Information Technology Reference
In-Depth Information
potential, a huge waste in available resources. Or unforeseen growth might occur, rendering
the data center inadequate so more money has to be spent on unplanned upgrades.
Cloud computing infrastructure, on the other hand, is built for scalability. Yes, it is true
that when you're building a cloud data center, you will often choose the maximum capacity
allowable by the budget, but that is where the similarity in capacity ends. Cloud data centers
are designed with future upgrades in mind simply because demand will only grow and the
data center has to scale for it. So in terms of a set initial capacity, traditional data centers
and cloud data centers are quite similar, but only until customer demand requires additional
capacity; it then becomes one of the biggest distinguishing points.
They also differ in how resources are being used. Traditional data centers are often
meant for a single corporation with a capacity that is often way beyond the needs of the
organization, but a cloud data center is often created for public use, with hundreds or
thousands of customers using the resources at the same time. In this regard, cloud data
centers often operate at close to maximum capacity at any given time, while a large por-
tion of a traditional center's capacity will often be wasted.
There are many things that distinguish cloud computing from “simple”
scalable, grid, or utility computing, but an easy way to determine whether
cloud computing is being used is to think of two words: scalability and
virtualization . If it is deployed with both of these present, chances are it
is true cloud computing.
Consider the purpose of a data center. If it was made for traditional IT, it would likely be
serving a multitude of applications across different departments, all of which are using just a
tiny fraction of the center's overall capacity and power. It is estimated that most data centers
are only utilized for 10 percent to 30 percent of their maximum potential. Of course, that is a
general estimate. These numbers are totally dependent on the type of use that an organization
has for its data centers; some, albeit only a few, are utilized at 90 percent capacity or more.
But for cloud computing data centers, public or private, it is most likely that there are only a
few dedicated cloud applications being run, and those are meant to perform multiple func-
tions and are being used by a lot of people to which the servers have to allocate resources.
This means that cloud data centers are often operating at optimum capacity.
Even with their major differences, however, in concept the general connectivity and com-
ponents remain the same. You have your servers, which are essentially your computing units,
and you have your storage units arranged in racks or cabinets. Everything is tied together
with the core data center backbone and connected to the outside with various networking
devices like switches and hubs. Figure 3.3 shows a simple design that is still the basic design
used for both traditional and cloud data centers. Each organization or provider will have
its own modifications to this simple design based on their needs, but again, the components
remain the same, although some may be assigned with a slightly different purpose.
Search WWH ::




Custom Search