Amazon S3 has some aspects that can lead to security vulnerabilities. This article goes through some of these, such as the ability to perform certain actions without authentication, alternative ways to retrieve object keys, incomplete multipart uploads that can be deleted but not downloaded, and the potential for accidental public exposure of buckets through misconfigurations.

Amazon S3 now supports conditional writes that can check for the existence of an object before creating it. This helps developers more easily prevent applications from overwriting any existing objects when uploading data. Conditional writes can be used to simplify how distributed applications with multiple clients concurrently update data in parallel across shared datasets. Developers no longer need to build any client-side consensus mechanisms to coordinate updates or use additional API requests to check for the presence of an object before uploading data. The feature is available at no additional charge in all AWS regions.

Amazon S3 can be used as a strongly consistent event store with the help of the new conditional writes feature. Conditional writes enable optimistic concurrency, which is necessary for managing data updates in distributed systems. This article provides a step-by-step guide on how to implement this strategy. It also goes over structured naming conventions and conflict resolution.