Object storage is an excellent secondary tier of storage. It can drive down the cost of storage, improve the quality of data retention and ensure that data is safe in the event of a disaster. Unlike other forms of storage, it can solve a variety of data challenges that the enterprise faces.
The problem is, it’s only one half of a complete solution. At some point, object storage needs to partner with other solutions in order offer a complete solution to IT professionals. Storage industry players finally joined forces with an approach that should help.
The Missing Puzzle Pieces
There are multiple use cases for object storage. The most obvious is for modern applications that can write directly to object stores. The problem is most of the data is stored on legacy storage systems, in formats created by legacy applications. And the large majority of these applications will never be updated to employ object storage. But they all could benefit from the cost effectiveness and data safety features of object storage.
Completing The Archive
The most obvious use case is archiving: the moving of inactive or infrequently used data from expensive primary storage to object storage. The cost per GB difference in the two tiers is justification enough, but that’s only part of the benefit. IT can realize additional gains because archiving also removes that data from the daily backup flow. This reduces backup time, cuts backup costs, eliminate extra data copies, and generally eases the data protection burden. The missing puzzle piece is the software to analyze, move and provide seamless recall of archived data.
Completing Data Protection Storage
Another object use case is data protection storage, an application where object storage makes a great backup target. In some cases the object storage system can be the sole backup target, in others it can be a secondary, retention tier. Which role it plays is driven by the recovery requirement. One requirement may be to “boot from backup.” In that case, if the primary volume fails the application will boot from a virtual volume on protection storage. Here, the backup target needs to deliver near-production performance, which will require a higher-end, more costly device. Ideally, that type of protection storage should be kept relatively small, with the older data pushed to a less expensive secondary tier – such as object storage — as soon as possible.
This missing puzzle piece for data protection storage is a backup application that can quickly protect production storage and applications, provide access to a relatively high-speed tier for “boot-from-backup” operations, but then move backups to object storage as they age.
Completing The File Data Solution
In many IT shops, the volume of application-generated unstructured data now exceeds that of user file data. Nonetheless, the user data is still a very critical aspect of the services IT provides. It is generated by humans, and humans are more vocal in their demands. They expect data to be available to them at any location, on any device, and they want it backed up. The also want to share data with colleagues and business associates.
These users however are not typically in need of high performance storage IO. Most files are rarely touched, with 60% of files remaining dormant for months or even years. For these files, object storage is a perfect fit. The missing puzzle piece is an enterprise file strategy. A solution that does more than file sync and share but also provides data distribution to remote offices, provides data protection of those offices as well as user end-points, all while giving IT complete control over how data is shared.
The missing puzzle piece here for file data is a solution that not only can provide file sync and share, it also needs to provide data protection, data distribution as well as IT control. In order to leverage the cost advantages and data integrity capabilities of object storage, it needs to interface with object storage, acting as a translation layer between common protocols like iSCSI, SMB and NFS and native S3 based object storage.
Complete Everything Else
There are many more use cases than just the traditional enterprise ones. Industries like healthcare, media and entertainment and energy all use industry specific applications that manage their very large data asset pools. These applications allow for advance cataloging and retrieval of this data when a user needs it in the future. The software applications in these industries all should be able to leverage object storage as their back end storage.
The Problem with Putting it All Together
One of object storage’s key values is its ability to fit into all of these use cases, and, in fact, accommodate all of them at the same time within a single storage pool. The problem is how to make this integration happen. Some object storage vendors have tried to create their own turnkey solutions for each use case. The reality is there are too many use cases — and too much specialized knowledge required — for a single vendor to build well thought-out products in each category. Look at the overall industry: vendors fall into camps, such as enterprise storage, or data protection, or media asset management. The ones who try to do it all rarely succeed.
More commonly, a vendor will deliver one solution for a specific platform, leaving the customer to go and find other solutions for their other data storage needs.
Another solution is for IT to take the whole burden on itself; a do it yourself approach to storage management. The problem with this approach is the time involved investigating all the available solutions and the time required to support them long term, especially if something goes wrong.
A Consortium Solution
A potential solution is for object storage vendors to create a consortium of vendors that will work with their object storage system. The consortium approach is what Cloudian announced recently with its Data Management Partners Program. In this program, Cloudian is partnering initially with Rubrik, Komprise, Evolphin and CTERA to provide fully tested and proven solutions to the various data management challenges IT faces. Members have direct access to each other so that if problems arise they can work together to resolve any issues.
With all-flash systems taking care of the performance issues that most data centers have, the single biggest challenge facing data centers is data management. Data management encompasses not only data archive but data protection, copy data management, file management and industry specific data preservation.
The problem for IT today is while object storage may be a consolidation point for storing all of this data, moving data to the object store is a fragmented mess. But given the variety of use cases and the business benefits achievable, the move is a necessary one consortiums like the one that Cloudian is building are an ideal way to help IT navigate these waters and implement complete solutions with the full confidence that it will work.