Understanding Cloud Characteristics - Deploying and Managing a Cloud Infrastructure - page 19

Relational database management systems (RDBMSs) primarily dominated storage,
read/write access, and transactions over structured data. The RDBMS would manage the
data stored across a networked storage system and consumers of this data would have
access to it through SQL or SQL-like queries. This was the norm pre-cloud. This is a par-
ticularly good scheme for data storage, but it only works with structured data with rela-
tionships and hierarchies explicitly defined. However, most of the data that is generated
every day is not structured. A huge chunk of the data being generated every day consists of
plethora of visual and audio files, billions of web pages on the Internet, and the plethora
of sensors within wearable devices, and so on, generating data at high frequency. We need
a better way to organize this data and enable fast access (read/write) to it.
REST APIs
REST APIs are the language of the cloud. As discussed previously, every single resource a
public or private cloud offers can be consumed through REST API calls.
Object ID
Since object-based storage means that every data element or object would be at the same
level (flat storage), every object gets a unique identifier, which is used to access the object
from the pool. The physical storage medium is completely abstracted from the consumer of
the data because you do not have to specify a network address or path to fetch a data ele-
ment. The system would know where the object is stored in the pool based on the ID you
provided. This is a huge step up from legacy data storage systems where you would have to
first ping the network storage system, then query it to fetch the address of the data elements
you wanted to access, and then take care of the minute details of the data transaction in
order to have the data read into the application. This has been replaced by a simplistic
object ID-based data object access.
Even though object storage has not been standardized like other technologies and frame-
works, every public cloud vendor incorporates the primary set of features for object-based
storage we've been discussing. There are several startups and enterprise data companies that
have rolled out object data storage system products that target customers looking into build-
ing private clouds. Often, as already mentioned, these enterprise companies would need to
have a gateway between their legacy data storage systems and the new cloud-based object
data storage system until they completely migrate to the cloud. Companies like Infinity
Storage and SwiftStack have object storage gateway products that can be used with object
storage products from HP, IBM, and a host of other companies. Covering the features of
object storage products from these companies is beyond the scope of this topic, but once
you're comfortable with the concepts of object-based storage, it would be easy to adopt and
deploy any product.
In the following sections, we'll cover the essential characteristics of object-based data
storage, and for this, we will use Amazon Simple Storage Service (S3) to explain the concept
in a practical manner. Other public cloud providers, like Microsoft Azure, Google Cloud,
Rackspace, and DigitalOcean, have their own object-based storage options. Then there are