| Summary: | 碩士 === 國立臺灣科技大學 === 電機工程系 === 107 === The global IP traffic is predicted to grow more than five times in the next five years. The current host-centric Internet architectures suffer from insufficient use of IP addresses. In Content-Centric Networking (CCN) the IP address is replaced with the content name and the cache memory is used to store data temporarily on each router. Therefore, as long as the content name can be known, the client does not need to know the IP address of the requested data. CCN consists of content store (CS), forwarding information base (FIB), and pending interest table (PIT). When an interest arrives at a CCN node, it will check whether there is a copy of the requested data chunk in the CS, i.e., the node searches the PIT entries for the same requested or named data chunk. If no, the node searches the PIT for an idle entry and a request is sent. If there is no idle entry, the request is blocked. If yes, the aggregation mechanism is performed, i.e., the interest is aggregated and no request is sent. To prevent the request from becoming out-of-date, the timeout mechanism is adopted. That is, a request in service will leave the server if the timeout time is exceeded. In this work, we focus on the performance evaluation of PIT in the cellular network. There are multiple cells in the cellular network and all nodes are mobile. Therefore, there are two classes of interests, new interests and handoff interests. To provide service differentiation, a reservation mechanism is adopted, i.e., there are servers reserved for handoff interests. Upon arrival, an interest is blocked if it encounters server congestion. We study two scenarios for PIT occupancy: new and handoff interests without retrials, and new and handoff interests with retrials. In the scenario with new and handoff interests without retrials, a blocked interest is cleared form the system immediately. In the scenario with new and handoff interests with retrials, a blocked interest is put into the retrial queue and wait to retry later. First, we derived the analytical models for the system considered. An iterative algorithm is developed to find the steady state probability distribution and the performance measures of interest. Second, the impact of various system parameters on the performance measures is studied. The system parameters include interest arrival rate, interest service rate, interest timeout time, interest aggregation probability, interest dwell rate, interest retrial rate, and interest retrial probability. The performance measures of interest are interest blocking probability, average number in service, average system delay, throughput, and incomplete rate. Last but not least, the computer simulation is written to verify the accuracy of the analytical results.
|
|---|
