Loading...

QoS-aware Dynamic Cache Updating Policies in a Wireless CDN

Kazari, Kiarash | 2018

508 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 51733 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Ashtiani, Farid; Mir-Mohseni, Mahtab
  7. Abstract:
  8. Caching is the process of storing part of popular contents near the users, which not only results in better traffic balance, but also reduces the delay of content delivery. Due to limited cache capacities and time-varying nature of contents popularity, using a proper replacement policy for cached contents during the heavy traffic hours is necessary. Many of the existing methods have ignored the dynamics of requests. Also, most of the previous works have focused on parameters other than QoS characteristics as the performance metric.We propose a cache updating policy for wireless networks considering dynamic popularity for file requests. Our model is a wireless heterogeneous network consisting of one base station (BS) and K access points (AP) which are deployed in its coverage area and each of them is equipped with a cache. Coverage areas of different access points may have overlap with each other, i.e., a user can be served by multiple APs. File requests arrive as a time-varying rate Poisson process at each access point. Cache-missed requests are forwarded to BS and for each of them, BS decides which one of APs should serve the request and send the requesting file to it. After receiving uncached files, AP can decide whether to cache it or not. Considering the overlapping coverage areas of APs, we model the service policy as a stochastic game in which players utilities are defined based on file delivery delay (as a QoS characteristic). By finding the Nash equilibrium of the game, we propose proper policies for both delivery and cache updating phases. Finally by comparing numerical results with some famous cache updating policies we show improvement of network performance in terms of file delivery delay
  9. Keywords:
  10. Caching ; Dynamic Popularities ; Stochastic Game ; Delay ; Heterogeneous Networks ; Service Quality

 Digital Object List

 Bookmark

No TOC