Client Terminal Representation Selection for Cache Hit Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
HTTP adaptive streaming technologies face network congestion issues due to cache inefficiencies, as caches along the transmission path are not effectively utilized, leading to increased load and reduced benefits from caching.
Innovation Solution
A method where a client terminal sends requests for multimedia content representations in priority order, starting with a first representation, and if not cached, iteratively requests alternative representations until one is found in the cache or all options are exhausted, using directives like 'only_if_cached' and 'altlist' to optimize cache hits and reduce network traffic.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a client terminal requests a specific representation of a content part in HTTP adaptive streaming, then the client receives the desired quality level, but the cache is not hit when alternative representations are available, leading to increased network load and congestion
Solution Approach 1:
Instead of requesting the preferred representation first and falling back to alternatives, the method inverts the approach by requesting alternative representations first (which are more likely to be cached) and falling back to the preferred representation if not cached. This inversion increases cache hit rates and reduces network traffic load.
Solution Approach 2:
The method changes the parameter of representation selection by introducing a set of alternative representations with different qualities. The client terminal dynamically adjusts which representation to request based on cache availability, transforming the static representation selection into a dynamic process that optimizes for cache efficiency.
2Productivity
If a cache is placed along the transmission path to reduce server load, then network resources are saved when cache hits occur, but caches are unable to improve the situation because HTTP adaptive streaming is not cache friendly
Solution Approach 1:
The method introduces dynamic representation selection based on cache availability. The client terminal adapts its representation requests in real-time based on cache responses, transforming the static HTTP adaptive streaming process into a dynamic system that is cache-friendly and can effectively utilize caching infrastructure.
Solution Approach 2:
The method implements a feedback mechanism where the client terminal receives cache hit/miss information and adjusts subsequent representation requests accordingly. This feedback loop enables the system to learn from cache performance and optimize future requests, making HTTP adaptive streaming compatible with caching infrastructure.
3Quantity of substance
If multiple client terminals share a transmission path with a cache, then caching benefits should be amplified, but duplicate transmission occurs when different representations are requested, annihilating the benefits of caching
Solution Approach 1:
The method performs preliminary action by requesting alternative representations first (which are more likely to be cached) before requesting the preferred representation. This preliminary check for cached alternatives prevents duplicate transmissions and reduces network traffic before the main data transfer occurs.
Solution Approach 2:
The method utilizes cached copies of alternative representations when available, instead of requesting duplicate transmissions of the preferred representation. By accepting a cached copy of an alternative representation, the system avoids duplicate transmissions and leverages existing network resources.
Data Source
Figure 1~2
Figure 3A~3B
Figure 4
AI summary
Method for retrieving, by a client terminal, a content part of a multimedia content. Method for retrieving, by a client terminal, a content part of a multimedia content, a cache being arranged along the transmission path between the client terminal and a server, several representations of said content part being available, comprising: A/ sending (S2) a request for a first representation of said content part belonging to a set of allowable representations selected among said available representations of the content part; B/ in case said first representation is not cached, receiving (S3) a response message from said cache indicating that the first representation is not cached; and wherein, upon receipt of a response message, steps A/ and B/ are successively repeated (S4) with an alternative representation of said content part, different from the first representation, until one requested alternative representation has been received by the client terminal (C) from said cache (R) or until all allowable representations of the set have been requested.