– Chris Gladwin, CEO of Cleversafe (www.cleversafe.com), says:
Related to storage, two trends in the last five years have increased storage for enterprises. These trends are disaster recovery, and unstructured content.
Legislation such as Sarbanes Oxley, HIPPA, and events such as 9/11 and Hurricane Katrina forced enterprises to become more diligent about protecting their digital assets. Protection comes at a cost, and that cost is increase in storage since typical solutions rely on replication technology to backup data offsite, resulting in multiple copies of data.
Further exacerbating storage is the increase in unstructured content – videos, images, audio files, powerpoint presentations, etc. These types of files have changed the game of storage for enterpises, and are pressing the storage systems and technologies many enterprises have today.
Coupling the increase of unstructured content and with replication has resulted in expensive storage systems. In examining their storage systems, enterprises are realizing the cost to operate and manage the storage is more expensive than the original hardware purchase.
What are some strategies enterprises can use to effectively cut the amount of storage they need?
Most enterprises are working several angles to get a better handle on their storage and storage management.
One approach is to implement virtualization. Historically, enterprise storage has been directly mapped to applications, resulting in inefficiencies because each of the independent applications isn’t necessarily fully utilizing the related storage. Virtualization basically operates above the storage appliances to abstract the location of the data from the physical hardware. This leads to better efficiencies and utilization of storage.
Another technique is to implement a Hierarchical Storage Management (HSM) approach. HSM typically uses policies to determine the frequency of use of files, and automatically moves data between high performing primary storage, and nearline disk storage and tape libraries. Enterprises also need to revisit any “store everything forever” decisions, and change policies to examine importance of data, and for data that will not be leveraged, include deleting it once they are legally not required to store it anymore.
Another area companies are looking at is the approach to data protection in using RAID and replication. Replication does work, but when coupled with RAID, can result in a 3-5 times increase of the raw capacity necessary. RAID typically increases the data by 30% (it depends on the configuration), and if copied offsite using replication once, that results in 260% raw storage necessary. Some organizations are making 2 offsite copies and are up at 390% raw storage.
Enterprises are looking for more efficient methods, such as dispersal, to lower the ratio between usable and raw storage. Dispersal, for example, results in 160% raw storage and provides superior data protection since it can tolerate multiple site failures, drive failures, and network outages that replication and RAID cannot.
Another technology geared at making storage more efficient is deduplication. Deduplication focuses on eliminating redundant data by finding duplicates, only storing one copy, and storing references to the one copy in all other instances.
Looking at storage from the operational cost perspective, enterprises look at data center consolidation, or outsourcing as another method to reduce cost. Also, as disk drives become more efficient, replacing existing storage across each tier can also reduce costs. Many storage products today utilize 1 terabyte drives and are much more power efficient than products available even 3 years ago. Replacing older equipment can reduce power consumption and floorspace.
Most enterprises realize that not one of these techniques is a silver bullet, and typically use them together for maximum efficiency. So, and example storage system may use virtualization for efficiency in application based storage, HSM to manage movement of data across storage tiers, deduplication to compress the data, and dispersal for data protection, and moving the data offsite.