February 2015 Summaries
1 posts from Firebase
Filter
Month:
Year:
Post Summaries
Back to Blog
Firebase push IDs are 20-character, client-generated identifiers designed for collaborative applications where many users may append data simultaneously without server round trips or naming conflicts. Each ID encodes 120 bits: a 48-bit timestamp that generally produces chronological lexicographic ordering and 72 bits of randomness that makes collisions extremely unlikely, with the random portion incremented for IDs generated by the same client within a single millisecond. Firebase uses a modified Base64-style alphabet so the encoded IDs sort correctly as keys, but their ordering is not fully guaranteed because client clocks may be inaccurate, although Firebase corrects clock skew after a client connects to its servers. Applications requiring reliable chronological ordering should store and query a server-resolved ServerValue.TIMESTAMP instead. Although push IDs are difficult to predict, Firebase advises developers not to use them as security tokens or rely on their obscurity for access control, recommending securely generated identifiers where unguessable paths are required.
Feb 11, 2015
922 words in the original blog post.