You are here

Distributed Coded Computing for Content Management at the Tactical Edge

Description:

TECHNOLOGY AREA(S): Electronics

OBJECTIVE: To develop a software solution for edge devices (such as smart phone, robot, UAS, IoT sensors, etc. to perform distributed computing and content based storage for efficiently processing, storing, and disseminating computational intensive tasks and information across available edge devices in a Disconnected, Intermittent, Low-Bandwidth (DIL) environment.The warfighters will benefits from increased situational awareness under contested environment in support of future Man-Unmanned Teaming operations and IoT sensor applications.

DESCRIPTION: Handheld mobile technology is reaching first responders, disaster-relief workers, and soldiers in the field to aid in various tasks, such as speech, image, video recognition, natural-language processing, command and control, decision making, and mission planning. However, these edge devices offer less computation power than conventional desktop or server computers and the tactical edge networks are bandwidth limited and suffer from intermittent connectivity to higher echelons in contested environment.In addition, exponential growth in mobile sensing technology is generating large amount of content exceeding the ability of individual edge devices to process/store and the network to disseminate them. This requires tactical edge networks to be able to process, store, and distribute content locally using available edge devices. Distributed computing in the commercial networks is based on centralized schedulers where the scheduler allocates tasks to computing devices and they all report back to the scheduler. Cloud or fog computing in commercial network requires fixed computing infrastructure and 24/7 access to the cloud. Today client server paradigm requires connectivity between edge devices and server hosting platforms. Once the dismounted soldiers get disconnected from the server hosting platforms, there is minimal-to-no capability to process, store, and share locally collected information between connected dismounted soldiers.Separately, in the commercial network, the content based networking approach uses client/server paradigm which is not suitable at the tactical edge. In the academic, there is research being conducted in distributed computing, but mostly following commercial network paradigm. To address the challenges at tactical edge, an innovative distributed computing approach is required to that perform computationally intensive tasks to reduce the collected information into meaningful content using machine learning techniques. The unique aspect of this research involves combining distributed computation with a content based networking paradigm to disseminate and store information efficiently across available edge devices for easy retrieval. Techniques within distributed computing that does not rely on centralized scheduler, works on resource constrained edge devices, addresses changes in number of computing nodes and associated resources and bandwidth limitations is required. New and innovative coded computing paradigms needs to be explored to address challenges of changing computing resources as well as metadata generation, distributed hashing, social networking, and role based encryption techniques to store and disseminate content efficiently across available edge devices should be explored.

PHASE I: Explore and design an architecture for distributed computing suitable for tactical edge network.The architecture shall include machine learning techniques to perform information processing such as object detection, and classification for extracting meaning information.The architecture design shall include content coding and dissemination techniques that considers various network constraints (i.e. computation resources, network bandwidth, and power).The implementation shall include distributed content storage mechanisms, content tagging and encryption technique for secure content dissemination and retrieval.The chosen approach and the algorithms should be substantiated by means of analysis, modeling and simulation or early breadboard prototyping.This task aims to explore the strengths and weaknesses of the architecture for Phase II.

PHASE II: Implement the above architecture and algorithm on COTS edge computing devices.Develop specification of the protocols which make use of the algorithms from phase I. Software implementation of the proposed protocols and algorithms to be implemented on a COTS platform.Demonstrate the system ability to process information using distributed coded computation and perform object detection, and classification to extract and store meaningful information relevant to the users within the network.Demonstrate and deliver capability in a network consisting of 10 node for laboratory assessment. Deliver a prototype system to CERDEC for further testing. Demonstration of capabilities using a network of wireless mobile nodes under a relevant scenario. Demonstration of the scalability properties of the proposed solution using a combination of COTS radio nodes and network emulation tools. Final demonstration shall be conduct in field environment in a network consisting 30 edge computing devices.

PHASE III: Development of distributed coded computation along with content based networking techniques can be integrated with Army’s Nett Warrior and Digital Warrior technologies to bring computationally intensive capability that extracts useful information to increase situational awareness capabilities to the foot soldiers on the ground. The proposed software system shall be integrated with hardware and software of Nett Warrior/Digital Warrior. In addition to military applications, this research is applicable for the First Respondent and the Homeland Security environments where distributed computation tasks to be performed are required especially in the events of natural disasters.

KEYWORDS: distributed computing, content based networking, DIL environment, classification, machine learning, cloudlets, fog computing, architecture, edge devices, UAVs, and UGVs

References:

1. “Command and Control in Underdeveloped, Degraded and Denied Operational Environments” Commercial Technology at the Tactical Edge http://www.dodccrp.org/events/18th_iccrts_2013/post_conference/papers/050.pdf; 2. “Mission-Centric Content Sharing Across Heterogeneous Networks” 2019 International Conference on Computing, Networking and Communications (ICNC), Tim Strayer, Ram Ramanathan, Daniel Coffin, Samuel Nelson; 3. “Content Sharing with Mobility in an Infrastructure-less Environment” Article in Computer Networks 144 · July 2018, Tim Strayer, Samuel Nelson, Amando Caro, Joud Khoury; 4. Cloud Computing at the Tactical Edge, October 2012 https://resources.sei.cmu.edu/asset_files/TechnicalNote/2012_004_001_28146.pdf; 5. “Coded Computing” Salman Avestimehr (USC), Songze Li (USC), Qian Yu (USC), and Mohammad Maddah-Ali (Bell-Labs) http://www-bcf.usc.edu/~avestime/papers/CodedComputingWeb2018.pdf; 6. “Coded distributed computing: Fundamental limits and practical challenges”, 50th Asilomar Conference on Signals, Systems and Computers, Songze Li, Qian Yu, Mohammad Ali, A. Salman Avestimehr

US Flag An Official Website of the United States Government