VOD Tree Network Server Placement and Program Assignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current Video-on-Demand (VOD) networks face challenges in efficiently distributing content due to high bandwidth requirements and excessive storage and processing costs, particularly in tree topology networks, where either a central server broadcasts all programs or multiple servers are installed at every node.
Innovation Solution
A dynamic programming optimization method is used to determine the optimal location of servers and assignment of programs in a VOD network with a tree topology, balancing server installation, program assignment, and bandwidth costs by installing servers at a subset of nodes and storing a subset of programs at each node.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a central server broadcasts all programs to all demand nodes, then all programs are available to all nodes, but bandwidth requirements on network links become very large
Solution Approach 1:
The patent segments the centralized content distribution into multiple distributed servers located at different nodes in the network. Each server stores a subset of programs and serves specific demand nodes, dividing the overall distribution load and reducing bandwidth requirements on individual network links while maintaining program availability through multiple access points.
2Speed
If a server is installed at every node and all programs that might be requested are stored at the node, then program access is fastest, but the number of servers and storage and processing costs become large
Solution Approach 1:
The patent applies local quality by making different nodes have different characteristics - some nodes have servers with specific program subsets while others don't. This creates a heterogeneous network where server locations and program assignments are optimized for each local area, providing fast access to frequently requested programs while avoiding the need for every node to have a complete server with all programs.
Solution Approach 2:
Instead of installing complete servers with all programs at every node (excessive action), the patent implements partial servers at selected nodes that store only the subsets of programs needed by their local demand nodes. This partial action approach achieves sufficient program access speed for local needs while dramatically reducing the total number of servers and storage requirements.
3Quantity of substance
If servers are installed at a selected subset of nodes and a subset of programs is assigned to each server, then bandwidth and server costs are reduced, but optimal configuration becomes complex to determine
Solution Approach 1:
The patent uses preliminary action by pre-calculating and pre-assigning optimal server locations and program assignments based on network topology and demand patterns. The dynamic programming algorithm determines the optimal configuration in advance, creating a predetermined assignment scheme that simplifies ongoing network operations while achieving optimal bandwidth utilization and cost reduction.
Solution Approach 2:
The patent applies dynamics by using a dynamic programming approach to optimally assign servers and programs. The solution dynamically determines the best configuration by evaluating multiple possible assignments and selecting the optimal one, allowing the system to adapt to different network topologies and demand patterns while providing a structured method for configuration.
Data Source
AI summary
A method provides for the optimal location of servers and the optimal assignment of programs to the servers in a video-on-demand (VOD) network with a tree topology. Each node may have demands for multiple VOD programs. The central server at the root of the network stores all programs, and each of the other servers may store some of these programs. The cost considered include cost of servers, cost of assigning programs to servers, and cost of link bandwidths used for broadcasting programs from servers to demands at various nodes. The demand for a specific program is served by the closest server that has this program along the path that connects the requesting node to the root of the tree network. The invention consists of a dynamic programming method that determines optimal server locations and optimal program assignments for minimizing the costs. Starting from the end-nodes of the tree network, the method determines optimal solutions to sub-trees, eventually reaching the root node, thus providing an optimal solution to the entire network.