Paul Meehan

HDS Converged Management (plus Infrastructure)

Blog Post created by Paul Meehan on Dec 22, 2014

Throwaway remark

 

One of my colleagues made a throwaway remark the other day regarding the Hitachi solution to convergence. The remark stated that “We’ve solved the converged problem by ensuring management is also converged”.

 

The Solution

 

With UCP Director for Hyper-V and vSphere built directly into the hypervisor management layer.

 

This means we provision storage, network, bare metal blades as well as VMs from directly inside the management plane (vCenter and SystemCenter).

 

There is NO SEPARATE GUI required for storage or most network activities, or other element management systems. Everything lives in the context of the vCenter or Systemcenter inventory object so a host has a specific menu which is different to the network level object.

 

And to clarify this is not a "Plugin" or anything like that. This functionality is deeply embedded and context-specific.

 

So the Hitachi strategy is to innovate the management layer as well as put the full stack together.

Furthermore, Hitachi engineers have asked why you need to rip up the datacenter completely and start with a new architecture. The key thing is to simplify and orchestrate the management which is where most of the $$$ get soaked up today.

 

So you can have legendary Hitachi Engineering excellence, support and heritage in form factors as small as our recently launched UCP Director 4000E (2 to 16 blade units)

 

Why have we done that ?

 

So the Hitachi converged solutions can scale without imposing an increasingly more complex management layer.

 

Let me give an example. If you scale your environment from 128 blades to 256 blades, with UCP Director you have the ability to perform orchestrated switch upgrades for Core and Aggregate switches as well as FC, making sure to maintain an available infrastructure at all times using intelligent decision making build into our code.

 

So you dont need to upgrade switches one by one. UCP Director does it, just like Update Manager performs orchestrated remediation of ESXi hosts.


it all seems so logical and simple

 

Here are three examples of stuff that seems like it should always have been there but is available within the UCP Director software, directly in vCenter.

1. Directly launching the KVM console from the hosts view in vCenter

2. Provisioning storage in two clicks, including creating/attaching/expanding a LUN, masking to hosts, zoning FC switches, scanning all hosts, mounting, creating etc.

3. Comparing and remediation of VLAN inconsistencies between virtual and physical switches and applying in one click, updating all physical switch ports in one operation. Who has been in this situation.

 

Best of Breed

 

When I look at Hitachi blades 6 months into my role in HDS, you could call UCP Director layer a true Software-Defined orchestration layer for VMware and Microsoft.

 

I believe this positions Hitachi as the clear technology leader in x86 blade server technology today.

 

  • The only vendor to support logical blade partitioning on an X86 blade.
    • Ask your Oracle customers how much they love this feature. Run 30 instances on a single blade and just license it for the two sockets.
  • We are the only vendor to support super-sized 80-core SMP blades running across four blades.
  • We support both hot and cold standby with complete identity takeover for our blades.

 

Conclusion

 

The flexibility and options our blades provide make them a compelling choice if you want to get a converged platform without the complexity and training you really need to manage such a complex piece of tin.

Outcomes