VPLEX

Hello Alok1947,

Definitely worth studying the free e-Learning that Kate mentions above, it should help set a good foundation for you.

As a brief summary for now, VPLEX brings additional storage virtualization features to the datacenter. Specifically:

1. Online/seamless data mobility
2. Logical Volume Management

Re:  1. Online/seamless data mobility

In a modern datacenter (supporting cloud, virtual compute, virtual network, 3rd platform technologies etc etc) it is not acceptable for data to be tied-down to a physical location (whether that be a set of disks, an entire storage array, or even a site/location). VPLEX allows us to move data in an online and seamless way with no impact to production hosts. For example, we could move an entire application/database from a VNX to a VMAX (any array combination from the 70+ heterogeneous on the support matrix). There are many use cases for this functionality.

Re:  2. Logical Volume Management

When we talk about LVMs we tend to think about host-based LVMs (eg. LVM2 in the Linux kernel, Veritas Volume Manager, and others), or array-based LVMs in the microcode of enterprise RAID arrays. I tend to think of VPLEX as an LVM at the fabric layer (we connect VPLEX engines to production fabrics). What can we do with this?  Well, for example, we can build a VPLEX virtual volume that is a RAID-1 mirror, so 2 mirror-legs, with each mirror-leg existing in separate physical storage arrays; so we are building additional redundancy across array chassis'. We can mirror a RAID-5 LUN in one storage array to a RAID-5 LUN in a separate storage array.... just for example.

To add an important feature. My RAID-1 example could be local, meaning that each mirror leg exists in the same datacenter. However, the RAID-1 mirror could be distributed, meaning that the two mirror legs can be separated by synchronous/metropolitan distances. The really interesting feature is that we support active-active for the distributed RAID-1, with writes at a mirror-leg being mirrored synchronously to the peer mirror-leg (note this is mirroring, not replication). Active-Active only makes sense in combination with a cluster-aware application running on top, such as an asymmetric failover cluster, a parallel cluster, or a cluster-aware filesystem. If that cluster-aware filesystem is VMFS (as one example), then we can vMotion virtual machines across metropolitan distances, whilst the machines continue to run and provide a service to clients.

Re. Your question: how to manage?

We do have a GUI for VPLEX, which is Unisphere for VPLEX.  However and importantly, VPLEX has a lot of functionality and flexibility. This always means that as well as creating best-practice and sensible configurations, you can also create configurations that are not such a good idea (eg a RAID-1 with each mirror-leg have different performance characteristic - we will only run as fast as the slowest mirror leg).

In addition to the eLearning that Kate mentions, we also have a 3-day class "VPLEX Management". (I know, attending classes can be challenging for many reasons).

Anyway, hope this helps.

Richard.