While many companies are still developing their cloud strategy, others have jumped in with both feet and have moved most of their unstructured data to the cloud. These are organizations in the media and entertainment, online application provider, scientific processing and energy markets. Despite some initial cloud success, as these companies scale their businesses and cloud storage requirements, they are re-thinking this all-in approach. The primary reason for the second look? The recurring cost of cloud storage. Many of these organizations are now looking for a way out, or a least a way to lower their cloud commitment.
The Cloud Storage Challenges
There are three key issues with the long term use of cloud storage. The first, as alluded to above, is cost. While the upfront cost of 100TBs is minimal in the cloud, the continued periodic payment for that 100TBs starts to add up. The second challenge is the predictable performance of cloud storage. The storage system that stores an organization’s data is shared with thousands if not hundreds of thousands of other customers. Unless precautions are taken by the provider of the cloud storage system, a few customers performing a massive ingest of data or a massive read of data, could impact performance. Finally, there is a chain of custody issue. If an organization decides to move data from one cloud to another, that data is susceptible to a “chain-break” during the copy.
What’s a Data Center to do?
The challenges of using cloud storage add up and organizations are looking for a way to do it themselves. Thus far, the cost and expense to store this data on-premises has been too high and the management of it too complicated, but as discussed above the cloud has expense and management issues of its own. The data center needs a simple cost effective way to store this data on-premises, but if an on-site storage option is selected it now needs to support S3 / Object Storage. Organizations looking for a way out of the cloud likely have already re-written their applications to be “cloud-friendly”, they don’t want to have to re-write them again.
An on-premises system also needs to meet or beat cloud storage pricing. High capacity, highly dense object storage systems are one way to achieve this goal, but tape is another. Ideally, the data center will use both; object storage for medium-term retention and tape for long-term retention. The problem is that using multiple storage tiers increases complexity.
To address the complexity issues caused by high capacity, multi-tier storage management, the on-premises solution also needs to be simple, enabling just a few IT professionals to manage multiple PBs of information. Essentially the solution becomes a cloud in a box, which abstracts the backend targets or tiers and presents a unified mount point. One of these mount points needs to be object storage so that companies that have jumped into the cloud with both feet can start wading their way out without changing any aspect of their applications. The “cloud-in-a-box” solution has to support native object and S3 protocols along with legacy protocols like CIFS and NFS.
Cloud storage looks very appealing up front and the small size of the first payment on 100TB’s of storage looks great, but the 50th or 100th payment does not. A cloud is a great place for an organization to start; it is not always a great place for an organization to finish. They need to have a well-developed exit strategy, just in case reality sneaks up on them. A cloud in a box solution that integrates modern and legacy protocols as well as modern and legacy storage devices may be just the place to start.
Sponsored by Fujifilm Dternity, Powered by StrongBox