 from two nearby Verizon 5G cells.


Table 2: SLICES-RI Beyond 5G BluePrint

For the last decades, testbeds have been designed in a bottom-up approach where resources are selected and implemented according to some hypothetical usages. As a result, true experimenters often have to redesign their experimental setup to be able to use the testbeds, or just use the testbed as a "free" pool of bare metal machines. SLICES-RI is driven by intent and proposes services that addresses fundamental scientific questions defined by the relevant research community. The intent is then transformed into a blueprint that can be deployed on the SLICES infrastructure and act as a benchmarking environment for studies targeting those questions.

This demo illustrates our approach targeting at first the 5G research community. We will explore the blueprint that covers the full experimentlifecycle by running a typical 5G experiment with xAPPs. The objective of the demo is not only to show a deployment but to motivate the research community to participate in the collaborative SLICES-RI project and to adopt a reproducible methodology supporting the full research life cycle.


Table 3: CC* Integration-Small: Integrating Application Agnostic Learning with FABRIC for Enabling Realistic High-Fidelity Traffic Generation and Modeling

Based on a model repository of custom models for immediate use, we will demonstrate our integration tool on FABRIC-hosted experiments through a bespoke matching system that aligns experiment resources with model requirements. The tool is built into the Jupyter notebook libraries for a seamless user experience when deploying on FABRIC slices.

Table 4: Analyzing the impact of congestion control on video conferencing over 5G networks

We conducted experiments to compare two congestion control protocols, namely GCC and SCReAM, in terms of video conferencing metrics such as frame delay and throughput. We used the POWDER indoor OTA lab, with the srs-Ran experiment including two COTS-UEs. This demo will show application-level logs with sent / received frames, and tcpdump captures.


P5 P6: Achieving Line Rate TLS SNI Inspection in the Data Plane using P4 and DPDK

A widely adopted approach to monitor HTTPS traffic leverages the Server Name Identification (SNI) extension of TLS. Generally, the hostname is transferred in plain text over the SNI field. Deep Packet Inspection (DPI) is used to parse the TLS header and extract the hostname. However, DPI is often performed on general-purpose processors, which might result in an overhead to the network, especially under high traffic loads. To this end, this presentation will describe an offloading scheme to extract the SNI hostnames from a TLS segment, using P4 and Data Plane Development Kit (DPDK). In the proposed scheme, a P4 switch is the first line of defense where most of the TLS traffic is processed. DPDK is the second line of defense that processes all TLS packets that require processing capabilities beyond what the P4 switch provides.


P7: Edge-Enabled Real-Time Data Processing in Power-Efficient Weather Stations using IBIS

In collaboration with the National Center for Atmospheric Research, we present an automatic weather station prototype. Our work utilizes the IBIS testbed to manage sensor collection on LOng-RAnge (LoRa) radio-enabled wether stations. Our work aims to enable data intensive experiments and reduce energy costs in deploying automatic weather stations.


Table 8: Continuous and Automated Testing of Open RAN Protocol Stacks on Colosseum

The deployment and testing of cellular networks present a complex challenge due to the numerous components involved, ranging from the core network to the Radio Access Network (RAN) and User Equipment (UE). Effective integration and constant monitoring of these components are crucial. The inherent characteristics of the wireless channel introduce additional complexities with its randomness impeding the repeatability and consistency of testing procedures. As a result, current solutions for private and public cellular systems are heavily reliant on manual intervention for tasks such as network reconfiguration, performance monitoring, and end-to-end testing. This dependence on human intervention significantly hampers the pace of innovation in the cellular systems domain. This demonstration shows the capabilities of Colosseum in performing automated and continuous testing of Open RAN protocol stacks, for which we consider OpenAirInterface (OAI) as a use case. Specifically, we perform continuous testing of latest weekly tags of OAI gNB and UEs, and automatically generate a test report that highlights code regressions in newer versions of this software. The pipeline that we developed has been integrated with the Eurecom continuous integration framework, and can be used to test OAI in large deployments and with different network and channel conditions emulated by Colosseum.

