March 2022 Summaries

Firebase Hosting expanded its origin server network to Europe, Asia, and two additional US locations, allowing content to be replicated closer to users worldwide. While its global CDN usually serves cached assets, cache misses previously required requests to travel to a US-only origin, creating higher latency for distant users. The new regional origins have reduced global time to first byte by about 20%, with uncached content in Asia seeing waiting times cut by roughly 50%. The expansion also supports faster delivery of dynamic content by enabling Cloud Functions for Firebase, beginning with Firebase CLI version 11.4.0, to use Hosting’s global origins through regional function deployment; Cloud Run rewrites outside the US were already supported.

#FirebaserFriday launches a monthly profile series highlighting Firebase team members worldwide, beginning with Frank van Puffelen, widely known as “puf.” He joined the Firebase team at Google after his Stack Overflow answers about emerging technologies, including Firebase, gained attention, and he continues to enjoy helping developers there seven years later. His current work includes talks on synchronizing game state and integrating Firebase into Flutter apps, collaborating with Flutter Google Developer Experts to improve Stack Overflow resources, producing monthly Firebase Release Notes videos, and answering community questions. Puf adopted his nickname as a teenager to avoid frequent variations of his surname, and he is also an avid reader currently exploring titles including The Lincoln Conspiracy, Carve the Mark, and The Brothers Karamazov.

Firebase Performance Monitoring helps Android developers assess app quality signals such as cold-start time, which spans from tapping an app icon until the app becomes responsive and its first frame is drawn. Because code from both the app and its dependencies can affect this interval, libraries that initialize early may increase startup time, although Firebase Performance Monitoring’s early initialization enables automatic collection of startup, CPU, memory, and network-performance data without additional developer code. The Firebase library performs tasks including dependency registration, configuration management, installation handling, startup tracking, and sampled system-metric collection during cold starts. Using Macrobenchmark and method tracing, engineers identified optimization opportunities in content providers and activity lifecycle methods, then adopted lazy initialization, moved nonessential work off the main thread, and delayed components such as event dispatch and remote configuration fetching. In tests on a simple Android app running on a 2019 Samsung device with Android API level 29, these changes reduced the library’s startup-time impact by more than 35%, and developers are encouraged to update to recent SDK versions for the improvements.