Within the Cloud: Technical Concepts of Cloud Computing - Deploying and Managing a Cloud Infrastructure - page 66

cloud computing but are certainly three of the top factors that organizations look into when
designing or simply considering cloud computing solutions: uptime, scalability, and security.
Uptime
For a cloud service, uptime is one of the most important features. For a customer, the
promise of uptime is quite simple: they log on to the service they are paying for, and if they
can access it, it is “up,” and if they can't, it is “down.” But for a service provider, it could
be much trickier. Downtime can be the result of a larger problem like server or disk failure
or even massive power outages with the backup generators failing as well. Some basic rea-
sons for outages, such as power requirements and network downtime, can be attributed to
larger planning issues, like the location of the data center itself.
As far as power facilities go, in many countries it might be impossible for data center
owners to find more than one electrical power provider, so they would have to spend a lot
more on uninterruptible power supplies, backup generators, and the fuel to power them.
That goes the same for networking and finding a major leased line provider. It's true that
in many countries there are often a few telecom companies operating, but chances are they
are all leasing on the same backbone provider, which means that when the giant everyone is
hanging on to takes a dive, everybody else falls with it. Without access to multiple line pro-
viders, an organization may be forced to spend millions to erect its own infrastructure or
accept the fact that it would not be able to offer certain higher availability levels.
Uptime seems like such a given that people take it for granted, but they have no idea how
much planning and effort it takes to deliver on that promise of a certain amount of uptime.
That is where the cost for adding more “nines” into your uptime percentage comes in, as in
90 percent to 99.9999 percent. As a comparison, consider that 90 percent (one nine) would
equal 36.5 days of downtime per year, 99 percent (two nines) is 3.65 downtime days, while
99.9999 percent (six nines) would mean only 31.5 seconds of downtime per year.
Although uptime is an important factor and it's quite enticing to rack up
those nines so you can have something to brag about to your friends, it is
not always the best investment. It gets exponentially more costly as you
add more nines. Besides, 3.65 days of downtime at 99 percent is negligible
when spread over a year's period.
As the number of nines goes up, so do the costs in subscription. This takes into account
the things mentioned earlier, like the costs of keeping infrastructure like power and network
up against all foreseen and unforeseen events, cyberattacks, and the occasional hardware
breakdown or malfunction.
Scalability
The word scalability is usually used to define cloud computing, and it is that single major
aspect that cloud data centers are designed and built upon. It sounds wonderful and looks
good on paper, but it can be a nightmare to implement and maintain. The idea of infinite
scalability means that the service provider must constantly increase available computing