Loading...

Network Traffic Reduction in Internet of Things Using Reconfigurable Cache based on Workload Characterization

Rezaei, Rezvan | 2019

353 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 53571 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Bayat Sarmadi, Siavash
  7. Abstract:
  8. Nowadays, with the advancement of internet of things, The number of nodes within the network has increased and there is huge traffic in these networks despite the small volume of data sent and received each node. The current architecture of the Internet network does not have the functionality required for Internet of things, as a result of designing a new architecture for Internet - based Internet of things. Content - based network is a new architecture introduced to the Internet, and its purpose is to change the network from host-centric to content-centric. One of the features of content-based networks is the use of in network caching that helps increase productivity and reduce the traffic in this network. The architectural nature of content - based networks has a good match with the internt of things network and therefore suitable for use in these networks. So far, extensive research has been done in content - based networks and different architectures have been proposed for the cache in these networks. Data on the Internet of things network are usually short - sized data and are short - lived data, on the other hand, the nodes in these networks face constraints such as storage and energy consumption. Also, the applications of internet of things is highly diverse and hence, the workload produced by these applications has different characteristics. In the present study, the effects of the workload characteristics are ignored in the policies that have been presented for in network caching. In this study, the real internet of things workloads and their characteristics have been studied. Then, to improve the performance of the caching policies, the LARC algorithm has already been used in the storage subsystem,has been considered as a primary idea. In this study, nodes in the network are not similar in terms of limitations such as memory and power consumption. For this reason, in the end, considering the characteristics of the workload and the specific constraints of each node in the network, a caching policy is presented for the cache of each node. evaluation of this method has been implemented using named data network simulator, which is implemented using python programming language. The results of the simulation show that the proposed caching policy has improved the network traffic by increasing the hit ratio in caches, decreasing network traffic and increasing the lifetime of nodes by reducing energy consumption
  9. Keywords:
  10. Workload Characteristics ; Internet of Things ; Cache Memory ; Content-Centric Networks ; Reconfiguration Memory ; Traffic Policing

 Digital Object List

 Bookmark

No TOC