RIDE LEARN DTN Phase II RIDE Robust Internetworking
RIDE & LEARN DTN Phase II RIDE: Robust Internetworking in Disruptive Environments (PARC) J. J. Garcia-Luna-Aceves Ignacio (Nacho) Solis LEARN: Learning Algorithms for Robust Networking (SRI) Mark-Oliver Stehr Carolyn Talcott 1
Use Storage, Processing, and Communication Opportunistically DTN Phase II Kickoff Meeting t 1 t 2 z t 3 2 Treat routes as functions of space and time Exploit longer-term storage of nodes Opportunistic “store-process-forward” August 2006
DTN Phase II Kickoff Meeting RIDE: Robust Internetworking in Disruptive Environments • Data should travel in a self-contained way. – If there is no connectivity, data can still be used. • Data should be referred to by name, not location. – Nodes can identify the data even if they don’t know the destination • Nodes should take advantage of their knowledge and resources – – – 3 Previously seen nodes and data Previous transfers Previous requests Link schedules/characteristics Available storage August 2006
DTN Phase II Kickoff Meeting August 2006 RIDE Technology • • • 4 Content-based opportunistic routing Adapt technologies from proactive, ondemand, and epidemic routing as well gossip and directed diffusion approaches Name data and assign attributes Maximize utilization of storage and processing Transfer named data based on a “virtual potential” (interest/resistance)
DTN Phase II Kickoff Meeting August 2006 Phase II Objectives • Design, Simulation and Evaluation – – • 5 Basic Content Naming and Attributes Signal distribution and scoping Content transfer and caching Content-based opportunistic routing Implementation of a routing algorithm module for the MITRE DTN plug-in architecture
DTN Phase II Kickoff Meeting August 2006 Phase III Objectives • Design, Simulation and Evaluation of routing with topology formation – – – • 6 Nodes have different roles Manage traffic Manage storage Develop more efficient data coverage techniques Advance “subscription” services and opportunistic transfers Incorporate new designs into the module for the MITRE DTN plug-in architecture.
DTN Phase II Kickoff Meeting DTN Phase II & III August 2006 Learning Algorithms for Robust Networking Robust Internetworking in Disruptive Environments A collaboration between SRI International and PARC Palo Alto Research Center Concept of Operations § Objective: reliable communication § in highly disruptive environments § without end-to-end connectivity § Guiding visions: § content-based networking § knowledge-based networking Approach § Opportunistic routing driven by virtual potentials § Learning-based routing with multi-level learning § Opportunistic virtual topology formation § Learning-based virtual topology formation § Hierarchical and agent-organizational techniques Strengths & Impact Interest in § Paradigm shift towards higher level objectives, Content e. g. from message exchange to content dissemination driven by application goals § Addressing is application-defined and resolved incrementally (symbolic routing) Core Technologies § Multiparty communication becomes an emerging concept of content-based networking § New content-based routing algorithms for § Technology independence enables seamless storage-rich disruptive environments interoperation with existing and future protocols § Distributed knowledge management and distributed learning as a cross-layer technology § Enables use of network storage, a valuable resource virtually unutilized by current protocols § Novel approaches to limit information flow § Wide-spread use facilitated by technology § Content-based algorithms for self-forming independence further increases available resources and 7 hierarchical virtual topologies Content & Dissemination Goals
DTN Phase II Kickoff Meeting August 2006 Thank You • PARC – J. J. Garcia-Luna-Aceves <jose. garcia-luna-aceves@parc. com> – Ignacio (Nacho) Solis <ignacio. solis@parc. com> • SRI International – Mark-Oliver Stehr <mark-oliver. stehr@sri. com> – Carolyn Talcott <carolyn. talcott> 8
- Slides: 8