TechPrufer
Get started
Grab
Grab logoGrab · Feature trail

Product Configuration Management SDK Optimization

2 posts · Oct 2024 – Nov 2024

What this trail covers

GrabX's product configuration management SDK has been optimized to reduce initialization time and resource usage. This involved phasing improvements: concurrent loading of subscribed and common data, concurrent download and in-memory loading of large configurations, and eliminating the disk persistence fallback mechanism. These changes have significantly reduced initialization times, with potential for up to a 90% reduction.

All Grab trails
  1. How we reduced initialisation time of Product Configuration Management SDKThe GrabX SDK initialization workflow was optimized in three phases: 1. Concurrent loading of subscribed services data and common data, reducing initialization time by ~80%. 2. Concurrent download and in-memory loading of large configurations within a subscribed service, further reducing initialization time by ~6%. 3. Elimination of the disk persistence fallback mechanism, reducing initialization time by ~50% compared to Phase 2. The cumulative reduction in initialization time is up to 90%.
  2. How we reduced peak memory and CPU usage of the product configuration management SDKThe GrabX SDK was optimized by partitioning configuration data by service, creating separate JSON files for each configuration to maximize S3 throughput, implementing a service-level changelog, and developing a service-based SDK that only subscribes to necessary configurations. This reduced memory utilization by up to 70% and maximum CPU utilization by over 50%.