Experimental Evaluation of Memory Management in Content-Centric Networking
read more
Citations
A Survey of Information-Centric Networking Research
WAVE: Popularity-based and collaborative in-network caching for content-oriented networks
Caching in information centric networking: A survey
A Survey of Mobile Information-Centric Networking: Research Issues and Challenges
Modeling data transfer in content-centric networking
References
Web caching and Zipf-like distributions: evidence and implications
Networking named content
Networking named content
I tube, you tube, everybody tubes: analyzing the world's largest user generated content video system
Generating representative Web workloads for network and server performance evaluation
Related Papers (5)
Frequently Asked Questions (12)
Q2. What future works have the authors mentioned in the paper "Experimental evaluation of memory management in content-centric networking" ?
The authors plan to further investigate its properties as future work.
Q3. What is the effect of the chunk popularity on the network?
At higher levels of the network the authors observe a “filtering effect” [16], where chunk popularity is modified by requests already served by lower level caches.
Q4. What is the role of the router?
At network level, routers are responsible for processing incoming Interests and send the corresponding chunk to the interface from which it was requested.
Q5. What is the popularity distribution of the class A1 and A2?
Class popularity distribution is Zipf with parameter α1 = 0.7 and α2 = 2.4 for A1 and A2 respectively, i.e. the probability to request a content of class k, k = 1, 2, ...,K, is q(k) = c/kα, where c is the nor-malization constant, c = (∑K k=1 1/k α )−1.
Q6. How does the algorithm guarantee a minimum amount of memory?
The weighted storage management technique, detailed in Algorithm 2, can guarantee a minimum amount of memory xi = xwi∑i=1,...,N wi ,∀i = 1, ..., N .
Q7. What is the popularity distribution of the content items in the ccnx?
Content items belong to two applications, A1 and A2, both counting 400 content items grouped in K=200 classes of decreasing popularity.
Q8. What is the impact of dynamic storage management on the hit probability of A1?
Fig. 8 highlights that under dynamic storage management, application A2 is able to use the memory not exploited by application A1, that is wasted in case of static partitioning, without affecting A1 hit probability.
Q9. What is the focus of the novel communication paradigm?
According to the novel communication paradigm introduced by Jacobson et al.[1], the focus is on the dissemination and retrieval of information rather than on the interconnection2 of endpoints.
Q10. What are the benefits of static partitioning?
Experimental results point out the benefits deriving from a static per-application cache partitioning and provide indications on how to perform an optimal sizing of cache partitions based on parameters like content popularity or obsolescence.
Q11. What is the impact of the cache size on the hit probability of the popular class?
In fact, if the cache does not distinguish among applications the hit probability depends on the global popularity which is strongly related to request breakdown.
Q12. What is the purpose of the CCNx prototype?
E. Implementation in CCNxThe authors extend the CCNx prototype to support a generic cache replacement policy and the management techniques described above.