Research:Scheduling Service Request for Improving User Experience and Balancing Resource Consumption
This page documents a research project in progress.
Information may be incomplete and change as the project progresses.
Please contact the project lead before formally citing or reusing results from this page.
In recent years, Wikipedia has faced increasing huge requests from all over the world. In some cases, it is difficult to response every user's request in a very quick time. Long network RTT and limited computing resource are two main factors responsible for this.
This research is aimed at dynamic launching some local service instances for quick response time when requests burst and stopping unnecessary ones for saving energy when requests drop such as weekend or holidays.
Methods
editStep 1: Define the objective function for different type of user. Mainly about expected response time.
Step 2: Measure service cost per request for calculating needed resources.
Step 3: Test average response time at acceptable load for meeting temporal demand of response.
Step 4: Dynamic estimating requests volume with geographical aggregation and maintain the RTT table.
Step 5: Allocate computing resources and scheduling requests under response time required.