Before you transfer your data into the cloud, it pays to understand how easy it will be to get it back out, should...
you need or choose to do so at a later date. Transferring your data out of the cloud tends to be more costly and time-consuming than bringing data into the cloud, so it's important to understand providers' terms and services before you commit your data to it.
The cloud market has come a long way in the last several years. New public cloud providers and compute/storage services have emerged. With the increase in competition, providers have made the process of moving data out of the cloud simpler and more cost-effective than it used to be. (Note that we're talking here about moving data out of the cloud permanently, as opposed to backing up or replicating data in the cloud to a different availability zone or provider.)
Here is a checklist of questions you should answer to evaluate the cost and effort required to move cloud-based data away from different providers:
1. What are the provider's terms for moving data out of the cloud? Any data, at any time?
Cloud providers should offer the ability for customers to permanently move their data out of the cloud at any time, without hassle. To be sure this is the case, read the provider's service-level agreement (SLA) for the service you're interested in. As one example, Amazon Simple Storage Service (S3) customers are able to transfer their data out at any time, incurring only a data transfer charge.
2. How can I be sure my data won't be stuck in a proprietary format, preventing me from using it outside of the provider's cloud?
This is the dreaded scenario of the roach motel: Data that goes in cannot come out -- at least not in a usable form. While this was a valid concern in the early days of public cloud computing and storage, it is no longer a big deal when you're using one of the major cloud providers.
In the case of cloud computing, major providers now offer a choice of multiple image formats for exporting volumes and instances from their service. For example, in the case of Amazon EC2, customers can choose from Open Virtual Appliance, Microsoft VHD and VMware ESX VMDK image formats. These options should make the exported instances and volumes usable in each of these environments.
In cases where a provider does not offer the specific data or image format you are looking for, you can typically find a third-party cloud migration service offering that will transfer and convert your workloads and data into the desired format (among such vendors are Unitrends, HotLink and RiverMeadow.)
3. Who owns the cloud-based data?
This is a question you should ask the provider before moving your data into the cloud. Providers should allow customers to retain full control and ownership of their data. Check the provider's SLA terms to be sure.
4. What methods are available for moving data out of the cloud?
Major providers tend to provide two approaches for moving data out of the cloud. One is the transfer of encrypted data over a secure network connection to your data center or another cloud (the provider typically covers the transfer from their cloud service to the Internet). The other is to export data via physical media (portable storage devices), which is useful for transferring out a large amount of data. If you desire, you can generally also contract with a third party that specializes in moving data to and from the cloud.
5. What are the costs of each method?
Digital, network-based transfer is usually the most cost-effective approach, and with the price wars among cloud providers, data transfer and export charges continue to decline. Amazon Web Services has been reducing its prices on a regular basis, and other providers typically are following. Currently, transferring tens of terabytes out of S3 in any given month can run in the low thousands of dollars, while transferring out hundreds of terabytes will run in the mid-to-high thousands of dollars. These charges are sure to keep dropping at a significant rate.
6. How can I be sure that all my data has been permanently deleted to protect my data privacy and security?
This is a tougher question, and you may not like the answer. Upon deletion of cloud-based data, the provider will typically "unmap" public links to prevent remote access, then release the disk space to another customer. Unfortunately, the provider generally does not wipe your data off storage media by degaussing it, nor does it destroy the media. Given that most providers now redundantly store data across multiple facilities for durability and availability purposes, this can be a big issue for data that needs to be protected for privacy or compliance reasons. Some third parties offer data wiping utilities or services, but cannot guarantee that your data will be fully wiped and forever inaccessible to other parties. If a full data wipe is a requirement for sensitive data that's subject to compliance rules, for example, then you should not move that data into a public cloud in the first place.
Cloud data retrieval and access is more complicated than it looks
What to watch for when switching cloud storage providers