Table 9: W-Fi & 5G-NR coexistence on unlicensed band

Due to increase in data traffic demand, 5G-NR is considering unlicensed band to operate (e.g. 5 GHz). The 5 GHz band is also used by Wi-Fi all over the world. We would demonstrate the challenges (in terms of performance: e.g. throughput) how each of these networks are affected, when they coexist in the unlicensed spectrum. The experiment will be performed on POWDER test bed (remotely accessing). 

Table 10: A New Architecture for Petabyte-scale File Transfer

Our project enhances error detection in scientific data transmission by adopting a Multi-Level Error Detection strategy. By integrating layer specific error control policies, we address the inefficiencies of traditional methods that rely only on host-based error detection and recovery. Our approach ensures higher data integrity and reduces recovery times

Table 11:  Pseudonymetry: Cooperative Spectrum Sharing Protocol for Passive/Active Wireless System Coexistence

Pseudonymetry is a cooperative spectrum sharing protocol for the coexistence of passive receivers and active wireless transmitters. Pseudonymetry uses RF signal watermarking and a database feedback loop to stop an offending transmission whenever measurable interference is detected at the passive receiver. In this demo, we demonstrate low power pseudonym detection at passive receivers by amplitude-based watermarking of OFDM signals in the active transmitters signals. 

Table 12: Experimentation with Mobile 28-GHz Phased Array Antenna Modules

We present experiments utilizing mobile 28 GHz Phased Array Antenna Modules, demonstrating their ability to perform beam steering with high granularity. We present an intuitive graphical user interface (GUI) for easy configuration of parameters such as beam shape, beam position, modulation scheme, and scheduling. We demonstrate this GUI by forming a wireless link between 2 mobile PAAMs and performing sweeping on the receive PAAM to find the angle of arrival from the transmitting PAAM. Tutorials are available online for users interested in experimentation with the 28 GHz PAAMs.

Table 13: Measuring the measurer

The goal of this project is to systematically investigate the resource consumption of a measurement system, with advanced use of MFLib, and compare it with the global FABRIC system, by using information from infrastructure-metrics, with the goal to determine how much impact measuring the system has on the global system. With this system the hope is to help and guide the development of measurement systems, allowing them to be more efficient and to make better decisions when it comes to allocating and using resources, while accurately measuring data. 


Table 14: Benchmarking SciStream on FABRIC

SciStream aims to resolve the challenge of facilitating high-speed (+100Gbps), secure memory-to-memory data streaming in scientific environments where data producers and consumers reside in different security domains. To establish and quantify the value of SciStream, we need to create benchmarks for streaming applications based on a variety of traffic patterns of real scientific applications. While designing and implementing various experiments using testbeds like Chameleon and FABRIC is a step in the right direction, the challenge is compounded by the necessity for reproducibility, which is difficult to achieve across shared network scenarios obtainable in testbed environments.

In this demo, we showcase a set of benchmarks and artifacts designed to precisely evaluate the performance of scientific streaming applications across diverse traffic patterns when running over the SciStream framework.


Table 15: Multimodal Model for Dynamic Data Collection and Analysis at the Edge

The distributed network of Sage nodes has collected thousands of images daily and analyzed each of the images in the cloud to extract meaningful information. With Large Language models (LLMs) becoming more accurate, faster, and less parameter-intensive, there is an opportunity to efficiently describe the information in these images directly on the edge. We have been prototyping a workflow of query-based image collection and analysis using a multimodal model deployed in Sage nodes. This development will lead to a fully functional search database and a system that alerts users when certain actions occur, enhancing the usability and efficiency of the Sage network.


Table 16: Real-Time Traffic Statistics Collection Using RDMA from P4 Switches in Live Networks

This project focuses on leveraging RDMA (Remote Direct Memory Access) from P4 programmable switches to collect live traffic statistics in real network environments. By utilizing the high-speed, low-latency capabilities of RDMA, the system captures and transmits real-time flow metrics directly from the P4 switch to a remote server for analysis. The system serves as a first step towards creating a digital twin network that can be used for research and infrastructure management. 


Table 17: Developing and Using Long-Lived FABRIC Services

There are an increasing number of long-lived experiments being developed that offer advanced services to researchers. We will demonstrate three example long-lived services (experiments) that provide researchers with access to (1) highly accurate one-way latency measurements, (2) link throughput measurements based on perfSONAR, and (3) a file transfer service with in-network caching.


