According to recent study from Cloudability, 86 percent of companies currently use more than one type of cloud services. But before you rush off to the great big service in the sky, it’s vitally important to understand both the pros and the cons and ask yourself all the right questions.
What Are the Challenges?
It’s no secret that data is increasingly growing larger in nature. As a result companies need to pose the question: How do we manage such large amounts of data?
There are a number of different factors that drive the use and popularity of big data, including new ways of linking datasets, creative approaches to visualizing data, and improved statistical and computational methods to name a few. Another challenge that big data has brought into consideration is when do you move the data to the processor (e.g. shipping data from the cloud into a data center) vs. moving the processor to the data?
The second challenge is making sure there are processes and procedures in place to manage the privacy and security of customer data. There is a surge in compliance-related issues, including data breaches, that are toxic to corporate reputations.
Another issue is retention time. How long should you store that data? Not all data is equal. Because there is so much data, containing different information, companies have a hard time determining how long they should store and have access to certain data, and when it’s okay erase it from its storage.
The cost of data storage is another consideration, so you may want to look at tiered storage. For data you need to constantly have access to requires a very fast storage platform, which is one of the most expensive types of storage platforms available. But, if you don’t necessarily need access to all of your data all of the time, you can use a less expensive option.
What Are the Risks?
According to the Ponemon Institute’s (News - Alert) December 2013 Study, the cost of data center downtime across industries is approximately $7,900 per minute. Whether it’s caused by a natural disaster, human error, or IT failure, how many organizations can really afford that kind of lost?
The big risk to an organization in terms of handling its reputation is to make sure customer-facing data are protected. A high risk area is when mission-critical data and apps lack a focus on business continuity and public disaster recovery. If you are down, your reputation will likely suffer if your competition is down less or not at all. Establishing criteria for recovery time and what the right recovery point is depends on the specific applications. You can’t have a one size fits all.
A business impact analysis forces organizations to look at infrastructure and realize what are the critical apps and data.
Overcoming These Challenges
Figure out the functions of the data, rather than focusing on the content of the data. Function shipping is the process of moving logical services to the physical areas of a system where they can be optimally run.
Move the functions to the cloud so they can be actionable on the data that is stored there. Moving large amounts of data to and from the cloud is a bad strategy – it’s hard to move, especially big data.
For an instant recovery, ship the functions. This means having a DR solution that moves functions back to the cloud along with your data for instant recovery. You want a recovery that allows for an automated, physical-to-virtual conversion so that the physical data center server can be moved as a virtual machine into the cloud.
Choose a solution for how the data will be used. If it’s long-term data storage, then you would go a different direction, such as Amazon Glacier, which is an extremely low cost storage service that provides data and backup for the infrequently accessed data, in which slow retrieval time is acceptable. If you are trying to achieve instant recovery, you will need a very different storage platform.
There are tons of standards around security, especially in the cloud. Ensure the solutions you consider are PCI (News - Alert) compliant, and don’t be shy about talking to auditors.
Finally, do a lot of testing. Test yourself to run your operations in a number of situations and test for security. You will need to test for when your app is running the cloud vs. the data center – any situation you might theoretically encounter.
John Gallagher (News - Alert) is vice president of marketing at Quorum (www.quorum.net).
Edited by Stefania Viscusi