Ä¢¹½Av

Ä¢¹½Av
 ›  Edge Computing @ CMU Living Edge Lab  ›  Articles and Blogs

Articles and Blogs from the Living Edge Lab

Friday, January 03, 2025

Unlocking Edge Computing in the new Living Edge Lab 5G Network

We've always planned to move the Living Edge Lab private LTE network to 5G to improve its performance. We hoped for the advertised sub-20ms 5G network round-trip-times. But, when we first built our LTE network, 5G CBRS technology and products were just not ready. Finally, in 2024, we moved to 5G. As a resource for the CMU research community and, with our partners at GXC, the Linux Foundation Magma Community, Kirby Electric, the CMU Manufacturing Futures Institute (MFI), the CMU Entertainment Technologies Center (ETC), and CMU Facilities Management Services, we planned, built, and commissioned a new indoor and outdoor private 5G CBRS network at the historic Mill 19 complex in Pittsburgh. In our new 5G network, we are seeing a reliable 10-15ms round-trip-time -- a 3X improvement over our LTE network!
Friday, December 20, 2024

Learning to Survive

What doesn’t kill you makes you smarter – is the key concept behind our work on Survival Critical Machine Learning (SCML). It uses Live Learning from the Hawk project in the context of a threat laden mission environment. Live Learning takes data collected from the environment during a mission to train new ML models for deployment during that mission. Our recent paper, Beyond Federated Learning: Survival-Critical Machine Learning, introduces a theoretical SCML model centered around autonomous scouts attempting to complete a mission in a hostile environment. The scout has sufficient capacity to train new machine learning models but starts the mission with limited accuracy in detecting the specific threats it will encounter. It has a limited pool of countermeasures to neutralize threats and a constrained connection to a remote human labeler and cloud AI capability. The SCML model balances the tradeoffs between many factors to maximize the scout’s survival time to end-of-mission. We evaluated the SCML model using the Hawk framework and the publicly available DOTA drone surveillance dataset. Our primary question was whether live learning could improve expected mission completion survival over a pre-trained model.
Wednesday, December 18, 2024

Flying on Auto Pilot at the Edge

Small drones – referred to as nano-drones or micro-drones – have become much more common, capable, and less expensive. In the US, the FAA allows licensed UAV pilots to fly drones under 250g over people without advance permission. That regulation has opened the door for applications in urban areas like news reporting, traffic monitoring, construction inspection, recreation, and others. But, today, all these uses require active control by a human pilot. At the Living Edge Lab, we believe that many more applications would be possible if these tiny drones – like their larger brethren – could be made to fly autonomously. This idea was the origin of our SteelEagle project. SteelEagle uses edge computing to supplement the limited computing capabilities of these tiny devices.. Imagine a mission to send a drone to a bridge, fly around it, record the condition of the key bridge components, and compare that condition with previous inspections -- all while avoiding obstacles (like running into the bridge itself!) and dealing with wind, traffic, and other disturbances. This mission is an ideal SteelEagle scenario.
Thursday, October 12, 2023

A Hotspot for the Edge of Nowhere

In a large-scale edge computing deployment, computing resources are built out by an edge data center provider, the access network is provided by a mobile network operator (or other telecom operator) and services and data are provided by both the edge and the cloud. But, what about a small-scale, urgent, and temporary deployment? Imagine using drones to inspect bridges and electric grid towers, mounting a search and rescue operation for a child lost in the woods, hosting a backcountry ski event for a weekend, and providing critical communication and application capabilities for forward military operations -- these are just a few situations where there is a need to deploy a modern compute and communication infrastructure very rapidly in areas with poor network coverage. We believe there are many use cases that could benefit from what we refer to as a “Just-in-Time Cloudlet”. The JIT Cloudlet bundles a mobile access network and edge-native applications into a small form factor. We have developed a reference architecture and a JIT Cloudlet prototype based on commercial off-the-shelf technologies, open-source software, and a “Cloud-native at the Edge” design paradigm.
Tuesday, August 29, 2023

A Ladder to the Shoulders of Giants

Much of today’s scientific research depends on digital laboratories – computing hardware and software that process the experimental data gathered by researchers. Yet, hardware and software evolve at a rapid pace and reproducing the results from a past experiment done even a couple of years ago is a tenuous proposition. Unless the original processing environment is maintained intact, even the original researchers may not be able to reproduce the exact results. And, the difficulty in creating an identical processing environment may preclude independent researchers from verifying the results of their peers. Our recent paper, Towards Reproducible Execution of Closed-Source Applications from Internet Archives, recounts the history of our ten year effort, known as Project Olive, to address the problem of processing environment archiving in service of reproducibility. It also discusses our recent work in this area which expands the problem of archiving to the problem of archive access – making it easy for a researcher to access and run the processing environment while still protecting the software and data from unauthorized use.
Tuesday, August 22, 2023

