Choosing the right enterprise storage solution | Jan19 Newsletter
What’s tiered storage?
There are two stories about enterprise storage. One is simple and always travels in the same direction: cost per megabyte keeps on falling. The other is more complex and rarely proceeds in the same direction for long: the seat of storage management keeps moving around, from server to storage subsystem to network controller to virtual storage controller.
The two are entwined and it's not at all clear that they’ve reached their destination. Most likely, they never will, but the journey is what matters to IT professionals.
Where are we now?
Currently, storage intelligence resides in hyper-converged infrastructure. This means data stores are commonly under a controller’s programmatic command. The controller will be closely coupled to server and network controllers to ensure the entire system works seamlessly, with everything virtualised for maximum flexibility and efficiency.
The question for IT and infrastructure managers is how best to select and deploy storage hardware that they will not directly control. The answer requires going into storage on a deeper level.
Tiered storage
Tiered storage is a common solution within a dedicated application delivery architecture, with separate storage tiers holding:
- Immediate transactional data
- Online analysis
- Short-term archive
- Long-term archive
Information moves from tier to tier in response to application needs and a data movement schedule. It's an architecture that balances performance and cost and is highly effective in most cases.
Tier 1 is fast and expensive, and hosts mission-critical applications that can’t tolerate downtime or latency. Tier 2 typically hosts cold data, historical information and classified files. Tier 3 is generally an archive. Typically, the lower the tier, the slower (and cheaper) the storage medium.
Hybrid solutions emerge
Most modern solutions use flash or hybrid flash storage for Tier 1, to support the need for fast read/write speeds. Tier 2 typically runs on lower-cost media, like a conventional storage area network (SAN). Tier 3 runs on even slower media, like tape or disk-based backup appliances.
With the continued rise of cloud-based storage, many organisations are moving to a two-tier model, with Tier 1 hosted on flash media and tier two hosted in the cloud.
Hyper-converged infrastructures and multi-app arrays
Hyper-converged architectures use a dedicated storage controller that operates in concert with the server hypervisor and network controller to automate building, deploying, and then tearing down instances of virtual servers.
That’s why running multiple application types on a single hyper-converged infrastructure isn’t recommended. When apps are kept consistent across a server, it becomes more stable and storage needs are more easily reckoned.
It's tempting to try a mix-and-match approach, and cloud services, in particular, offer great flexibility. But regardless of your particular needs and solution preferences, the key is to think strategically about your ecosystem and have a clear understanding of the transaction volumes and other requirements (like security, privacy and backup/redundancy) you need.
This will help you get the benefits you want on the storage tiers you can afford.