Had a fascinating chat with Hayden the other day about something that keeps cropping up in my work: data protection and disaster recovery in the wonderfully messy world of multi-vendor storage. You see it everywhere now; organisations have bits of their data estate scattered across on-premise SANs, public clouds like AWS or Azure, maybe some private cloud deployments, and often even older NAS filers hanging around. The challenge? Making all of that play nicely when disaster strikes.
“It’s a right headache, isn’t it?” Hayden started, after I’d set the scene. “Everyone wants best-of-breed, or they’ve grown organically, but suddenly you’re staring at a patchwork quilt of storage solutions, each with its own backup methods, replication technologies, and recovery procedures. Forget about a single pane of glass!”
He’s spot on. Think about it: your primary database sits on a high-performance, on-premise SAN. Great. But your archives are on a cheaper, cloud-based object storage. And your development team uses a different cloud provider entirely. How do you ensure that, if your on-premise SAN melts down, you can recover everything with minimal downtime, and in a consistent state?
That’s where the real problem lies: consistency. You can back up each system individually, sure. But can you orchestrate a recovery across all those systems, maintaining data integrity and application dependencies? That’s a whole different ball game.
We talked a lot about the appeal of the ‘one-vendor’ approach initially. In theory, it simplifies things. One support number to call, presumably seamless integration between storage tiers, and a single set of tools to learn. The reality, as Hayden pointed out, is often very different.
“Vendor lock-in is a real concern,” he said. “You’re betting the farm on one provider, and if their technology stagnates, or their prices skyrocket, you’re stuck. Plus, best-of-breed really does exist. No one vendor is the absolute best at everything. “
He’s right again. I’ve seen companies painted into corners by single-vendor strategies. They end up paying exorbitant support fees, tolerating mediocre performance, and missing out on innovations from other vendors simply because switching is too painful.
So, what’s the answer? For me, and for Hayden, it boils down to platforms that can embrace the multi-vendor reality. These platforms need to offer a few key capabilities:
-
Storage-Agnostic Backup and Replication: The ability to back up data from any source to any target, regardless of the underlying storage technology. Think about tools that use standard protocols like NDMP or APIs to interface with various storage systems. This way, you’re not tied to the proprietary backup mechanisms of each vendor.
-
Orchestrated Disaster Recovery: The key here is the ‘orchestrated’ part. You need a tool that can automate the entire recovery process, from failing over applications to switching network routes. This requires a detailed recovery plan that’s regularly tested and updated.
-
Centralised Management: A single console to monitor backup status, replication progress, and recovery readiness across all your storage environments. Forget about logging into a dozen different management interfaces just to see if your backups are running.
-
Application Awareness: Understanding the dependencies between different applications and data sources. A simple file-level backup isn’t enough; you need to back up the application in a way that ensures it can be restored and run correctly in a disaster scenario.
Hayden mentioned an insurance company he worked with recently. They had a classic multi-vendor setup: an on-premise SAN for their core systems, cloud-based storage for archives, and a separate cloud provider for disaster recovery. They implemented a disaster recovery platform that could replicate data from their SAN to the cloud, allowing them to failover their entire environment in minutes in the event of a disaster. The key was that the platform understood the dependencies between their different applications and data sources, ensuring a consistent and reliable recovery.
Building a robust data protection and disaster recovery strategy in a heterogeneous environment isn’t easy, but it’s essential. Ditching the allure of a single vendor and embracing a multi-vendor-friendly approach can open doors to solutions that are not only more flexible and cost-effective, but also better suited to the unique needs of your organisation. By focusing on storage-agnostic backup and replication, orchestrated disaster recovery, centralised management, and application awareness, you can build a system that provides robust protection, reduces recovery times, and allows you to sleep soundly at night, regardless of what the future holds for your storage infrastructure.