Safekeeping Faces at the Edge

As so often happens, we can create technology far faster than our social and legal systems can adapt. And, as is so often necessary, we must adapt technology to meet compelling societal needs as they emerge. For the ethical questions raised by face recognition, a more direct intervention to align with emerging norms and policies will likely be required. But, there are legitimate uses for the implicated technologies. Interventions shouldn’t excessively impede these legitimate uses. We started the Silent Witness project at the Living Edge Lab in anticipation of these requirements. We believe that, while norms and policies will vary around the world, four computer vision technology principles will be common.
Friday, April 14, 2023

Easing the Pain of Creating Anywhere

The Covid-19 pandemic forced most members of the university community to work from their homes and other locations with highly variable last-mile connectivity. Since access to campus was limited, they also became more dependent on Virtual Andrew to perform their academic and administrative work. The change from on-campus LAN-access to off-campus WAN access exposed limitations of VDI as a remote service. These limitations led to a three-way collaboration between VMware, the CMU Living Edge Lab, and CMU Computing Service to investigate how Edge Computing could enable a better VDI experience. Our recently published technical report, Deploying Edge-based Virtual Desktop Infrastructure, details this collaboration (so far).
Tuesday, October 25, 2022

Hiding the Devilishness of Detail

As edge computing emerges, it introduces new complexity. Now, a user’s proximity to edge computing resources matter. And, edge computing economics and market structure mean that the edge computing provider and the resources it provides can be quite variable over time and space. For the edge-native application provider, finding the “right” edge computing node or cloudlet to serve a specific user for a specific session can be non-trivial. Here at the Living Edge Lab, we believe that this complexity calls for a new abstraction at the boundary between edge-native application providers and edge computing infrastructure providers. Our Sinfonia project, led by Jan Harkes in collaboration with Meta and ARM, focuses on the development and evaluation of a framework for this abstraction.
Friday, June 10, 2022

Nudging the Planets to Align

At the Living Edge Lab, we’d like to have the benefits of both remote working and serendipity and believe that technology could enable this. Our recent concept paper, Balancing Privacy and Serendipity in Cyberspace, with Nigel Davies of Lancaster University and Nina Taft of Google, from this year’s HotMobile Conference explores technology approaches to bringing serendipitous chance encounters to the virtual workplace. Our question: “Can we create technology-mediated serendipity for coworkers who are not collocated?” The specific use case we examine is generating chance encounters between coworkers who are at different physical workplaces other than at-home
Wednesday, May 25, 2022

Opening Edge Computing’s Black Box

We want a mobile network with round-trip latencies less than 20ms. The lower we can get, the more edge-native applications become feasible. In our first step toward continual improvement in round-trip latency, we launched the Network Latency Segmentation Project led by students Sophie Smith and Ishan Darwhekar. They implemented network measurement probes at various points in the network – at the user equipment (UE), between the Radio Access Network (RAN) and the Evolved Packet Core (EPC) and the between the EPC and the Cloudlet to measure the latency of the network at each segment.
Wednesday, April 27, 2022

Accelerator Calculus at the Edge

This work investigates the use of an image decoding accelerator embedded in an edge storage system used for image analytics. Edge cameras can generate massive amounts of visual data that can’t be economically stored in an uncompressed form at the edge or fully transmitted to a backend cloud. However, city-scale video analytics requires application access to decoded data from large datasets of encoded, compressed stored images or videos at the edge.
Monday, April 18, 2022

Throwing a Softball at Search and Rescue

PhD student Haithem Turki of the Ä¢¹½Av Living Edge Lab has taken the search and rescue scenario as the motivation for his research work on Neural Radiance Fields or “NeRFs”, first developed at UC Berkeley in 2020. NeRFs are the newest technique for 3D scene capture and rendering. They use a type of neural network known as a multilayer perceptron (MLP) to represent the scene as a function. Once a NeRF has been trained from the input data, it embodies the 3D information contained in the scene. The rendering stage extracts the NeRF information and displays it to the user from a specific viewpoint.
Wednesday, July 21, 2021

The Quest for Lower Latency: Announcing the New Living Edge Lab Wireless Network

In 2017 to support our edge computing research, we built our first private LTE network. Working with Crown Castle and DT, we created a small 4G LTE network using an experimental Band 42 3.5GHz license. The network had three outdoor sites and provided limited coverage near the Ä¢¹½Av campus. While we were able to get better latencies than from commercial carriers, the number and types of commercially available devices for that frequency band, limited coverage, and limited support for the equipment made it difficult to use for much of our edge-native application research. Fast forward to late-2020. By this time, unlicensed GAA CBRS spectrum was available and commercial 4G LTE network and user equipment for this frequency band was increasingly available. We began a plan with our partners to upgrade the existing network to a modern, supportable network in the CBRS band, covering a broader geographic area with lower latency and greater throughput.
Tuesday, April 27, 2021

The Waiting is the Hardest Part
Wednesday, April 14, 2021

OpenScout
Tuesday, March 02, 2021

Connecting the Dots at the Edges
Tuesday, January 19, 2021

Edge-Native App Design When You Can’t See Behind the Curtain
Wednesday, September 23, 2020

Putting It All Together With Cognitive Assistance
Wednesday, August 26, 2020

When the Trainer Can’t Be There
Articles and Blogs