Streaming Server Local Storage for Archiving Bottlenecks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing content streaming systems face issues with file segmentation and transcoding during server-side archiving, leading to performance degradation and incomplete on-demand content streams when content providers reconnect after a disruption.
Innovation Solution
A scalable content streaming system with ingest clusters, remote storage clusters, and distributed edge servers, where each ingest cluster has shared local storage allowing streaming servers to append to existing files, eliminating the need for transcoding and minimizing bottlenecks by recording streams locally before transferring to remote storage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If content providers host or distribute their own streams, then control over content distribution is improved, but bandwidth costs and storage resource requirements increase significantly
Solution Approach 1:
The patent introduces a content streaming system as an intermediary between content providers and end users. The system includes ingest clusters that receive content from providers, storage servers that archive the content, and edge servers that distribute it to users. This intermediary infrastructure relieves content providers of the burden of providing their own distribution bandwidth and storage resources while maintaining their control over what content is distributed.
2Volume of stationary object
If streaming servers record streams to remote storage directly, then storage capacity is improved, but network traffic bottlenecks and recording performance degrade
Solution Approach 1:
The patent segments the storage function by introducing a two-tier storage architecture: local storage attached to each streaming server in the ingest cluster, and remote storage for long-term archiving. The streaming server first records to fast local storage, avoiding remote storage bottlenecks during the critical recording phase. Only after local storage is available does the server transfer data to remote storage, separating the high-speed recording function from the high-capacity archiving function.
3Adaptability or versatility
If content is transferred from local storage to remote storage during live streaming, then storage scalability is improved, but system complexity and data consistency issues increase
Solution Approach 1:
The patent implements preliminary action by having streaming servers record content to local storage immediately upon receiving it, before any transfer to remote storage occurs. This ensures data is captured locally with full control, and only after successful local recording does the system initiate transfer to remote storage. This preliminary local recording action simplifies data consistency management by establishing a known good state before replication.
4Adaptability or versatility
If multiple streaming servers access remote storage simultaneously, then storage accessibility is improved, but NFS mount overhead and performance degradation occur
Solution Approach 1:
The patent applies local quality by giving each streaming server in the ingest cluster its own dedicated local storage resources. Instead of having all servers compete for access to remote storage over the network, each server has high-speed local storage directly attached, providing optimal performance for its specific recording operations. This localized storage quality eliminates NFS mount overhead while maintaining accessibility, as each server accesses storage locally without network protocol overhead.
Data Source
AI summary
Some embodiments provide a scalable content streaming system that performs server-side archiving. The content streaming system includes a set of streaming server clusters, a remote storage server, and a set of distributed servers. Each streaming server cluster includes at least one streaming server and a local storage server. Each streaming server of a particular cluster distributes a content stream that is uploaded to the streaming server while also recording the content stream to the local storage server of that particular cluster. The remote storage server stores files from the local storage servers that are transferred when recording to those files is complete. The set of distributed servers distribute a live stream of content from a particular streaming server when the content is being uploaded to the particular streaming server and distribute a recorded stream of the content from the remote storage server when the upload is complete.


