Developing affordable disaster recovery strategies enable organizations to maximize data protection and meet the myriad of threats facing today’s data center. These threats include natural disasters, hardware failures, accidental or deliberate data destruction, and ransomware attacks. Appropriately protecting the organization’s data from these threats is IT’s most important responsibility. The challenge is living up to this responsibility while keeping the disaster recovery (DR) total cost of ownership (TCO) at a minimum. The best DR plan in the world remains only a plan if the organization can’t afford to execute it.
At the heart of any disaster recovery strategy is primary storage replication. Most vendors claim to have replication, and many include it for “free.” In reality, most replication solutions are a “bolt-on” to the vendor’s snapshot capabilities, making affordable disaster recovery more difficult to achieve and compromising DR readiness. Extending snapshots for DR doesn’t meet the five requirements necessary for affordable disaster recovery while still delivering maximum data protection.
Top 5 Requirements for Affordable Disaster Recovery with Maximum Data Protection:
- Granular – Setting snapshot and replication policies per volume and per site
- Tiering – automated movement of DR data to the least expensive storage
- Flexible – Replicate from an All-Flash Array to a Hard Disk Array at the DR site
- Maximum Protection – Snapshots and Replication integrated but independent
- Cloud – Replicating to the Cloud can reduce secondary site costs.
You can only meet these five requirements if you take a platform approach to storage infrastructure. Our Storage Engine is a complete rewrite of obsolete storage algorithms and flattening of the legacy storage IO stack. It is a new single-layered storage engine that enables our customers to benefit from minimal storage TCO, maximum data protection, and optimal workload placement. It also enables our customers to benefit from the full promise of classic storage services finally and not “put up with” their limitations any longer.
S1:Replicate is an essential storage service from which our S1 engine extracts the full potential so that you can develop an affordable disaster recovery strategy while not compromising disaster recovery speed or integrity.
In part one of this series, I will cover why the first three requirements are so critical and how StorONE’s S1:Replicate meets the demand. In part two, I’ll take a deeper dive into snapshots and replication to explain how these services should and should not interact as part of your DR strategy. I’ll also cover if, when, and how you should integrate the Cloud.
1. Affordable Disaster Recovery Requires Granular Protection
There are various types of data classes within an organization. Some of this data is more important to the business than other data. Databases and virtual server environments like VMware usually need the highest level of disaster recovery, which may mean a synchronous mirror on-site and then asynchronous replication to a remote DR site or the Cloud. Other data, like data stored on a NAS share, may just need off-site protection. Affordable disaster recovery allows IT to set policies per workload even if those workloads share the same media.
With S1:Replicate, you can specify the replication policy per volume. Mission-critical data volumes can replicate synchronously within your data center or campus and then asynchronously to a remote data center to protect against site disaster. Business essential data may not need synchronous copying, and its policy may only be asynchronously replicating to the remote DR site. Finally, all data can asynchronously replicate to the Cloud or another data center for long-term vaulting.
2. Affordable Disaster Recovery Requires Flexibility
Even though it is included “free,” many organizations cannot use legacy system replication because of the cost, complexity, and inflexibility of legacy storage arrays. The problem is that the legacy storage system vendors usually require that the replication target be the same configuration as the primary storage system and that it comes from them. While popular with the storage vendor salespeople, buying two identical systems forces the organization to purchase expensive AFAs as replication DR targets that sit idle until a disaster strikes. To get a better understanding of the actual cost of legacy-based replication, see our white paper, “The Total Cost of Storage Replication.”
Affordable disaster recovery requires hardware and location flexibility. With S1:Replicate, customers can use different hardware vendors for both servers and storage media. As part of the DR strategy, the customer may determine that only specific critical workloads will return to service until the disaster passes. As a result, they may reduce costs by using a less powerful server, lower-performing flash drives, and less flash capacity.
3. Affordable Disaster Recovery Requires Optimal Workload Placement
Most modern organizations long ago adopted expensive all-flash arrays (AFA) and flash drives (SSD) as their primary storage tier to ensure maximum performance for their various application workloads. Legacy storage vendors use all-flash arrays to hide their inefficiencies that slower, more affordable media like hard disk drives reveals. For a closer look at why AFA vendors avoid tiering, see our article, “Are All-Flash-Vendors Afraid of Tiering?” The right storage software can not only leverage tiering to lower the cost of primary flash storage, but also by using flash optimally for full performance.
Allowing “cold” data to continue to reside on primary storage wastes precious capacity and forces the organization to purchase additional expensive flash storage to store ever-increasing amounts of new data. It also makes it significantly harder to develop an affordable disaster recovery strategy.
An all-flash primary data center forces IT to have an all-flash array in their DR site that they may only use a couple of times a decade. An affordable disaster recovery strategy requires leveraging flash to flash replication of active data and tiering data quickly to hard disk drives at the DR site to reduce disaster recovery TCO.
S1:Replicate, because it is an integral part of the Storage Engine, can leverage our intelligent auto-tiering to move data received at the DR site or Cloud to a hard disk tier to reduce the DR TCO. Flash is needed for only the most active data, and the bulk of data is transparently stored on hard disk drives, which are 10X less expensive than flash media.
Conclusion
StorONE’s S1:Replicate feature solves one of the biggest problems of using primary storage for DR; the cost. It also makes DR operationally simpler, which reduces TCO. When a disaster occurs, the DR site has instant access to production data in its native format. Recovery consists of connecting applications to the volumes on the DR storage system and promoting them to production.
In the next part of this series I’ll discuss the fourth requirement “maximum protection – Snapshots and Replication integrated but independent,” and compare snapshots with replication to explain why snapshot based replication jeopardizes your ability to effectively recover from disaster.
In the meantime, join us tomorrow for “5 Ways to Minimize the TCO of Disaster Recovery,” a live discussion on developing an affordable DR strategy and a live demonstration of S1:Replicate. Register now and we will send you a copy of our latest whitepaper “The Total Cost of Storage Replication,” soon after you register.
