Co-located Service Router Reducing Network Bandwidth in Distributed Storage
Solution Overview
Problem
Distributed service provider systems face challenges in efficiently and reliably processing data access requests due to high network bandwidth consumption and potential failure points introduced by multiple communications hops in distributed storage systems, leading to suboptimal reliability and end-user perception issues.
Innovation Solution
Implementing a service provider system architecture where a service and router are executed within the same physical machine, reducing network hops and bandwidth consumption by half, and enhancing reliability through in-memory data access handling and deterministic routing techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If distributed storage techniques are used to store enormous amounts of data, then data storage capacity and scalability are improved, but network bandwidth consumption increases and reliability decreases due to multiple communication hops
Solution Approach 1:
The patent combines the service function and router function into a single computing node, eliminating separate communication hops between services and routers. This merging reduces network bandwidth consumption while maintaining distributed storage capacity, as the service and router co-locate on the same physical machine rather than communicating over the network
2Adaptability or versatility
If multiple communication hops are introduced in distributed storage access, then data routing flexibility is improved, but system reliability deteriorates due to increased potential failure points
Solution Approach 1:
By merging the service and router into the same computing node, the patent eliminates intermediate communication hops that introduce failure points. The router on the same node can still provide routing flexibility for accessing distributed storage, but without the reliability penalties of network communication failures
3Adaptability or versatility
If services are distributed to different computing systems, then system adaptability and load handling are improved, but data access latency increases due to network communication overhead
Solution Approach 1:
The patent merges service and router functions on the same computing node, eliminating network communication overhead for service-to-router data access. This reduces data access latency while preserving the adaptability benefits of distributed services, as each service node independently handles its own routing without network round trips
Data Source
AI summary
A method and apparatus for executing data access requests in a distributed storage system are described. The method can include initiating, by a service application executing on a first computing node, a data access request to manage service data using a plurality of distributed data storage nodes that store service data for the service. The method also includes communicating, by the service application to a router application executing on the first computing node, the data access request, and determining, by the router application, at least one data storage node from the plurality of distributed data storage nodes that can satisfy the data access request. The method also includes transmitting, by the router application to at least one data storage node, the data access request for fulfillment of the data access request on behalf of the service application.