Wide Area Network Aggregation System Parallel Request Fan-Out
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Wide area networks often incur high latency and unreliability, which can cause performance issues when communicating across physically disparate sites, such as in large datacenter installations supporting thousands of virtual remote desktops, where multiple WAN operations are required to locate resources.
Innovation Solution
The Wide Area Network Aggregation System (WANAS) enables aggregated communication by transforming a single request into parallel site requests, utilizing pre-established connections and a connection manager for failover and load balancing, allowing local requests to be processed in parallel within each site, thereby reducing latency and improving reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple WAN operations are used to locate resources across disparate sites, then resource availability is improved, but latency increases and performance deteriorates
Solution Approach 1:
The patent pre-establishes WAN connections between datacenters before they are needed for resource location. Connection brokers proactively open and maintain TCP connections across the WAN, so when a resource location request arrives, the communication path is already ready, eliminating connection establishment latency and improving response time while maintaining resource availability across disparate sites.
2Adaptability or versatility
If multiple WAN operations are performed to find resources across sites, then comprehensive resource search is improved, but reliability decreases due to network instability
Solution Approach 1:
The patent introduces connection brokers as intermediary components that manage WAN connections between datacenters. These brokers handle connection establishment, maintenance, and failure recovery, shielding the resource location logic from WAN instability. The brokers maintain persistent connections and can retry or failover if connections fail, thereby improving reliability while enabling comprehensive resource search across sites.
3Ease of operation
If sequential WAN operations are used to locate resources, then connection simplicity is maintained, but productivity decreases due to slow resource location
Solution Approach 1:
The patent pre-establishes WAN connections between datacenters before they are needed for resource location. Connection brokers proactively open and maintain TCP connections across the WAN, so when a resource location request arrives, the communication path is already ready, eliminating connection establishment latency and improving response time while maintaining resource availability across disparate sites.
Solution Approach 2:
The patent transforms the resource location process from a single-site sequential operation to a multi-site parallel operation. By establishing connections to multiple datacenters in advance and enabling simultaneous resource searches across them, the system adds the dimension of parallelism, dramatically improving resource location speed while maintaining operational simplicity through the connection broker abstraction.
Data Source
AI summary
Methods, systems, and techniques for federating operations, in an optimized way using wide area networks are provided. Example systems provide an API for generating and handling federated requests as an aggregation. In one example Wide Area Network Aggregation System, WANAS provides an API, a connection manager, and connection iterators to manage inter-site connections and requests, and inter-pod requests and messaging. These components cooperate to distribute a task across multiple physically disparate sites using a representative connection to the site and fan out of requests to other pods within the site.