Migrating Your Business to the Cloud – Precautions and Infrastructure Needs

2012 has been the year of transitioning to the cloud, even though it’s a considerably old Internet infrastructure model for businesses. Most executives understand that web space (or cloud space) is not a new invention, but it has new relevance. Looking back, store products showcased in physical boxes with software inside ready for manual downloading were impractical and wasteful delivery methods. The cloud could have provided immediate access and purchases to such software. The same principle applies for internal businesses seeking to streamline processes with data usage.

Retailers slowed the market growth of crowdsourcing, network sharing and software cloud storage. The infrastructure burden is no longer shifted to the consumer. It’s now mostly managed by the server or host. This is a sign of progress, streamlining and internet efficiency. Security protections have matured for server support systems, which give data heavy businesses more reason to migrate to the cloud.

Instead of purchasing additional processors as needs grow, you can now purchase more processing power using a host’s hardware. Ever wondered why Jeff Bezos and the Amazon team began purchasing rural land in your region? They are contributors to the new market of web space hosting. Their servers power the web, like many other providers. The burden of hardware has been minimized, and this can cheapen operations. With sound purchases, your business makes your personnel more easily and more quickly able to eliminate different data types, without ever having to acquire hardware or allowing hardware to sit idle.

The recommendations and fact points below are based on our exposure to IT departments transitioning, or migrating, to cloud based software systems. Part of our research as graduate students focusing on technology management and entrepreneurship was to define the processes and systems involved in this migration process.

Two new technologies are converging, the result of which promises to make businesses that require access to data more productive and profitable: cloud computing and data mining.
Cloud computing is the delivery of computing as a service rather than as a product. Computing resources are maintained offsite in the same way utilities such as electricity, Internet service and telephone service are.

Data mining, as the name implies, is the ability to ascertain desired information from a set of data. Today, it is possible to enable the extraction of knowledge from large, complex data sets—and this is essentially what data mining entails.

Before exploring these two technologies and how businesses will benefit from them, it’s important to understand the current, typical state of affairs for data-dependent businesses.

Firstly, infrastructure is Key.

Businesses that offer software as a service such as Sales Force, for example, typically warehouse disparate types of data. Typically, information such as name, company, title and address is housed concurrently in a single database. Customers log in, and either enter new data, or purge and query existing data through a user interface provided in a web browser over any Internet connection.

This requires hardware and software infrastructure that is expensive to maintain, difficult to ensure uptime, and requires safeguards against corruption. As each customer’s database grows ever larger, such companies must continually expand the infrastructure in the form of things like additional software licenses and servers.

Some companies also have data subscriptions such as litigation and financial data that customers can access through a search engine to perform things like e-discovery, or business intelligence research. In those cases the data is typically growing, which further compounds the need for a growing infrastructure and can complicate otherwise linear business decisions.

In other words, for some businesses, infrastructure growth is directly proportional to business growth. For those that rely on data subscriptions, the equation has many more variables. In both cases, though, data must be first managed, and then mined. This is best possible through cloud integration (cloud computing).

The key benefit of cloud computing is something called elasticity. Elasticity is the concept of on-demand provisioning, which enables cloud computing subscribers to access additional computing resources as they are needed.  A conundrum faced by IT departments is to understand the future need of hardware and software infrastructure.  They must be able to plan for and procure what’s needed according to a forecasted schedule.

At the beginning of the fiscal year, for example, HR and executive management typically require a forecast from hiring managers regarding the number of employees they will hire and when they will hire them. IT can then submit a budget requirement and schedule for hardware and software purchases. Cloud computing can eliminate or diminish this requirement.

This capability becomes exceedingly valuable to data-reliant businesses. Engineering and IT personnel can reduce planning efforts and resources, and focus more on providing the right infrastructure for individual internal consumers. Migration to a cloud computing environment also yields an additional benefit for data-reliant companies: the ability to dynamically diminish computing resources when they are not needed.

Infrastructure requires a non-trivial capital and human investment. Someone has to buy it, someone has to install it, and someone has to maintain it. The intrinsic flexibility of cloud computing translates to flexibility for both engineering and external customers. If data mining from a new data set is temporarily required, additional infrastructure that will afterward be wasted isn’t necessary.
All of this assists in both business planning and execution, through simplification and reduced cost. This is important for understanding the cloud’s potential, but what about upgrading systems for data usage specifically?

Data mining requires a few considerations before migration from a traditional IT environment to a computing cloud. First, it’s important to understand that you can’t simply flip a switch and dump your data directly into the cloud. Second, it’s important to understand that an orderly migration effort is important.

Cloud computing introduces security risks that don’t exist when data is behind your own firewall and your migration plan must account for that reality. There are numerous companies that are well vetted to establish cloud security and monitor your system, which should be researched and accounted in the transition expenses.

Data behind your firewall is subject to two concerns: 1) privacy for each customer from all other customers, and 2) external threats. In a cloud, although you no longer retain the responsibility for external threats, you retain the requirement of protecting customer privacy—and you introduce an additional risk.

You must also contend with preventing the data from being sniffed or compromised as it is transported from the cloud to its destination. Encryption therefore becomes a critical concern.  This typically goes hand-in-hand with another concern: cost optimization.

You want your data to have the smallest footprint possible because a bigger footprint translates to greater computing resources which increases cost. Data compression therefore is highly desirable. What’s needed is a solid set of encryption and compression algorithms.

Data-reliant companies typically also have data hygiene requirements. It’s a simple fact: data is dirty.  It is often mis-formatted, incomplete and inaccurate.  The data cleansing and verification processes and algorithms already in place are much more important in a cloud than behind your firewall. Because of the extra transportation step from your environment to the cloud, you must know that you have clean data from the start. Pre-transfer and post-transfer testing is critical, and additional attention to monitoring is also necessary.

Cloud migration could be adopted almost universally by businesses that rely on data.  Businesses that don’t migrate will lose to competitors who adopt the cloud because the cost of infrastructure maintenance behind the firewall is much greater than operating in a cloud environment. However, that doesn’t mean cloud migration should be attempted rashly. It requires thoughtful planning that involves both engineering and IT.

For example, you should plan to migrate, “mission critical,” operations last to reduce the probability of business disruptions. Infrastructure should be maintained throughout the migration to ensure that it can be rolled back if necessary. Also, careful documentation and testing should be performed during the early stages to ensure mission critical operations are not disrupted during the final phase.

* This post was co-authored with Brandon Chicotsky.