Loading...

Aging Based Cache Policy for The Network of Caches

Shafaei, Shayan | 2022

137 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 56715 (52)
  4. University: Sharif University of Technology, International Campus, Kish Island
  5. Department: Science and Engineering
  6. Advisor(s): Rezazad, Mostafa; Ghahramani, Amirali
  7. Abstract:
  8. The basis of the current IP-based networking architecture is on addressing locations (IP addresses). This way of locating and searching content does not meet the existing applications’ demands for generating and consuming larger data. In order to confront this problem, Named Data Networking (NDN) has been proposed that allocates names to data instead of the location. This fundamental change improves the functionality of the network in terms of various aspects such as smoother mobile routing, higher security, intrinsic multicast, broadcast system, and enables caching at the network level. As data directly receives an address in NDN, it can reside anywhere in the network. Furthermore, content addressing allows caching in the network devices. Since the request’s address is the same as the content’s name, network devices can index the data in their memory for later similar requests. Also, since the throughput of the networking devices is enormous, and the availability of high-speed memories in such devices is scarce to cache the whole data passing through the devices, such caching architectures’ performance is crucial. Thus, it is imperative to efficiently design an algorithm to cache more essential and popular data instead of other data. There are different policies to enhance the caching space in the routers. The main drawback of some algorithms designed for caching data are as follows: (1) The adjacent routers cache the same contents that have caused a large number of redundant data in the network and reduced the diversity of data. (2) It does not consider the hierarchy of the routers in the routing path. Such shortcomings lead to wasting the considerable space of storage on the routers. (3) Caching data based on the amount of popularity or using other routers imposes a vast overhead on the network. According to the novel algorithm proposed in this research, a critical factor that significantly influences the caches’ performance is considered. As the routers’ memory size is too limited to cache all data, which are passing through the routers, the lifetime of the content in a cache is more important than any other factor. Any caching policies would be futile if the content did not stay in a cache long enough to receive the second request. Therefore, the proposed algorithm assigns different ages for the content cached at varying levels of hierarchies. Through this algorithm, the content is likely to stay long enough to receive its potential hit and distribute various content throughout the routing path. According to the obtained results, the proposed algorithm outperforms the other related algorithms with about 40% more cache hit rate, and brings data closer to the client’s 10% more approximately
  9. Keywords:
  10. Cache Network ; Admission Policy ; Performance Evaluation ; Cache Placement ; Named Data Network (NDN)

 Digital Object List