CDN Defragmentation via Historical Delivery Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Adaptive bitrate (ABR) content delivery networks face issues with unfair bandwidth allocation and content fragmentation due to opportunistic client behavior and uneven network conditions, leading to inefficient resource utilization and poor user experience.
Innovation Solution
Implementing a system that pre-provisions ABR assets based on historical delivery trends and network conditions, using delivery rules to determine which content representations to pre-populate at each content delivery node, optimizing defragmentation by analyzing manifests and segment files to retrieve missing content from other nodes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If ABR streaming is used to distribute content across CDN nodes, then content delivery flexibility is improved, but unfair bandwidth allocation and greedy client behavior cause inefficient resource utilization
Solution Approach 1:
The system performs preliminary actions by pre-provisioning content segments at CDN nodes based on historical delivery patterns and predicted popularity. This advance preparation ensures that when content becomes popular, the necessary segments are already available at appropriate nodes, preventing greedy clients from monopolizing bandwidth and improving overall resource utilization efficiency.
2Reliability
If content is pre-provisioned at all CDN nodes, then content availability is improved, but network bandwidth is wasted on unpopular content representations
Solution Approach 1:
The system applies local quality by provisioning different content representations at different CDN nodes based on local historical delivery patterns. Each node receives only the specific content segments and representations that are relevant to its local user base, ensuring content availability for local users while avoiding the waste of provisioning all possible representations at all nodes.
Solution Approach 2:
The system uses partial action by provisioning only a subset of content representations at each node based on predicted popularity and historical patterns. Instead of provisioning all possible content representations at every node, the system selectively provisions only those segments that are likely to be requested, reducing bandwidth waste while maintaining adequate content availability.
3Stability of the object's composition
If complete segment collections are maintained at all nodes, then content integrity is improved, but storage space and bandwidth are wasted at nodes that don't need all segments
Solution Approach 1:
The system applies local quality by maintaining different segment collections at different nodes based on their specific roles and local delivery patterns. Edge nodes maintain complete collections for their local content, while other nodes maintain only the segments relevant to their function, ensuring content integrity where needed while reducing overall storage consumption.
4Stability of the object's composition
If defragmentation is performed frequently to maintain complete content collections, then content completeness is improved, but network traffic and processing overhead increase
Solution Approach 1:
The system performs preliminary action by proactively transferring content segments to nodes before they are needed, based on predicted popularity and historical patterns. This prevents fragmentation from occurring in the first place, reducing the need for frequent defragmentation operations and thereby reducing network traffic overhead and processing burden.
Data Source
AI summary
A system and method for optimizing defragmentation of content in a content delivery network (CDN). In an example, a manifest available to the content delivery node with respect to a particular content as well as segment files stored in a database cache of the content delivery node are analyzed to determine if any segment files referenced by the manifest are absent from the database cache. If so, one or more delivery rules associated with the content delivery node based on historical ABR asset delivery patterns may be applied to determine representations of the absent segment files that satisfy the one or more delivery rules. After determining compliant representations of the absent segment files, they may be pulled from another content delivery node of the CDN (e.g., a parent node or an edge server node).


