Tags:
create new tag
view all tags

Pilot Grant: Explorations on Redundancy Elimination Capabilities

Purpose The focus of the effort to study the practical potential for Redundancy Elimination (RE) as a pilot grant for our effort entitled Whirlwind. Whirlwind takes the approach whereby devices can pre-cache or exchange caching information when in close proximity (ex. D2D exchanges). A pre-requisite for that effort is showing that there is enough redundancy present whereby those opportunistic exchanges would make sense. This pilot study attempts to analyze various scenarios and to marshal evidence either for or against opportunistic D2D exchanges and in a broader sense RE.
Sponsor National Science Foundation
Principal Investigator Aaron Striegel, University of Notre Dame
Students Xueheng Hu, Lixing Song

Progress

We have logged / captured / analyzed data across a variety of venues. Data is anonymized and explored only for signatures / cache strategies.

  • South Shore train line from South Bend to Chicago
    • Several days of WiFi connectivity on the South Shore line via Verizon and AT&T hotspots.
  • University Relations Tent
    • Tent with several hundred individuals for 3-4 hours before game time
    • Fall 2014 season, Fall 2015 season
  • Classroom / student union

We have the following trials planned for Spring / Summer 2016:

  • Joyce Center / Purcell Pavilion and Compton Family Ice Arena for 3-4 home games
  • Additional classroom and train runs
  • Fall run with University Relations tent for 2016

Data

We are in the process of preparing a paper for mid-Spring 2016. Current highlights of the data include (results are preliminary):

  • The percentage of RE has been decreasing
    • RE percentage ranged at the highest up to 20-30% in Fall 2014 for the UR tent. Most venues are sub 10%. The latest UR tent run also echoed that same trend.
    • We are looking to see if that is a function of our caching strategy or a larger trend. Our preliminary thought is that it is a trend.
  • There may be some opportunity for pre-staging
    • Rather than looking at D2D exchanges opportunistically, there appears to be some support for effective pre-staging.
    • We are adapting our research to look at broader pre-staging and high-speed placement at queueing points (i.e. ticketing, concessions) to push said pre-staging data.

We also are planning to anonymize our NetSense Bluetooth ranging data and to share said data for mobility monitoring / opportunities for interaction in a dense environment. That data should be released in early summer 2016 and was the basis for our Mobicom paper.

The paper will be shared via arXiv in mid-Spring 2016 and then submitted to an appropriate venue (likely IMC).

Code

One of the outgrowths of this work as well has been the creation of robust code for the implementation of various caching strategies on an open-source platform. The code can be found via Subversion in our repository for ScaleBox in the src/whirlwind directory.

Edit | Attach | Watch | Print version | History: r1 | Backlinks | Raw View | More topic actions
Topic revision: r1 - 2016-02-02 - AaronStriegel
 
This site is powered by the TWiki collaboration platform Powered by PerlCopyright © 2008-2018 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback