20TB Hard disk drives are available now. They promise to significantly reduce the cost per TB and floor space consumption of backup and archive storage solutions. George Crump, CMO, StorONE, addresses three common concerns so that organizations may choose their storage solutions wisely and benefit from all that 20TB drives have to offer.
Many IT professionals have grave concerns about using the new drives, and most vendors are not offering them in their solutions, despite the potential savings. Lengthy RAID-rebuild times, the inability to use hot spares, and utilizing all available capacity are all challenges of using these high-density drives.Â
Here we will review and address three primary issues so enterprises can realize all the benefits of 20TB drives.
1. Extended RAID Rebuild Times
One of the top concerns of the 20TB platform is the lengthy RAID rebuild times. If you have a drive failure, the rebuild process may already be taking days with 8TB or 14TB drives. With 20TB drives, you could be looking at two weeks. A next-generation RAID algorithm needs to focus on protecting data, not drives. It needs to understand what data was on the failed drive and use an advanced, distributed parity to rebuild the data quickly. The algorithm should be set to redistribute the reconstructed data in parallel across all the remaining drives instead of bottlenecking to a single spare.Â
In addition to a new RAID algorithm, the core storage software must provide the highest performing read and write operations. After all, read and write operations are a critical component of the RAID rebuild process. However, most companies have not invested in improving the low-level read/write performance of hard disk drives, by their software, in over a decade, despite the advances in hard disk technology. The primary reason for the low investment is that most companies don’t own the code for their foundation.Â
More advanced systems on the market employ unique algorithms to address the RAID rebuild speed issue. As a result, they can rebuild a volume made up of high-density drives in less than three hours (instead of three weeks). In some cases, the algorithm can be extended to flash drives, where rebuild times are less than five minutes. By resolving the high-density drive rebuild problem, the adoption of 20TB drives can be made without compromise.Â
See More: Why You Should Be Shipping Software Updates Like a Cloud Developer
2. Problems Using Hot Spares
When 20TB drives enter your data center, the concept of using hot spares needs to exit it. Legacy storage systems use hot spares so that if there is a drive failure within a RAID group, they can start the recovery process immediately. As discussed above, the administrator will need to begin this process right away because their rebuild times are so slow.
Forcing customers to use hot spares with 20TB hard disk drives causes significant problems. First, hot spares aggravate the already slow rebuild process even further because all the writes of the rebuild are directed at one drive, which further bottlenecks the rebuild process. Second, hot spares waste available capacity. Best practices are to use at least two spare drives per drive size. If you have a storage system with 16TB, 18TB, and 20TB drives, legacy storage vendors will make you create at least one volume and dedicate two drives per drive type.
The workaround is to look for a modern RAID implementation that can use all the drives without hot spares. If there is a failure, the idea is to have the data from the failed drive redistributed across the surviving drives to bring the volumes back to a protected state quickly.
See More: Is Synthetic Data Set To Become a Mainstream AI Strategy?
3. Utilizing the Newfound Capacity
A typical 4U, 84-bay drive enclosure, using 20TB drives, will deliver almost 1.7PBs of raw capacity. However, effectively utilizing this newfound capacity is another critical requirement of a modern storage solution. Unfortunately, most legacy storage software lacks intelligent data placement and metadata management, which means that as these systems’ capacity exceeds 50%, they will start to suffer drops in performance. As the capacity reaches 60% utilization, the performance impact becomes severe. At that point, most customers are forced to add additional hard disk drives to maintain performance.
In backup storage, where most of the data is never accessed, the lack of running at high-capacity utilization levels means that customers cannot utilize 300TBs to 400TBs of storage for each PB of capacity. The investment in high-density drives, for the most part, goes to waste. Look for a backup storage solution that can run at 90% capacity utilization without performance impact. You are paying for 100% of the capacity; you should use all of it, not half! Storage algorithms that use intelligent data placement and efficient metadata management can maintain high performance even when utilization exceeds 90%.
What challenges have you faced in consolidating storage, and what advice would you give for more efficient storage management? Share with us on LinkedInOpens a new window , TwitterOpens a new window , or FacebookOpens a new window . We’d be thrilled to hear your thoughts!