Why should you consider using cloud storage for at least some of your data and applications? Well, unless you've
been living under a rock for the past six years, you probably have a good idea of the potential advantages that cloud-enabled storage can offer.
Public cloud storage can provide significant economies, as you pay only for the capacity you need and use, and thereby save on the Capex costs of new storage systems as well as the space, power and cooling costs to house and run them. Most providers offer virtually unlimited scalability, allowing your data storage to cost effectively expand to meet the needs of your growing business. Depending on the cloud storage service (not true with commodity cloud storage like Amazon Simple Storage Service, but certainly true with some higher-level services), you may also be able to offload data protection tasks such as backups and replication along with other administrative tasks to the provider, leading to reductions in associated hardware, software and management costs.
But we frequently hear IT managers asking a more difficult question: How can I best decide between cloud versus traditional data center storage when it comes to new initiatives, major growth and expansion, or periodic technology refreshes? Should I stick with the networked (or sometimes, direct-attached) storage approaches my organization has been deploying and managing internally, or should I plan to store at least some of my data in a public cloud?
While there's no single correct answer to this question, it's useful to consider some of the tradeoffs that may lead you to take one approach or the other. Here are eight major decision points to take into account before your organization uses cloud storage:
1. What is the primary use case? Is it backup, disaster recovery (DR), collaboration or primary storage? This is the first and most pivotal question, since the answer will narrow your focus and streamline your decision-making process. For example, the requirements a provider will need to meet for cloud-enabled backups will likely be far less stringent than those for primary storage.
2. What type of information will I be storing? Is it user and application data? Will my applications also run in the cloud? If it's the latter, you may be looking at a storage service packaged with cloud compute, such as Amazon Elastic Block Store (EBS), which is used to support Amazon Elastic Compute Cloud.
3. Am I looking at object storage for cloud-architected applications, or block or file storage for traditional applications? The former will likely be better suited to storage in a public cloud with access via REST-like or SOAP APIs, while the latter will require access via block or file storage protocols.
4. Will public cloud storage meet my needs for data availability? This means knowing the provider's service-level agreement (SLA) terms, and what recourse you have if the terms are not met.
5. What about my data security and privacy requirements? You'll want to review the provider's encryption capabilities, both for data in transit and data at rest. What types of access controls and authorization methods are in place?
7. Can a provider meet my recovery times? If I'm using the cloud for backup and recovery, will the provider's SLA allow me to meet my recovery time and recovery point objectives?
8. What about primary storage? If at least some of the data will be primary storage, does the provider offer a gateway appliance or service to provide the feel and accessibility of local storage? Can you meet peak requirements for IOPS and response time?
If you're not satisfied with the answers to these questions for particular data sets and use cases, then you should go with in-house storage. More than likely, you'll find that some use cases and data sets are suitable for cloud data storage, while others are not. For example, many users are already using cloud storage for backup, especially for data sets they consider to be "non-critical." Remember that cloud storage gateways can serve as a "bridge" between cloud and in-house storage, providing advantages such as secure access, a local cache to boost performance, and support for traditional protocols such as Network File System (NFS) for files and iSCSI for block storage. Some gateways are targeted at specific use cases, such as backup or DR, while others are designed to accommodate a range of storage needs.
More cloud tips from Arun Taneja
The top ten questions to ask when comparing cloud storage services
How to compare cloud storage pricing schemes
While this article has focused primarily on public clouds, you may also want to consider private cloud storage. There are not a lot of options on the market, but you can find them if you look. For example, Nirvanix offers CloudComplete, a managed private cloud storage offering. Most private cloud products today are provider-managed, with the storage nodes and capacity usually residing in your own facilities. You gain the scalability, ease of management and most of the cost benefits of public cloud storage, including equipment Capex savings. However, you will likely end up paying operating costs.
But with storage on the premises, you will enjoy greater control and data security compared to a public cloud deployment, along with an easier path to compliance with industry regulations. Nirvanix also offers the ability to scale out your private cloud storage by adding nodes in other locations, with all the nodes federated under a single namespace. One additional advantage of private cloud storage is that it's a stepping stone to a hybrid model, in which data can be stored and easily moved as needed between private and public storage clouds. Private cloud storage can be implemented with software from the likes of Caringo, EMC Atmos, Mezeo Software or Scality (all allow the use of commodity hardware). While using private cloud storage is still in an early stage of evolution, it is something to investigate if some of the answers to the above eight questions don't add up to an obvious choice.
Arun Taneja is a regular SearchCloudStorage.com contributor and the founder and president of the Taneja Group, an analyst and consulting group focused on storage and storage-centric server technologies.