Dynamic Edge Cache for Query-Based Services
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current caching solutions for query-based API services are inefficient, leading to increased traffic and processing at the API service tier, as they fail to effectively cache dynamically chosen content, especially when using API calls, resulting in slower page loading and higher operational costs, particularly on mobile devices with limited resources.
Innovation Solution
Implementing a dynamic edge caching system that translates fine-grained API requests into coarse-grained API calls, using an API cache manager with components like a CDN connector, request evaluator, query API service connector, instruction generator, dictionary generator, and cache-control directive generator to cache domain service API responses efficiently, reducing data transfer and processing overhead.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If fine-grained query-based API requests are handled directly by the API service tier, then the API service can fulfill specific user requests accurately, but the traffic and processing load on the API service tier increases significantly
Solution Approach 1:
The patent introduces a CDN edge cache as an intermediary between users and the API service tier. The edge cache stores coarse-grained API responses and serves fine-grained requests directly from the edge, acting as a mediator that reduces traffic to the API service tier while maintaining request fulfillment accuracy through dictionary-based response reconstruction.
Solution Approach 2:
The patent segments the API response handling into two parts: coarse-grained data retrieval (handled by CDN edge cache) and fine-grained request processing (handled by API service tier). This segmentation allows the majority of data transfer to occur at the CDN level while only minimal processing occurs at the API service tier.
2Loss of energy
If coarse-grained API responses are cached at the CDN edge, then data transfer and processing overhead are reduced, but the cache must store larger data blocks
Solution Approach 1:
The patent segments cached data into coarse-grained API responses stored at the CDN edge and fine-grained request specifics handled by the API service. This segmentation allows the CDN to cache only the necessary bulk data while maintaining the ability to fulfill specific fine-grained requests without caching every possible variation.
Solution Approach 2:
The patent changes the granularity parameter of cached data from fine-grained (individual API calls) to coarse-grained (bulk API responses). This parameter change reduces the total volume of data that needs to be cached while still enabling efficient fulfillment of fine-grained requests through dictionary-based reconstruction.
3Loss of time
If fine-grained API requests are cached individually, then specific requests can be served from cache, but the complexity of managing and matching cache entries increases
Solution Approach 1:
Instead of caching individual fine-grained requests and their responses, the patent inverts the approach by caching coarse-grained responses and reconstructing fine-grained responses on-demand using dictionaries. This inversion simplifies cache management while maintaining fast response delivery.
Solution Approach 2:
The patent uses dictionary-based copying where a master copy of the coarse-grained response is stored in the CDN edge cache, and fine-grained responses are created by copying and extracting relevant portions from this master copy based on the specific request. This eliminates the need to manage multiple individual cache entries.
Data Source
AI summary
Systems and techniques for dynamic edge cache for query-based services are described herein. A request for web content may be received that includes a fine-grained application programming interface (API) call. Instructions may be generated for at least one coarse-grained API call (e.g., a full dataset) to retrieve data requested by the fine-grained API call (e.g., a data element or subset of the full dataset). A dictionary may be generated to allow the fine-grained API call to retrieve data using the at least one coarse-grained API call. The dictionary and the instructions for the coarse-grained API call may be transmitted to a content delivery server. This technique allows the content delivery server to cache a full dataset via the coarse-grained API call and service fine-grained requests from the edge-cache. This may reduce data transfer and decrease web content delivery times.