Cloud Computing and SOA Convergence in Your Enterprise: A Step ...

Moving Onward 227
Performance related to network latency typically arises around large data
sets being sent to and from the cloud computing provider. The larger the data
set, the more likely network performance issues will come into play. You can
take steps to mediate the issues around network performance by understanding
the network performance issues that may arise through performance modeling
and testing and by designing the applications, services, data, and processes to
minimize the use of the network. Since you have to pay for bandwidth, you save
money as well as make your overall architecture perform better.
Network performance modeling is the concept of creating models that
determine how the network will perform under an increasing load and using
that model to determine saturation points. It is a good idea to perform this
step, at least at a rudimentary level, to understand any network performance
issues that will affect your cloud computing solution before they happen.
However, given the “bursty” nature of the Internet and the wide variations in
network performance from time to time, this is not an exact science.
As an option, some cloud computing providers allow you to leverage
dedicated connections into the data centers that host the cloud computing
platform, providing a dedicated pipe between your enterprise and the cloud
computing provider. This, of course, tremendously decreases the chances
that you will have performance issues. However, it also increases the cost.
Platform performance centers on the performance of the cloud computing
platform itself. Most cloud computing platforms leverage a shared, multitenant,
virtual architecture. You have your own virtual space, or virtual
machine, but you share processors and storage space with hundreds, perhaps
thousands, of other users on that cloud platform. It is conceivable that the
cloud computing platform will become saturated from time to time, and
thus performance issues could arise.
There is not much you can do about this, other than work with your
cloud computing provider to insure that you get the performance you require.
Many enterprises insist on service-level agreements, or SLAs. These
are legal agreements that require the cloud computing platform to provide a
specific level of performance.
Other things you can do to protect yourself include extensive testing of
the cloud computing provider at different times of the day before you accept
the platform. Or, you can simply call existing customers of that cloud computing
provider and ask them specifically about performance. There are no
guarantees here, and one of the downsides of leveraging a cloud computing
platform is that you do not own or control those systems.