About the Post

Author Information

Georg Singer is a researcher in the area of cloud computing at the University of Tartu in Estonia. His research focuses on ROI, economic and business aspects of cloud computing.

Risks and rewards of moving applications to the cloud

We have recently come across an interesting discussion group called “Cloudusecases.org”. It has published two white papers titled “Cloud Computing Use Cases White Paper” and “Moving to the Cloud – Risks and rewards of moving applications to the cloud“. Both papers are very interesting and informative. Especially the second one suggests the following very straight forward three-step process for evaluating cloud computing regarding a business and its requirements and needs:

1.   Classify Your Information Assets: Understand the function and value of
the organization’s applications and data and the risks to the organization if
they are lost or compromised.
2.   Determine Your Requirements and Risks: Define the requirements of the
organization and determine if a cloud provider exists that is capable of
delivering those requirements while keeping the risks at an acceptable level.
3.   Calculate Your Return on Investment (ROI): Using the organization’s
needs, assets, risks and requirements, calculate the cost of moving to the
cloud and compare that to your existing costs.

They also propose the following list of candidate applications for moving to the cloud on page 6 of their document:

♦   Pilot Projects: A cloud pilot project is a good way for an organization to
evaluate cloud services to see how useful, reliable and cost-efficient they
can be. A pilot project should be a non-critical application that has a limited
scope, a short time frame and loosely defined estimates of its ROI. Building
a pilot project has the added benefit of giving the organization a chance to
learn how to use cloud services.
♦   Variable Workloads: Some workloads have low requirements the majority
of the time, with occasional periods of very high requirements. An
organization must buy resources to handle the maximum workload, even
though most of the time those resources will be idle. Moving that workload
to the cloud can free the organization to buy only the resources to handle
its normal requirements. When the workload peaks, the organization can
use cloud computing to provision the resources it needs, then release those
resources when the workload returns to normal.
♦   Non-Essential Tasks: Certain applications and data are essential to an
organization’s core mission; they typically have much higher requirements
for resources and much tighter restrictions on how and where they are
used. If there are low-risk applications and data that could be processed
off-site, moving them to the cloud would free resources for the rest of the
organization.
♦   Data Mining: Data mining typically requires substantial hardware to find
patterns in massive amounts of data. Done in-house, the organization must
buy, maintain, power and cool all of that equipment. Moving that task to the
cloud can provide substantial savings. The machines required can be
virtual machines that run only when needed.
♦   Development and Test: Development and testing require substantial
resources when done on in-house systems. Developers must have the
same level of development tools on their machines. Testers must maintain
many different machine environments and test applications on all of them.
Moving development tools into the cloud ensures that all developers are
using the same level of tools, and upgrading the organization to a new
version of the tools requires one upgrade in one place. Doing testing in the
cloud allows the organization to define a single set of virtual machines for
testing; those virtual machines can be started when needed, then shut
down when the tests are complete.

This list is pretty much in line with our findings which we recently published in a blog post titled “To Cloud or not to Cloud – as a function of WAN costs“. We do not agree with the candidate called “Non-Essential Tasks”. Maybe the title is a little bit misleading. Just think about an Internet start-up. The perfect candidate for starting purely in the cloud – little initial investment, scalable computational resources. It certainly has its core tasks running in the cloud right from the beginning. And it makes perfect sense this way.

At the end the white paper also touches the whole aspect of cloud ROI. Calculating the return on investment of a possible move to the cloud is the most important part of a profound and sober decision whether to migrate – or not.Yet this ROI part is a little thin and we are looking forward to an enhanced and updated version of the white paper.


Tags: ,

No comments yet.

Leave a Reply