Cloud storage has been one of the hot buzzwords of the past couple of years. But like any phrase subjected to the effect of industry hype, confusion has reigned over what is actually meant by the term. In this Storage Basics briefing, we'll clarify the different types of cloud storage services and infrastructures that exist.
Using Amazon S3 and Nirvanix CloudNAS for data storage
In some cases cloud storage is precisely that: data storage in a remote cloud with little in the way of bells and whistles. Typical of this is Amazon's Simple Storage Service (Amazon S3). It holds data for customers ranging from consumers to blue chip enterprise names (such as The New York Times, which uses Amazon S3 as an archive to deliver articles) and the NASDAQ exchange.
Pricing for Amazon S3 starts at a few pence per gigabyte and decreases with the amount of data held. Amazon requires customers to use its APIs to connect to its storage using REST and SOAP Web services protocols.
Nirvanix offers a different method of sharing its cloud storage resources. Instead of using an API to access its network, customers install Nirvanix CloudNAS software on a Linux or Windows server that acts as a virtual NAS gateway to its cloud using common file systems and protocols such as NFS, CIFS and FTP.
Cloud backup services
A number of companies offer dedicated cloud backup services. These include the kind of features you'd expect to see in a traditional backup product -- scheduling, reporting and file recovery -- the only difference being the backup target is the vendor's cloud.
Cloud backup services include Carbonite's Online Backup, EMC's Mozy, EVault (a Seagate venture), Iron Mountain Digital and Symantec's Online Backup. A Web search for "online backup" will show numerous companies offering that service, many of whom repackage the large vendors' online backup services.
Another variation on backup to the cloud is the option to set the cloud as a storage target. For example, earlier this year CommVault added cloud storage as a data backup target to Simpana. Simpana uses the REST protocol to integrate with cloud offerings from Amazon, Iron Mountain, Microsoft and Nirvanix.
Other vendors also offer the cloud as a storage target option. EMC's NetWorker, for example, offers a Cloud Backup Option to back up to storage services based on EMC's Atmos infrastructure.
With a foot in two camps -- backup products that offer cloud as a target as well as cloud backup services -- Symantec offers the Symantec Protection Network (SPN), as well as the ability to specify it as a backup target from Backup Exec.
Cloud archiving services
Some vendors offer a cloud storage service tailored to the needs of data archiving. These include features such as search, guaranteed immutability and data lifecycle management.
Iron Mountain's Virtual File Store (VFS) cloud archiving service offers customers the ability to send data for retention in case of a disaster recovery (DR) scenario. VFS uses a customer site appliance to move data to Iron Mountain's data stores.
Features include the ability to mark data as read-only or write once, read many (WORM), as well as to set lifecycle management policies, access controls and reporting capabilities.
VFS meets the regulatory need for long-term, verifiably immutable data retention; however, it lacks a search function, and requires customers to develop scripts or use a third-party data mover to migrate data to Iron Mountain.
Autonomy Zantaz's Digital Safe archiving service also uses an on-premises appliance, but unlike Iron Mountain it connects to customers' data stores via APIs and allows search.
Digital Safe provides read-only, WORM, legal hold, e-discovery and all the features associated with enterprise archiving. Its appliance carries out data deduplication prior to transmission to the Zantaz data repository. It is, however, considerably more costly than any other cloud storage offering.
Internal cloud/Private cloud
Industry definitions of what constitutes an internal cloud are not yet settled. And PR people have tended to try to shoehorn all kinds of technologies into the cloud buzzword, ranging from scale-out NAS to file virtualisation.
But possibly the best candidates for the internal cloud or private cloud storage title are the likes of Bycast's StorageGRID (acquired by NetApp in April), Cleversafe's dispersed storage and EMC's Atmos. These offerings are designed as cloud infrastructures for use by an organisation behind its firewall and allow companies to build storage on a massive scale. They are characterised by dispersed physical storage and the use of object-based addressing to direct data rather than the use of a file system.