Personal tools
Wireless WormholesFrom CUSystems
[edit] Radio Relays for Wireless WormholesUniversity: University of Colorado - Boulder Professors: Dirk Grunwald, Douglas C Sicker, Tim Brown and Peter Mathys Department: Computer Science NSF Funding : NeTS-FIND: Radio Wormholes for Wireless Label Switched Mesh Networks
[edit] Project OverviewThis project addresses two related problems in mesh networking, quality of service and traffic engineering, by advocating a new wireless network technology that builds on existing technologies applied to optical and wired networks. Mesh networks, which share properties of fixed multipoint wireless and ad hoc wireless networks, typically have lower throughput than fixed multipoint wireless networks because the same spectrum is used to deliver service to stations and relay traffic. This proposal seeks to extend and integrate three technologies to make mesh networks more useful. Overall, the proposal involves building “radio circuits” that use “cut through” switching similar to optical lambda switching. Orthogonal frequency multiple access (OFDMA) serves as the underlying PHY layer. Individual mesh nodes are implemented using a software-based radio switch, which allows transiting packets to be rapidly forwarded. Lastly, the Generalized Multi-Protocol Label-Switching (GMPLS) protocol provides a management plane. It is expected that this research will enable mesh and ad hoc networks to decrease latency, increase throughput and enable quality of service capabilities as well as provide mechanisms for traffic engineering. Wireless distribution is one of the three options for high speed broadband access (with cable modem and DSL being the other two). Effective mesh distribution networks will enable wireless to be a competitive alternative in broadband access. It will also provide a reliable network technology for rapidly deployable networks needed for disaster response and recovery. This project presents a novel approach to implementing such community scale mesh networking. Recent evaluation of ad hoc networks has shown that current techniques do not scale well with either network size or desired throughput. We argue that the greatest opportunity for improvement is in the integration and coupling of different network layers. To this end, we introduce a ``relay network`` design in which end-to-end path resources are pre-allocated to avoid the cost of channel acquisition and contention at every hop. The routing, congestion control, and MAC functions are coupled -- contention informs the routing and packet admission decisions, and demand determines link capacity allocation. We propose using spatial frequency division multiple access (SFDMA) to reduce interference and propose a forwarding layer using label-switching to allow a combination of SFDMA as well as statistical multiplexing. Normally, SFDMA methods are difficult to allocate due to a paucity of channels; we provision our network using Orthogonal Frequency Division Multiple Access (OFDMA) channels, avoiding re-synchronization delays on each packet. This project uses a combination of technologies to demonstrate that our proposed design is both effective and realistic. We have simulated the network organization and are implementing the PHY layer using a software defined radio platform. [edit] More Details
[edit] FundingThis project is partially funded by NSF NeTS Project #0435297 - "NeTS:ProWiN: Programmable Radio Platforms for Highly Dynamic Networks" and "NeTS-FIND: Radio Wormholes for Wireless Label Switched Mesh Networks" Resources for this project are provided by NSF Project #0454404 - "CRI: Wireless Internet Building Blocks for Research, Policy, and Education" and NSF Project #0435452 - "NeTS - ProWiN: A Programmable Wireless Platform For Spectral, Temporal and Spatial Spectrum Management" Views |