Print this article | Return to Article | Return to CFO.com
Data comes in three broad categories, and taking each one to the cloud demands different strategies to account for the different risks and costs inherent in each. Here's what you should be asking your IT department.
Rob Livingstone, CFO.com | US
September 28, 2011
Fact: People never really know how much stuff they have until it's time to begin packing it all up to move. Similarly, most organizations never really know how much legacy data they have (or where and what it is) until it's time to move it from one platform to another. This is even a bigger challenge when it comes to shifting it all to the cloud.
Most organizations that have been around for a while have tons of legacy data residing on their on-premise IT infrastructure. Broadly speaking, legacy data falls into three categories: structured, semistructured, and unstructured.
Structured data includes sales, shipping, payroll, and accounting transactional data. This data typically is very precisely defined, and is locked-up within enterprise applications such as CRM, ERP, payroll, leasing and finance systems, and so on. It consists of many tables, typically in the form of relational databases.
Semistructured data, on the other hand, is tagged or labeled so that it can be grouped or categorized. One example is e-mail: individual e-mails do not fit into a pre-defined data format, but each e-mail is tagged with sender, date, and so on, which allows them to be sorted and categorized.
Unstructured data includes myriad separate files - often Word files, Excel sheets, multimedia files, and others - spread across network drives and various servers, PCs, and network storage systems. This data has no master catalog or management interface. This is the black hole of organizational data.
So, what should CFOs be thinking about before deciding to shift legacy data to the cloud?
With structured data (because there are no universal data-interchange standards) your organization will be faced with the challenge of mapping and transforming it to a format the cloud system can accept. Even then, there may not be an exact match for some of the data. Specialized data-mapping and migration software is now available, but that adds cost and complexity.
And remember: due to the standardization that makes cloud business possible, most Software-as-a-Service (SaaS) systems have limited-to-no ability to modify their database to accommodate your legacy data. Full data conversion can be difficult and costly, bound by the restriction of the provider's platform's integration offerings (if they exist). On-premise database systems (CRM, ERP, etc.) will rarely be compatible with a
cloud-based SaaS system unless it came from the same vendor, in which case the vendor typically will provide you with a migration path.
Absent that, what conversion process will you need to implement? Have you (or your IT department) realistically costed the migration effort and, if so, how will it affect your project's bottom line? Consider the migration of your structured data carefully. Get a full rundown from IT: the cost could be a showstopper.
The most pervasive form of semistructured data is e-mail, and almost all e-mail systems have standard export/import tools that allow you to shift e-mails from one system to another. This is not a major headache for the average organization.
Unstructured data often represents the largest volume of data within an organization and, surprisingly, it can be the easiest to shift to the cloud. It's essentially simply a job of cut and paste to your cloud storage. However, unlike the traditional shift from one server to another in your own data center, the time it takes to move large volumes of legacy data over the Internet to cloud servers can be hours, days, or weeks, depending on your network and Internet speeds. Together with IT, you have to decide if this is a problem for your organization and plan accordingly.
As with any change to complex enterprise applications and databases, the devil is in the details and comprehensive due diligence is key. If your primary driver in moving to the cloud is to reduce operating costs, you need to be sure (which means IT has tested it) you will be able to shut down your legacy system as you cut over to the cloud. If you can't, you'll be adding a cloud cost to your existing IT costs. Can you recreate information (reports, invoices, contracts, and so on) using your legacy data after you move? If you can't, you've got trouble.
With the exponential growth of data within organizations, and the increasing cost of processing, managing, securing, and storing it all, a move to the comparatively lower costs of cloud platforms and infrastructure-
as-a-service (PaaS and Iaas) probably is inevitable. But between now and then it's important that you accurately balance the cost, risk, and governance considerations involved in moving your legacy data to the cloud.
The cloud may be a pass to get off the support cost escalator; it could be a path out of the legacy data black hole or, if it goes wrong, it could be your one-way ticket out of a job. You owe it to yourself and to your organization to understand the risks.
Rob Livingstone, an experienced CIO, is the author of the book Navigating through the Cloud - A Plain English Guide to Surviving the Risks, Costs and Governance Pitfalls of Cloud Computing. Visit Rob at http://www.navigatingthroughthecloud.com or e-mail him at firstname.lastname@example.org.