Given the criticality of primary storage and the affordability of alternative solutions, IT must consider developing a standby storage strategy. Standby Storage is a storage solution that can, in the event of a primary storage system failure, (either because of a hardware issue or firmware bug) “stand-in” for the primary storage.
IT should look for several critical capabilities in these solutions:
- Backup Class Affordability
- Production Class Availability
- Production Class Performance
- Maximum Flexibility
Why You Want Standby Storage
If it delivers the above capabilities, a standby solution enables IT to protect itself from the worst case disaster—the complete failure of a storage system —which forces hardware replacement and full recovery from backups. While most organizations will buy a four-hour response, it is important to realize that it is a four-hour response, not a four-hour resolution. Even after the primary storage system is returned to an operational state, it may take a day or more to restore critical applications and return them to operation. In total, a primary storage system failure is typically a two-day outage, which is not enough to cause a business to go out of business but enough to cost the organization significant loss of revenue and productivity.
A storage system failure is in some ways more problematic than a total site disaster. With a storage system failure, everything else in the data center is working. Users and servers are available, but they can’t access data. There is also the genuine concern of rushing through the recovery process only to find out that the supposed fixes didn’t work. Even if the restoration does work, the pressure to recover quickly means that IT can’t spend the time necessary to diagnose what went wrong.
Recent advancements in backup and replication software make preparing a standby storage solution more practical than ever. IT can easily position and re-instantiate virtual machines on the standby system while improving their standard backup and recovery process. Developing a standby storage strategy should be part of a modern disaster recovery plan.
Standby Storage Needs Backup Class Affordability
Standby storage needs to be much more affordable than the primary storage system it plans on supporting. Otherwise, it makes more sense to buy a different storage system from your primary storage vendor, which most enterprises can’t fit into their budget and is why they are exposed to a storage system failure. Vendors can make standby storage affordable by first making sure the upfront cost of the system is affordable. These systems should leverage a hybrid configuration, not all-flash. Most of the data that resides on them will be dormant until there is a failure event.
Another way for these solutions to demonstrate affordability is to have a “day job” plus the ability to extend into production class capabilities. An ideal example is a backup storage target that can help reduce backup costs, shrink backup windows and improve recovery times while also being ready to become production storage.
Standby Storage Needs Production Class Availability
IT can’t risk failing over crucial infrastructure components and then have the standby storage system fail. The standby solution needs dual controllers for high availability, and it needs protection from media failure like RAID. But as we will discuss in our webinar, 4 Reasons RAID is Breaking Backups and How to Fix Them, legacy RAID won’t work for the standby solution. These standby systems’ “day job” is to be a backup storage target, and they need to be affordable, which means they should and will use high-capacity hard disk drives. However, using legacy RAID may mean days of rebuilds if a drive fails. A superior protection method is required.
Standby Storage Needs Production Class Performance
Performance is an area where legacy backup storage targets fall woefully short. While the standby storage solution will leverage hard disk drives to keep costs down, it should have a small flash tier so that during an instant recovery or replication recovery, it can deliver performance equal to the application and users’ expectations. The challenge is tricky since most legacy storage systems require dozens of flash drives to deliver high performance. The software that drives the standby storage solution has to be more efficient than the primary storage system. It has to be able to extract maximum per drive performance.
The efficiency of the software that drives the standby storage solution also enables it to provide maximum flexibility, enabling you to support multiple generations of primary storage systems from a single standby storage platform. The standby storage system should extend from being a backup and standby storage solution to supporting production-class workloads of its own like file serving, virtualization, and even high-performance databases.
Flexibility also means adapting to new hardware as it comes to market. The standby storage system should adapt to support new drive densities and intermix those densities with existing drives without sacrificing capacity or forcing you to create new volumes. It should also support new storage protocols like NVMe-oF as they become available.
How to Get Started
The best place to start with standby storage is by investigating a storage solution for your backup storage. This first step can lower costs while immediately improve backup and recovery performance across multiple backup applications, essentially consolidating your backup software’s data to a single storage platform. An alternative is to design a smaller solution just for your most mission-critical workloads and either replicate to or direct backups at that system. We are happy to work through both options with you to see which is the best fit. You can contact me directly using our Contact Us form. I’ll personally walk you through the options.
You can also attend our next Whiteboard Wednesday session on September 22nd, where we will be “Architecting an Affordable Standby Storage System.” During the session, we will help you in developing a standby storage strategy.