Dropbox, Carbonite and many others have accelerated the use of Cloud storage in the consumer market for backup of music, photo, file sharing and more recently with various social networking needs and media sites. Similarly, Box.net has established its strong presence within the business segment. Techaisle’s SMB and mid-market business MarketView forecast shows that cloud storage will be a US$1.1 Billion market by 2016 growing at a 37 percent CAGR.
It would not be out of place to say that businesses are always very concerned about regulations and business policies requiring them to retain data for longer periods. Add to this the intricacies involving data protection via backups and replication, data scalability requirements and data availability needs across multiple geographies - the complexities and cost with storing massive amounts of data that is generated across the enterprise becomes huge.
Though it has limitations, an object based Cloud storage solution addresses many of the business challenges above – a scalable, easily replicable, pay-as-you go solution that is geographically accessible through public internet solution, thereby meeting businesses’ most demanding requirements with respect to their data storage policies.
Given the cost competitiveness, scalability and security attributes, backed by enterprise grade SLAs, Object Storage in the Cloud is an extremely viable option for the small and medium businesses (SMBs) looking to migrate their datacenters into the Cloud. RAID arrays are mostly used by mid-sized businesses but that is no protection against a disaster or any malware attacks.
What is Object Storage and why it is significant for mid-market businesses?
An object Storage solution breaks storage data into distinct segments, or ‘Objects’, each containing a unique identifier (or metadata) that allows data retrieval.
Valet parking is often cited as an analogy for Object Storage. When parking at a garage, the attendant gives a claim ticket that identifies the car that allows the driver to pick up the car later. The driver is not concerned where the car is parked as long as it is identifiable when it is time for pick up. Object Storage, likewise, stores data (objects) and retrieves when required based on its unique identifier.
Object Storage differs from traditional Storage Area Network (SAN) or Network Attached Storage (NAS) in that the former is ‘Object Based’ and has the following characteristics:
- Each object has its own ID, metadata, data protection policy and is unlike any file system where files often inherit attributes from their parent containers/files/directories.
- There is no limit on volume restriction of size of file systems – unlimited scalability, not limited on infrastructure capacity maximums.
- Data is accessible anywhere over HTTPs - availability of the service can be anywhere on the internet. Hence this is latency sensitive.
- Data is typically retrieved via a RESTful or SOAP based API Web service. Storage vendors have their own proprietary API platforms – e.g. Amazon S3, Nirvanix APIs, OpenStack or EMC Atmos platforms. Any programming language that supports web service-based API calls to remote systems can be used to build applications around the storage solutions.
- Price, which is normally based on usage, is much lower than traditional Block and File storage solutions. Businesses typically pay a per-gigabyte rate for upload and download and a per-gigabyte fee for monthly storage. In addition, some providers charge for each data access request based on reads, writes, etc.
- Data Archiving and backup: Retain data that needs to be easily accessible and always available but is not constantly used in real-time.
- Data Compliance Requirements: Must keep data safely and reliably for audits, reporting, regulatory compliance, discovery, backup and restore, or disaster recovery.
- Data accessibility: Have a library of content and/or media files that employees in many locations must access to download items. Also need employees in many locations to be able to upload or download files.
- Web 2.0 and Social Media: Manage exploding data growth or have fluctuating data storage requirements - Object storage systems have massive scale and provide moderate performance at low cost.
Medical Imaging and medical records applications also have massive use for Object Storage due to sheer volume of data storage requirements. Healthcare Vertical, particularly, have shown high adoption rates for Object based storage.
The market is yet very fragmented though Amazon AWS Simple Storage Service (S3) is considered the leader in this space. Though it has challenges, AWS is a highly innovative service and has created AWS Storage Gateway that enables hybrid storage architectures that span both on and off premise storage options. Nirvanix, a pure play Cloud Storage provider that offers public, hybrid or on-premise Nirvanix-powered storage services which are priced for various support levels. There are other big names such as Google, Windows Azure Blob, Rackspace CloudFiles, AT&T’s Synaptic Cloud Storage and recently Savvis has also entered the Cloud Object Storage space.
Vendors seek to differentiate themselves on price, quality of services (QoS), SLAs and hybrid architectures. At the same time they tend to gravitate towards some established storage and compute platforms to enable standardization, achieve economies of scale and allow for ecosystem build-up. This allows their business customers to combine their storage solutions with any third party solution that uses a similar platform. For example, AT&T Synaptic and Savvis are aligned with EMC Atmos Storage Platform, whereas providers like HP, Rackspace CloudFiles and SoftLayer are aligned wtih OpenStack platform. This enables any third party solution that is based on the above platforms to be combined with storage solutions offering any custom configurations. AmazonS3 is an exception that is based on its own AWS Storage Gateway.
Justifiably, there is a great deal of hype today around Object Storage, especially relating to its Cloud, Social Media and Big Data applications. However, it is important to understand the specific use cases and workloads Object Storage can be useful for, given its limitations such as Latency sensitivity, lack of standardization among object storage interfaces and in some specific uses where the stored data is modified frequently and hence not suitable for an Object Storage solution.
Upcoming report: Techaisle's Cloud Object Storage Competitive Landscape report.