The idea of fixing storage challenges by simply adding a piece of software may sound too hard to believe. But it is true: it is that easy. Let us show you the 2 most common types of storage virtualization available today.
Keep in mind that your existing SAN is also made up of software + hardware. With storage virtualization, you are merely upgrading the software piece and keeping the same hardware. That makes this architecture easy to deploy and migrate to.
Therefore, you need to determine if you only want to virtualize existing storage or if you want to add new storage and get rid of your old legacy arrays.
Option 1: Storage Virtualization Node
The first option consists of a storage virtualization controller node. This is typically a 1u x86 server with CPU, memory, network ports, and a raid-1 for the operating system to host the virtualization software.
The next step is to interconnect your existing SAN arrays via iSCSI or FC connections to the new storage virtualization controller. Then you can connect your hosts next and begin to present virtual disks to each of the hosts.
At this point, you only need to open a single management console for your daily storage admin tasks. You still get all the enterprise-level storage services your virtualization software offers, even if the underlying software is not licensed for those features and functions.
If you ever need to bring a newer array and remove an old array, you simply click on “remove” to get rid of the old one and click on “add to pool” to start using the new storage. You can add a second node with dedicated storage to build a highly available storage group with 2 redundant active/active copies of your data.
Option 2: Converged Server SAN
This second option will need a bigger 2u x86 server with higher hardware specs to produce more punching power. You really want to max your ROI by supersizing your disks, memory and networking ports.
If you get the backplane with 24x 2.5” drives, you can pack a lot of capacity using a mixture of 4TB NVMe drives and/or SAS 10k drives. In some cases, we have seen solutions hosting up to 60TB or even 80TB of usable capacity per 2u server.
You still get the same benefits mentioned with the first option, but now you are adding a new twist by mixing new internal disks and existing external SAN arrays under the same virtual pool. You are getting the best of both worlds and still pushing 100k plus IOPS with sub-millisecond response times.
Most organizations start out with the storage virtualization node and eventually convert it to a converged server SAN without incurring a huge CAPEX when adding the internal drives for capacity.
With this setup, all of your performance issues will be a thing of the past, and you can continue to scale by adding more arrays or nodes.
- More redundancy, check.
- Lower costs, check.
- Managed storage from a single pane of glass, check.
- Performance and scalability, check.