Welcome to Storage Virtualization 2.0
by Hu Yoshida on Feb 3, 2008
Back in the early days of storage virtualization, the Storage Network Industry Association, SNIA, defined virtualization as:
“The act of integrating one or more (back-end) services or functions with additional (front-end) functionality for the purpose of providing useful abstractions. Typically, virtualization hides some of the back-end complexity or adds or integrates new functionality with existing back-end services.”
Recently I had occasion to revisit the SNIA site to see if this definition had changed over the last few years. I was pleased to see how they expanded the definition into two parts.
The first part defines virtualization as “the act of abstracting, hiding, or isolating the internal function of a storage system or service from applications, compute servers or general network resources for the purpose of enabling application and network independent management of storage or data.”
The second part adds “The application of virtualization to storage services or devices for the purposes of aggregating functions or devices, hiding complexity, or adding new capabilities to lower level storage resources.”
The first part of the definition is covered by most storage virtualization solutions, whether they are server, network, array, or controller based virtualization. They all provide application and network independent management of storage and data. However, there are different limitations on what storage and data they can virtualize, depending on where they are based. Server based virtualization is limited to the storage and data attached to that server. Network based virtualization is limited to storage that is attached to the network, and array based virtualization is limited to the storage and data contained within the array. Only control unit based virtualization can support, virtualization of heterogeneous FC storage attached to the control unit and to any system that attaches to the control unit through, FC, SAN, ESCON, FICON, iSCSI, and gateways for VTL, NAS, and CAS.
The second part of the definition marks a clear differentiation between what IDC and Gartner call Virtualization 1.0 and Virtualization 2.0. This part is defined by the ability to aggregate functions or devices and add new capabilities to lower level storage resources, while hiding complexity.
Lower level storage resources like JBOD arrays, can be enhanced with the performance and functionality of an enterprise storage control unit through control unit based virtualization An example of this would be to virtualize SATA disks for attachment to mainframe ESCON channels in order to provide a lower cost tier of storage. Without this ability to aggregate SATA with ESCON, mainframe cycles would be required to compress the data in order to reduce the cost of storing this to ESCON connected disks. Another example would be the ability to enable thin provisioning for all storage that is virtualized behind the control unit.
This is not limited to aggregation of functions and devices that sit behind the control unit, it can also aggregate functions that are attached to the front of this platform, with the functions contained within the control unit and attached storage resources. For instance it can aggregate a VTL gateway with control unit replication for remote vaulting between heterogeneous storage systems, or leverage CAS with thin provisioning and remote replication.
InfoStor recently published a case study of a Midwest financial institutions’ experience with Virtualization 2.0 through the use of Virtualization and thin provisioning. The key benefit of this virtualized storage infrastructure was the ability to easily and quickly provide storage and connectivity to a rapidly growing server farm and put off purchasing a new storage array of about 40TB.
Welcome to Virtualization 2.0
[...] Hu Yoshida, the CTO of Hitachi Data Systems, explained clearly about this definition here.. [...]