What is Cloud Storage?

The IT industry (manufacturers) sometimes gets ahead of itself. Time
and again it's guilty of introducing a technology, generating an
enormous amount of discussion, and then moving on before the actual
implementors, (IT professionals) get a chance to really understand it.
Cloud storage is one of these cases, where many IT professionals are
still asking "what is cloud storage and how should I use it?".

Cloud storage means different things to different people depending on
how it's implemented. The most common implementation is a 'public
cloud', which is essentially storage capacity accessed through the
internet or a wide area network (WAN) connection, and purchased on an
as-needed basis. Users can expand capacity almost without limit, by
contacting the provider, which typically operates a highly scalable
storage infrastructure, sometimes in physically dispersed locations.

Wednesday, November 18, 2009

This expansion ability drives a common expectation that cloud
providers have massive storage infrastructures. As a result most cloud
storage hardware and software systems developers have also focused on
scalability and ease of management. A grid architecture is common,
with storage being made up of clusters of individual servers or nodes
that are coupled together to present a single storage area or single
management point.

This clustering can take two forms, either tightly coupled or loosely
coupled. In tightly coupled clusters, identical, or very similar nodes
are combined to form a single storage pool or file system with
centralized execution of storage functions. Loosely coupled clusters,
on the other hand, can be largely dissimilar nodes coupled together by
a global file system with storage functions being individually
executed across the nodes. Each has their pros and cons and will be
detailed in an upcoming Storage Switzerland article.

The interface to the cloud storage is what the storage manager or the
user will see. This is typically some sort of appliance or software
application that runs locally and then sends data to the cloud. It can
be as simple as presenting a file interface to the cloud storage area,
meaning that it looks like a network mounted drive or it could be
integrated into a backup or archive application.

Cloud storage is a new distribution model, however, with the potential
for economies of scale. Aside from cost, its benefits are outsourced
operation, simple, unlimited growth and 'enterprise' features for
smaller users - like high availability, security, data protection,
etc.

Finally, there is a lot of discussion about the second implementation,
internal or 'private clouds'. This essentially is taking the
capabilities of a public storage cloud, like scalability and cost
effectiveness, and bringing them to the large data center. The use
case could be to create a cloud storage-like service for the
organization that can be accessed from anywhere or it could be as
simple as leveraging a cloud storage solution to build a highly
scalable, easy-to-use NAS.

With the infrastructure in place the next logical question is "what
are the best uses for this storage platform - archive, backup,
collaboration?". This is something that we'll examine in the next
article.