Research On Issues Of Caching Thchnologies For Content-Centric Networking

The design principle of the Internet is proposed in the 1960s and 1970s, when the cost of computer hardware is expensive. The emergence of the Internet made it possible to share hardware resources via the Internet. As the most widely used Internet architecture, TCP/IP has received identification worldwide. After more than 50 years, the Internet has become important foundation supporting the development of the society. However, the TCP/IP architecture still acts as the core role on the Internet, which has never been changed. Nowadays, Along with the reduction of production cost on the computer hardware, the demand of sharing hardware resources has been fade, and instead, people focus more on the service and information that the Internet can provide. With the emergence of new network technologies such as Cloud Computing, Social Network and Internet of Things, and the rapid growth of content traffic such VoD and Time Shifted TV, the basic demand of Internet users has been transformed from end-to-end network interconnection to acquiring the user-driven information. Up to now, CDN and P2P are the primary technologies used for solving the problem of content delivery and file sharing based on TCP/IP network. However, due to that CDN and P2P are acted as overlay network, the transmission efficiency is still low.In order to adapt to the development of Internet technology and the change of basic demand of users, Content-centric Networking (CCN) was proposed as novel future network architecture. In CCN, users only focus on the contents themselves rather than the location of contents. Users can send content request directly to the network and the content request can be satisfied by any content provider in the network. To achieve the design goal of "content-centric", nodes in CCN have in-network caching ability and can provide ubiquitous and transparent cache service. In-network caching can reduce the repeat traffic and the download time for content users. Hence, caching technology is one of key technologies in CCN.This thesis presents research achievements in caching technology in CCN. Cache replacement, cache placement and cache locating are emphasized in the thesis as important issues. In a single node, how to replace stale contents by the new coming ones is related to the cache replacement issue; the cache placement strategy decides where to place the contents in the network due to the ubiquitous caching; while how to utilize the cached contents in the network is related to the cache locating issue. These issues have important implications for allocating cache resources, improving cache hit ratio and reducing repeat traffic in the network. Major achievements of this thesis are as follows:(1)In Chapter 3, cache replacement strategy based on guard time is proposed. Due to the temporal locality of the traffic, the same content may be re-requested in a period of time with greater probability. However, if the stack distance of the content request sequence is larger than cache size, popular contents may be replaced frequently in traditional LRU strategy. This phenomenon can be avoided by attaching guard time to cached contents. During the guard time, related contents are prohibited to be replaced. The new arrival contents can only replace contents of which the guard time has been expired. If cache hit event occurs during the guard time, the value of guard time will be reset. In this thesis, the replacement process is modeled based on semi-Markov theory. The sojourn time of cached contents of different popularity and the updating time of cache status are derived. Theoretical analysis and simulation experiments prove that the proposed strategy can be aware of the content popularity and traffic characteristics. In order to avoid the lag of updating cache, a dynamically adjusting algorithm for guard time is also proposed, with which each node in the network can autonomously adjusts the guard time based on the change of cache hit ratio in local, maximizing the cache performance.(2) In Chapter 4, cooperative cache placement strategy is researched to improve the utilization efficiency of cache resources and to avoid the homogenization of cached contents in the network. Firstly, a cache placement strategy based on genetic algorithm is proposed, in which all cache status and traffic characteristics are collected and analyzed together with the topology property. The cache placement process is modeled as a 0-1 integer programming problem, which can be solved by the genetic algorithm. And then, a random cache placement strategy is proposed, in which caching nodes are selected based on the hop distance to the content provider. The expression of cache hit ratio is derives as the function of content popularity, content request rate and cache size. Simulations show that proposed strategy can improve the utilization of cache resources and reduce the average hops to get contents effectively.(3) In CCN, content requests can be satisfied by any content providers in the network, however, the network layer has to bear the task of routing content request to the correct content provider. In order to utilize the ubiquitous cache resources sufficiently, the issue of cache locating is addressed in Chapter 5. Firstly, Ant Colony Optimization is introduced to solve the issue. Cache nodes can send content probe ant to search for the optimal path based on Ant Colony Optimization, meanwhile, cache nodes can broadcast the announcement of popular contents initiatively, assisting to complete the cache locating. Furthermore, a Tracker-based cache locating mechanism is proposed, in which cache collaborative network is established for each popular content file. Tracker nodes can integrate the information submitted from other cache nodes and route the content request to the nearest cache node directly.