Table 18: FabFed: Tool-Based Network Federation for Testbed of Testbeds

FabFed is a tool that allows FABRIC experimenters to connect their slices with external testbed and cloud resources. We will demonstrate common use cases for cross-testbed experiments using FABRIC with Chameleon, CloudLab, SENSE, AWS, and GCP. We will also describe how FabFed can connect external testbeds using FABRIC as a transit network.


Table 19: MERIFs in Education

Thousands of students have gained hands-on experience with networks, distributed systems, cloud computing, and wireless communications through educational use of testbeds. In this demo, we will showcase ready-to-use teaching materials on three topics - computer networks, network security, and wireless networks.

Table 20, 21: ARA PAWR: Enabling Wireless Experiments with Commercial Off-the-Shelf RAN and X-Haul Platforms

To demonstrate the example experiments enabled by ARA and to stimulate the research community participation, working on pressing problems in wireless communication systems, and applications. The demo will highlight example experiments and key results obtained through real-world deployment of the commercial Off-the-Shelf (COTS) RAN and X-Haul platforms in the field showcasing the potential of ARA for advanced wireless innovation and applications. 

Table 22: ARA-Enabled End-to-End Open-Source Prototyping of 5G/O-RAN Solutions for Predictable Low-Latency Communications

In the demo, we will demonstrate a predictable low-latency communications algorithm and the multilayer latency measurement on ARA's end-to-end open-source wireless living lab. Our goal is to showcase how ARA's experimental infrastructure supports the design, development, and testing of new algorithms using open-source software, such as OpenAirinterface.

Table 23: Combining Attestation and Provenance to Improve Reproducibility and Debugging on FABRIC

Experiments involving Software-Defined Networking (SDN) have an important role in research, both in SDN and also to support other experiments through fine-grained network reconfiguration. The subjects of these experiments include security and efficiency. Testbeds like FABRIC allow researchers to easily build topologies and test new ideas; however, when running at scale it can be difficult to debug and to reproduce important behavior. My work lays the foundation for a system that allows researchers to both easily generate artifacts for analysis and diagnose issues with their experiments.


Table 24: Enhancing perfSONAR Measurement Capabilities using P4 Programmable Data Planes

PerfSONAR is a tool used to monitor and troubleshoot problems in high-speed networks such as Science Demilitarized Zones (DMZs). This presentation will describe a scheme that offloads the traffic monitoring to a P4 programmable data plane (PDP) switch. The scheme integrates a PDP switch with perfSONAR, where the switch continuously collects network measurements (e.g., latency, throughput, packet loss rate) and periodically reports the measurements to the perfSONAR archiver. This integration significantly enhances the granularity, visibility, and troubleshooting capabilities of perfSONAR. Additionally, the scheme automates the reporting period according to the variability of the monitored measurements, which eliminates the need of human intervention observed in today’s networks.


Table 25: 5G in Practice: Measuring Emerging Rural Wireless Technology for Edge Devices in Distributed Computation Workloads

In this work, we conduct experiments over a distributed network of Raspberry Pi 4 Single Board Computers (SBCs) to determine network capabilities and to perform standard data processing benchmarks using Apache Hadoop. We measure 5G rural wireless technologies and implement MapReduce, a versatile and widely used server technology, making our results broadly applicable. Our analysis, conducted with equipment from the NSF-supported ARA Wireless Living Lab, ensures reproducibility of our 5G experiments through open-source Docker images, supported by the FLOTO fleet management software for large-scale field deployments.


Table 26: Real-time Multi-Camera Analytics for Traffic Information Extraction and Visualization

We integrated two street-level cameras and a programmable edge computing node within the COSMOS testbed in New York City with a central management platform provided by Kentyou. The cameras have perpendicular view of an intersection. Our pipeline collect and analyze the video streams from both cameras and obtain real-time traffic and pedestrian-related information. The pipeline merges the obtained information from both cameras in a unified top-view of the intersection, and send the results to a dedicated dashboard for real-time visualization and further assessment (e.g., accident prevention).