Analysis

posted on 12 May 2008 14:00

VMware has got storage companies falling in line behind it again, 3PAR, Dell, EMC, FalconStore, HDS, IBM, LeftHand and NetApp are all going to create replication adaptors so that they can talk to VMware's Storage Recovery Manager API.

Suppliers such as DataCore, Pillar and many others will probably follow too. Say some 30 individual replication adapters are going to be built and tested for these suppliers and their products.

Let's imagine what the other virtual server suppliers: Microsoft; Citrix; and Virtual Iron, are going to do. They will probably introduce their own automated disaster recovery capability as well and then ask storage companies to write adapters to plug their replication functions into the automated DR APIs of Hyper-V, Xen Server and Virtual Iron. That means another 90 adapters.

There are no standards here and storage companies with more limited storage resources might jib at doing for Microsoft, Citrix and Xen what they have done for virtual server leader VMware. EMC might be against this effort because of a preference for VMware anyway.

As virtual server infrastructures become extended the ability of all storage companies to support all virtual server environments will become stretched and will probably become limited. They may decide not to do it for Virtual Iron, or request cash from the virtualization vendor for the effort.

What some company might also do is to write a layered VMware plug-in that has separate front and back-end modules. The front-end talks to the hypervisor and the back-end to the storage array. Then front and back ends could be modified separately and one piece of code re-used over and over.

This would be more economical than storage companies reinventing the VMware SRM replication adapter wheel for other hypervisors themselves. Maybe a clever European storage software company might do this?

^{[Chris Mellor.]}