Amazon S3 provides a simple web services interface that can be used to store and retrieve any amount of data, at any time, from anywhere on the web. It gives any developer access to the same highly scalable, reliable, fast, inexpensive data storage infrastructure that Amazon uses to run its own global network of web sites. The service aims to maximize benefits of scale and to pass those benefits on to developers.

Before you can get started with Zend_Service_Amazon_S3, you must first register for an account. Please see the S3 FAQ page on the Amazon website for more information.

After registering, you will receive an application key and a secret key. You will need both to access the S3 service.

The Zend_Service_Amazon_S3 class provides the PHP wrapper to the Amazon S3 REST interface. Please consult the Amazon S3 documentation for detailed description of the service. You will need to be familiar with basic concepts in order to use this service.

Zend_Service_Amazon_S3 provides the following functionality:

Once you have registered with Amazon S3, you're ready to store your first data object on the S3. The objects on S3 are stored in containers, called "buckets". Bucket names are unique on S3, and each user can have no more than 100 buckets simultaneously. Each bucket can contain unlimited amount of objects, identified by name.

The following example demonstrates creating a bucket, storing and retrieving the data.

require_once 'Zend/Service/Amazon/S3.php';

$s3 = new Zend_Service_Amazon_S3($my_aws_key, $my_aws_secret_key);

$s3->createBucket("my-own-bucket");

$s3->putObject("my-own-bucket/myobject", "somedata");

echo $s3->getObject("my-own-bucket/myobject");

Since Zend_Service_Amazon_S3 service requires authentication, you should pass your credentials (AWS key and secret key) to the constructor. If you only use one account, you can set default credentials for the service:

require_once 'Zend/Service/Amazon/S3.php';

Zend_Service_Amazon_S3::setKeys($my_aws_key, $my_aws_secret_key);
$s3 = new Zend_Service_Amazon_S3();

All objects in S3 system are stored in buckets. Bucket has to be created before any storage operation. Bucket name is unique in the system, so you can not have bucket named the same as someone else's bucket.

Bucket name can contain lowercase letters, digits, periods (.), underscores (_), and dashes (-). No other symbols allowed. Bucket name should start with letter or digit, and be 3 to 255 characters long. Names looking like an IP address (e.g. "192.168.16.255") are not allowed.

The object is the basic storage unit in S3. Object stores unstructured data, which can be any size up to 4 gigabytes. There's no limit on how many objects can be stored on the system.

The object are contained in buckets. Object is identified by name, which can be any utf-8 string. It is common to use hierarchical names (such as Pictures/Myself/CodingInPHP.jpg) to organise object names. Object name is prefixed with bucket name when using object functions, so for object "mydata" in bucket "my-own-bucket" the name would be my-own-bucket/mydata.

Objects can be replaced (by rewriting new data with the same key) or deleted, but not modified, appended, etc. Object is always stored whole.

By default, all objects are private and can be accessed only by their owner. However, it is possible to specify object with public access, in which case it will be available through the URL: http://s3.amazonaws.com/[bucket-name]/[object-name].

It is possible to get and put objects using not stream data held in memory but files or PHP streams. This is especially useful when file sizes are large in order not to overcome memory limits.

To receive object using streaming, use method getObjectStream($object, $filename). This method will return Zend_Http_Response_Stream, which can be used as described in HTTP Client Data Streaming section.

$response = $amazon->getObjectStream("mybycket/zftest");
// copy file
copy($response->getStreamName(), "my/downloads/file");
// use stream
$fp = fopen("my/downloads/file2", "w");
stream_copy_to_stream($response->getStream(), $fp);

Second parameter for getObjectStream() is optional and specifies target file to write the data. If not specified, temporary file is used, which will be deleted after the response object is destroyed.

To send object using streaming, use putFileStream() which has the same signature as putFile() but will use streaming and not read the file into memory.

Also, you can pass stream resource to putObject() method data parameter, in which case the data will be read from the stream when sending the request to the server.

In addition to the interfaces described above, Zend_Service_Amazon_S3 also supports operating as a stream wrapper. For this, you need to register the client object as the stream wrapper:

require_once 'Zend/Service/Amazon/S3.php';

$s3 = new Zend_Service_Amazon_S3($my_aws_key, $my_aws_secret_key);

$s3->registerStreamWrapper("s3");

mkdir("s3://my-own-bucket");
file_put_contents("s3://my-own-bucket/testdata", "mydata");

echo file_get_contents("s3://my-own-bucket/testdata");

Directory operations (mkdir, rmdir, opendir, etc.) will operate on buckets and thus their arguments should be of the form of s3://bucketname. File operations operate on objects. Object creation, reading, writing, deletion, stat and directory listing is supported.