valentinT - Fotolia
Downtime is an unavoidable reality of every IT service, but availability concerns seem more acute when they come from an outside vendor. Cloud storage providers are no different. System failures and connectivity issues across the Internet can affect storage services. Evaluate storage availability requirements for stored data and verify that the cloud storage provider can guarantee availability in its service-level agreement.
Microsoft offers 99.9% availability for major Azure storage options; that allows 8.76 hours of downtime per year, or more than 43 minutes per month. By comparison, AWS offers 99.99% availability for stored objects, which allows 4.3 minutes per month. If the provider's service availability doesn't meet your requirements, find another cloud storage vendor with acceptable availability or keep mission-critical data on-premises.
Availability, legal and regulatory requirements can make it difficult to transfer important business data to Cloud-based storage. However, an evolving range of products integrate on-premises and cloud storage resources. The idea is to bring local, remote and cloud storage together into a single unified architecture. Move less-critical data to the cloud outright and keep critical business data on-premises; replicate it to cloud providers if necessary. Tools that connect on-premises data and cloud storage include the AWS Storage Gateway virtual appliance, ownCloud Enterprise software and Oxygen Cloud Enterprise.
Cloud storage services tailored to specific tasks or applications are also an option for many enterprises. Azure blob storage suits most unstructured data like music, backups, big data sets, documents and logs, among others, while Azure table storage is best for databases like NoSQL. Amazon Glacier works well for archival data, elastic block storage is intended as block storage for EC2 workloads, and CloudFront stores content for delivery. Google SQL offers storage for MySQL databases, and Datastore handles NoSQL databases. Ultimately, matching tasks to services provides the best performance with the associated applications.
While the focus often is on better availability, there are many times when storage requirements allow lower storage availability. Data archives, system backups and other data that's accessed infrequently work well with lower-availability services, such as Google Durable Reduced Availability (DRA) storage. The main advantage of lower availability is cost savings. Google DRA is priced at $0.02 per GB versus $0.026 per GB for standard availability storage. Amazon's Reduced Redundancy Storage option for S3 offers lower-redundancy storage for simple tasks where data is already protected elsewhere -- such as file distribution -- and is about 20% less expensive than standard S3.
Stephen J. Bigelow is the senior technology editor of the Data Center and Virtualization Media Group. He can be reached at email@example.com.
Netflix model the gold standard for high availability
Cutting downtime with cloud resiliency calculations
Common standards to ensure cloud uptime
Dig Deeper on High availability and disaster recovery
Related Q&A from Stephen J. Bigelow
Regression tests and UAT ensure software quality and both require a sizeable investment. Learn when and how to perform each one, and some tips to get... Continue Reading
Learn the meaning of functional vs. nonfunctional requirements in software engineering, with helpful examples. Then, see how to write both and build ... Continue Reading
Just because software passes functional tests doesn't mean it works. Dig into stress, load, endurance and other performance tests, and their ... Continue Reading