Store and Retrieve Files with Amazon S3

What is Amazon S3?#

Amazon Simple Storage Service provides secure, durable, highly-scalable object storage in the cloud. Using the AWS Mobile SDK, you can directly access Amazon S3 from your mobile app. For information about S3 Region availability, see AWS Service Region Availability.

The AWS Mobile SDK allows you to consume the S3 service in your mobile application via the S3 Transfer Utility, which is replacing the S3 Transfer Manager as of AWS Mobile SDK for Android v 2.2.4.

For information on migrating from the S3 Transfer Manager to the S3 Transfer Utility, see Migrating from the Transfer Manager to the Transfer Utility on the AWS Blog.

For a complete sample that shows how to use the TransferUtility class to perform download and upload tasks, and manage the tasks, see Running S3TransferUtility Sample



You must complete all of the instructions on the Set Up the AWS Mobile SDK for Android page before beginning this tutorial.

Create and Configure an S3 Bucket#

Amazon S3 stores your application's resources in buckets—cloud storage containers that live in a specific region.

Create a Bucket#

  1. Sign in to the S3 Console and click Create Bucket.
  2. Enter a bucket name, select a region, and click Create. Each S3 bucket must have a globally unique name.

Grant Access to Your S3 Resources#

The default IAM role policy grants your application access to Amazon Mobile Analytics and Amazon Cognito Sync. In order for your Cognito identity pool to access Amazon S3, you must modify the identity pool's roles.

  1. Navigate to the Identity and Access Management Console and click Roles in the left-hand pane.

  2. Type your identity pool name into the search box. Two roles will be listed: one for unauthenticated users and one for authenticated users.

  3. Click the role for unauthenticated users (it will have unauth appended to your Identity Pool name).

  4. Click the Create Role Policy button, select Policy Generator, and then click the Select button.

  5. On the Edit Permissions page, enter the settings shown in the following image. The Amazon Resource Name (ARN) of an S3 bucket looks like arn:aws:s3:::examplebucket/* and is composed of the region in which the bucket is located and the name of the bucket. The settings shown below will give your identity pool full to access to all actions for the specified bucket.

  6. Click the Add Statement button and then the Next Step button.

  7. The Wizard will show you the configuration that you generated. Click the Apply Policy button.

For more information on granting access to S3, see Granting Access to an Amazon S3 Bucket.

Configure Your Environment#

To start using S3 in your application, you need to do the following:

  • Add the correct import statements
  • Declare the S3 TransferUtility service in your manifest files
  • Instantiate a Cognito Caching Credentials Provider, an Amazon S3 client, and a Transfer Utility

Declare the Service in AndroidManifest.xml#

Add the following declaration to your AndroidManifest.xml:

  android:enabled="true" />

Instantiate an S3 Client#

Pass your credentials provider to the S3 client constructor, like so:

// Create an S3 client
AmazonS3 s3 = new AmazonS3Client(credentialsProvider);

Instantiate TransferUtility#

You will use the TransferUtility class to upload and download files from S3. Pass the S3 client and the application context to the Transfer Utility, like so:

TransferUtility transferUtility = new TransferUtility(s3, APPLICATION_CONTEXT);


For a complete working sample, see S3 Transfer Utility Sample.

Upload an Object to S3#

To upload a file:

TransferObserver observer = transferUtility.upload(
  MY_BUCKET,     /* The bucket to upload to */
  OBJECT_KEY,    /* The key for the uploaded object */
  MY_FILE        /* The file where the data to upload exists */

Uploads automatically use S3's multi-part upload functionality on large files to enhance throughput.

Upload an Object to S3 with Metadata#

Create a ObjectMetadata object:

ObjectMetadata myObjectMetadata = new ObjectMetadata();

//create a map to store user metadata
Map<String, String> userMetadata = new HashMap<String,String>();

//call setUserMetadata on our ObjectMetadata object, passing it our map

Then, upload an object along with its metadata:

TransferObserver observer = transferUtility.upload(
  MY_BUCKET,        /* The bucket to upload to */
  OBJECT_KEY,       /* The key for the uploaded object */
  MY_FILE,          /* The file where the data to upload exists */
  myObjectMetadata  /* The ObjectMetadata associated with the object*/

To download the meta, use the low-level S3 getObjectMetadata method. For more information, see the API Reference.

Download an Object from S3#

To download a file:

TransferObserver observer =
  MY_BUCKET,     /* The bucket to download from */
  OBJECT_KEY,    /* The key for the object to download */
  MY_FILE        /* The file to download the object to */

Tracking S3 Transfer Progress#

With the Transfer Utility, the download() and upload() methods return a TransferObserver object. This object gives access to:

  • The state (now specified as an enum)
  • The total bytes transferred thus far
  • The total bytes to transfer (for easily calculating progress bars)
  • A unique ID that you can use to keep track of distinct transfers

Given the transfer ID, this TransferObserver object can be retrieved from anywhere in your app, including if the app is killed. It also lets you create a TransferListener, which will be updated on state or progress change, as well as when an error occurs.

To get the progress of a download or upload, call setTransferListener() on your TransferObserver. This requires you to implement onStateChanged, onProgressChanged, and onError. For example:

TransferObserver transferObserver = download(MY_BUCKET, OBJECT_KEY, MY_FILE);
transferObserver.setTransferListener(new TransferListener(){

    public void onStateChanged(int id, TransferState state) {
        // do something

    public void onProgressChanged(int id, long bytesCurrent, long bytesTotal) {
        int percentage = (int) (bytesCurrent/bytesTotal * 100);
        //Display percentage transfered to user

    public void onError(int id, Exception ex) {
        // do something


Pause an S3 Transfer#

If an app is killed, crashes, or loses Internet connectivity, transfers are automatically paused. If the device running your app loses network connectivity, paused transfers will automatically resume when the network is available again. If the transfer was manually paused, or the app was killed, transfers can be resumed with the resume(transferId) method.

To pause a single transfer:


To pause all uploads:


To pause all downloads:


To pause all transfers of any type:


You can also query for TransferObservers, which contain the transfer ID, with either getTransfersWithType(transferType) or getTransfersWithTypeAndState(transferType, transferState). This means that if your app is killed or crashes during a transfer, you can manually determine if there are any paused transfers when the app resumes and handle those as you see fit.

Resume a Transfer#

To resume a single transfer:


Cancel a Transfer#

Canceling an upload or download is simple. Just call cancel() or cancelAllWithType() on the Transfer Utility object.

To cancel a single transfer:


To cancel all transfers of a certain type: