repository·master·Indexed 25 days ago
https://github.com/aws/aws-sdk-java-v2A complete rewrite of the Java SDK designed for improved performance via non-blocking IO and pluggable HTTP implementations. It includes Maven archetypes for quickstarting client applications (archetype-app-quickstart) and Lambda functions (archetype-lambda), support for GraalVM Native Image, and an alternate syntax for event streaming APIs to simplify bi-directional communication over HTTP/2.
The SDK metrics system is split into two primary modules:
core. Because sdk-core depends on metrics-spi, it is included automatically in your project.cloudwatch-publisher, csm-publisher). You must add these dependencies manually to use them.The DynamoDB Enhanced Client is a high-level library designed to provide a more idiomatic Java experience compared to the standard DynamoDbClient. It simplifies data access by handling conversions between Java objects (POJOs) and DynamoDB items, supporting common Java data types (like Instant or Number) that the standard client requires manual conversion for.
Key capabilities include:
Note: This client is a 'Level 2' library, meaning it is service-specific and built on top of the generated DynamoDbClient. For full implementation details and usage, refer to the DynamoDb Enhanced Public Preview Library.
The S3TransferManager is a high-level, asynchronous, and non-blocking library built on top of the standard S3 client. It is designed to be the preferred solution for uploading and downloading S3 objects because it is more intuitive and generally more performant than using the low-level S3 client directly.
Key features include:
{Service}BatchManager (e.g., SQSBatchManager) to facilitate batch operations on low-level service clients. This utility is separate from the standard service client and operates similarly to TransferManager. It is designed to handle batching logic, including manual flushing and buffer management, without replacing the core client methods.The AWS SDK for Java v2 is introducing a new progress listener mechanism for TransferManager-initiated uploads and downloads. This feature aims to provide an intuitive way to track transfer progress (e.g., for displaying progress bars) without the confusing overloaded parameters found in the Java SDK v1 ProgressListener.
Key goals for this implementation include:
TransferManager.ExecutionInterceptor system.The SDK uses a MAJOR.MINOR.PATCH[-QUALIFIER] versioning format. Unlike strict semantic versioning, PATCH releases may include new features. Use the version components to assess upgrade risk:
PREVIEW).Request presigning allows a signature creator to use their secret signing credentials to generate an AWS request. This presigned request can then be executed by a separate signature user within a fixed time period without requiring additional authentication.
There is a distinction between a presigned request and a presigned URL:
GET HTTP method.content-type or x-amz-* headers.For example, an S3 GetObjectRequest is only a presigned URL if it excludes specific headers like x-amz-server-side-encryption-customer-algorithm, x-amz-server-side-encryption-customer-key, x-amz-server-side-encryption-customer-key-MD5, or x-amz-request-payer. Including these would cause a signature mismatch in a browser because the browser would not send those headers.
To generate presigned URLs for S3 operations (like GetObject or PutObject), use the {Service}Presigner pattern. The SDK provides specific request and response shapes for each operation to ensure type safety and ease of use. You can use either a builder pattern or a lambda-based Consumer approach to configure the presign request.
// Using the Builder pattern
s3.presignGetObject(GetObjectPresignRequest.builder()
.getObject(GetObjectRequest.builder().bucket("bucket").key("key").build())
.signatureDuration(Duration.ofMinutes(15))
.build());
// Using the Consumer pattern
s3.presignGetObject(r -> r.signatureDuration(Duration.ofMinutes(15))
.getObject(go -> go.bucket("bucket").key("key")));