Online Social Media Content Delivery by Zhi Wang & Wenwu Zhu & Shiqiang Yang
Author:Zhi Wang & Wenwu Zhu & Shiqiang Yang
Language: eng
Format: epub
ISBN: 9789811027741
Publisher: Springer Singapore
Benchmark. For benchmarking the performance, we measure the delay between the time a viewer issues a request and the time it receives the first 1, 000 KB of the video stream, which is the typical size of the video data before a viewer starts the playback. This delay consists mainly of the RTT between the request and service regions, in addition to the queueing and processing delay at the server.
3.6.2 Experimental Results
3.6.2.1 An Overview of Performance Comparison
First, we compare the performance of different upload bandwidth reservation schemes: (1) our scheme given in Algorithm 4; (2) a random allocation scheme in which a small amount of upload bandwidth is progressively allocated to a randomly selected region until the total expense exceeds the budget; (3) a proportional allocation scheme in which each time, a small amount of upload bandwidth is allocated to a region selected randomly according to a probability that is proportional to the number of views in that region, i.e., more upload bandwidth is allocated to regions with more viewing requests. Under each upload bandwidth reservation scheme, the same video deployment algorithm (Algorithm 5) is employed.
In Fig. 3.9, we observe that our bandwidth reservation scheme can achieve much smaller delays for the viewers, given that not only the number of viewing requests in each region but also the user preference of different regions are incorporated in our design.
We also study the request load of these service regions under different upload reservation schemes. In Fig. 3.10, each sample represents the fraction of requests served by a region versus the region’s rank. We observe that in our design, the request load is neither very uniformly distributed, as achieved by the random scheme, nor highly skewed, as in the proportional approach. Instead, our upload bandwidth allocation scheme supplies adequate bandwidth in each region that matches the number of requests in that region. The reason is that our reservation is based on the microblogging prediction framework, which provides a better estimation of future requests in a region than the proportional scheme.
Fig. 3.9Comparison under different upload bandwidth reservation schemes
Download
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.
Algorithms of the Intelligent Web by Haralambos Marmanis;Dmitry Babenko(8309)
Test-Driven Development with Java by Alan Mellor(6795)
Data Augmentation with Python by Duc Haba(6713)
Principles of Data Fabric by Sonia Mezzetta(6456)
Learn Blender Simulations the Right Way by Stephen Pearson(6362)
Microservices with Spring Boot 3 and Spring Cloud by Magnus Larsson(6232)
Hadoop in Practice by Alex Holmes(5965)
Jquery UI in Action : Master the concepts Of Jquery UI: A Step By Step Approach by ANMOL GOYAL(5813)
RPA Solution Architect's Handbook by Sachin Sahgal(5631)
Big Data Analysis with Python by Ivan Marin(5397)
The Infinite Retina by Robert Scoble Irena Cronin(5319)
Life 3.0: Being Human in the Age of Artificial Intelligence by Tegmark Max(5159)
Pretrain Vision and Large Language Models in Python by Emily Webber(4362)
Infrastructure as Code for Beginners by Russ McKendrick(4130)
Functional Programming in JavaScript by Mantyla Dan(4044)
The Age of Surveillance Capitalism by Shoshana Zuboff(3964)
WordPress Plugin Development Cookbook by Yannick Lefebvre(3843)
Embracing Microservices Design by Ovais Mehboob Ahmed Khan Nabil Siddiqui and Timothy Oleson(3646)
Applied Machine Learning for Healthcare and Life Sciences Using AWS by Ujjwal Ratan(3618)
