Global Latency Benchmark for Serverless Redis
This single page application shows the latency for three different Upstash set up:
- Single Region Database in AWS US-West, edge caching
disabled - Single Region Database in AWS US-West, edge caching
enabled - Global database, edge caching
disabledWe have implemented an AWS Lambda function which reads a value from an Upstash database using the REST API. We have deployed this function to 8 different regions all over the world.
This page calls the 8 Lambda functions and each function makes three requests to the database. First request has Edge Caching enabled, the second one is without Edge Caching, third one is onto the Global database. Each function records the latencies of both requests.
Here the latency of your request:
⟳Database Region:US West (N. California)
Global Database Regions:US West (Oregon), US East (N. Virginia), South America (São Paulo), Asia Pacific (Singapore), Europe (Frankfurt)
Client Region | Latency - Edge Disabled | Latency - Edge Enabled | Latency - Global |
---|
US West (N. California) | ? ms | ? ms | ? ms |
US West (Oregon) | ? ms | ? ms | ? ms |
US East (N. Virginia) | ? ms | ? ms | ? ms |
South America (São Paulo) | ? ms | ? ms | ? ms |
Asia Pacific (Hong Kong) | ? ms | ? ms | ? ms |
Asia Pacific (Tokyo) | ? ms | ? ms | ? ms |
Europe (Ireland) | ? ms | ? ms | ? ms |
Europe (Frankfurt) | ? ms | ? ms | ? ms |
As the distance between the server and the client increases, the impact of edge caching increases. See
the blog post for the details.
The latency numbers collected so far:50th Percentile (latency in milliseconds):
Loading Chart
99th Percentile (latency in milliseconds):
Loading Chart
99.9th Percentile (latency in milliseconds):
Loading Chart
See
the blog post|
the source codeLoading...