Hardware Management - Deploying and Managing a Cloud Infrastructure - page 226

found in a server. This is basic knowledge, but even though some of these parts are server
grade and designed to work with any other brand and module out there, that does not
always mean they actually do. There are often inconsistencies in compatibility, where one
part works well with a certain model from a certain manufacturer but not with another.
Sometimes this problem does not manifest itself until later, so firmware or driver patches
have to be made to make the parts fully compatible and remove any bugs. This is where
research comes in, because these things are not advertised or not really generally known.
We have to learn from the mistakes, experiences, and observations of others.
Even prebuilt systems such as ones being sold by Dell and HP sometimes come with a
little conflict between the individual parts, which has to be addressed by patches or fixed
in later models.
Storage Media
We have entered the era of big and even bigger data in terms of scale and analytics. We are
being flooded by a lot of unstructured data as well, mostly from the high growth of digital
archives for compliance-related data (which is mandated by government and standardization
entities), multimedia files (such as video, audio, and digital images), some fixed content, and
the ever-growing data related to social networking.
It is difficult to predict and define what capabilities and benefits we get from cloud storage
technology as it continues to evolve. But the benefits of cloud storage is already obvious in
two areas: storing archival data and storing unstructured data.
A lot of archival and compliance data is unstructured because of the lack of a proper and
universal naming conventions and tagging and indexing schemes. Therefore, unstructured
data is stored as files and not as blocks, so capacity more than management is required for
this type of data.
For decades, economic factors have been major influences on storage initiatives and
storage infrastructures that organizations consider and deploy. Now, cloud computing
provides the flexibility and potential cost savings, with both public and private clouds,
for storing massive amounts of data.
Let's look at a few types of storage media being used in cloud computing.
Tape
Tape storage has traditionally been the go-to media for storing large archival and backup
data, mainly because of economy, with the relatively low cost per gigabyte of capacity. But
it has always been viewed by many as an outdated medium; you have probably seen it being
used in movies to indicate to viewers that they are watching a scene that occurs in the '60s
or '70s. It would probably surprise a lot of people to know that tape storage technology has
continued to evolve. We just don't hear about it a lot, especially if we don't work in large
archival data centers. It certainly isn't advertised in the mass consumer market.
But tape media has continually evolved to keep up with the archival needs of large corpo-
rations. We already have enterprise-class tape robotics , which allow tape drives to be rapidly
changed and selected, and the Linear Tape File System (LTFS), which allows us to search and
index files in a tape drive.