Rich Client Connection Throttling for Server Peak Traffic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Electronic marketplaces face significant challenges in managing abnormal peaks of traffic during promotional events, leading to server overload, bad customer experience, and potential perception of unreliability, with conventional solutions being costly and resource-intensive.
Innovation Solution
Implementing a distributed computing environment with a 'rich' client application that caches content and manages connections by setting time intervals, allowing users to interact with pre-downloaded content during peak traffic, and using contests or games to gradually throttle user participation and reduce server load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the electronic marketplace uses enough servers to handle abnormal peak traffic, then the server capacity and reliability are improved, but the cost and resource waste increase significantly
Solution Approach 1:
The system dynamically adjusts server capacity by implementing auto-scaling that adds servers during peak traffic periods and removes them during normal periods. This dynamic resource allocation maintains reliability when needed while avoiding permanent resource waste, as servers are only provisioned during abnormal traffic peaks rather than being permanently maintained.
2Ease of operation
If the Web server accepts all connection requests during peak traffic, then the customer service quality is improved, but the server overload and system failure increase
Solution Approach 1:
The system implements connection throttling that selectively accepts or rejects connection requests based on current server load conditions. During peak traffic, not all connection requests are accepted - the system partially accepts connections while rejecting others to prevent overload. This partial action maintains server stability while still providing service to a portion of users, avoiding complete system failure.
3Reliability
If the electronic marketplace refuses connection requests during peak traffic, then the server stability is improved, but the customer experience deteriorates
Solution Approach 1:
The system implements connection queuing that accepts connection requests in advance and holds them in a queue when server capacity is limited. Instead of immediately refusing connections, the system preliminarily accepts them and processes them sequentially as server capacity becomes available. This preliminary action maintains server stability while significantly improving customer experience by avoiding immediate rejections and providing eventual service.
Data Source
AI summary
Users are allowed to communicate with a server via a local client even when the server experiences an abnormal peak of traffic. The local client is configured to send a new connection request only after a specified time interval passes from a previous connection. The local client presents content, such as a game, that makes it appear to the user as if the user is interacting with the server while the local client is waiting for reconnection to the server. The user interactions during the time interval are locally stored and later provided to the server for further process.