Analyst Briefing: Virtualization
The Driver Behind the Next Phase of Consolidation
Hamish Macarthur, CEO and Co-Founder, Macarthur Stroud International (www.macarthurstroud.com)

We are living in changing times. Not only are there movements in the financial markets, but there are also changes afoot in the computing infrastructure. The key driver to these infrastructure changes is the need to reduce the cost of managing and operating systems, bringing cost targets in line with the business.

Information management must contribute to the profitability of the business. Supported by an adaptable system infrastructure to scale and flex with business priorities, computing services must be continuously available, robust, and secure so that no data is lost and service levels are assured at all times.

The first phase of system consolidation was supported by the introduction of storage networking in the mid-nineties. The number of servers was reduced, while storage arrays were rationalized to deliver better utilization. Significant return on investment was achieved in the process.

However, there are still further gains to be realised and virtualization is enabling this. It delivers greater levels of consolidation through better utilization of servers and storage. This helps to simplify management of the servers and storage platforms, contributing to a reduction in operational costs and capital expenditure.

Further consolidation with the implementation of virtualization leads to a reduction in the number of servers, storage arrays, and other system devices. This in turn means that a smaller footprint is required, reducing estate costs. Equally important in today’s world of high and increasing energy prices is the resultant reduction in power and cooling costs. And everyone in the IT department can make their contribution to the environmental policies of their organizations.

Where Virtualization Resides
Virtualization resides in five levels of the system architecture:

• Desktop
• Application
• Server
• Storage
• Network

Implementing virtualization is a journey. It is important to realise that the consolidation and management issues will be achieved step by step; at every stage the resources will be better utilized, the service levels will be maintained, and costs will be contained or reduced.

In this article, we are concentrating on storage virtualization. The initial view of storage virtualization is the ability to establish a single pool of storage. But there are several facets to storage virtualisation, from allocating and provisioning through to security, data protection, and archiving.

Storage Virtualization
Storage virtualization supports the next wave of consolidation, new levels of data protection, migration of data to lower cost storage, a reduction in volume of stored data, and a significant contribution to ‘green’ credentials.

Building on the concept of a single pool of storage, the other areas of storage virtualization all contribute to better management of resources and continuous business operations.

Dynamic Thin Provisioning of Virtualized Storage - thin provisioning allows the implementation of the minimum amount of physical storage, yet creates the illusion to the application that it has enough or limitless storage available. As the demand for capacity grows, the virtualized allocation of storage can automatically grow to the maximum threshold capacity requested by the application. Once the appropriate rules, threshold, and permissions have been set up, subsequent management tasks are considerably simplified.

Delivering continuous data availability - virtualization has enabled new techniques and technology practices to be deployed that will move most businesses to a continuous data protection model, enabling users to rewind their systems or recover information with the minimum risk of data loss. Examples of how this is delivered

• RAID provides automatic rebuilding of data in the case of one or two disks failing
• Point-in-time copies or snapshots take a copy of data that has changed
• Virtual Tape Libraries (VTLs) enable disk-to-disk (D2D) back-up, reducing back-up times and allowing for faster restore times
• Data replication processes will form the basis of new continuous system operations

With virtualized servers, one issue to consider is how to replicate or move a system and the data to then run it on a new or different system platform. This process can also enable critical system management functions such as:

• Upgrading systems to new servers and disk arrays
• Internal system testing, ensuring that new software versions and patches can be fully tested before they are deployed in the production environment
• Moving the complete image of a system without having to worry if this is being moved to, or being moved from, a physical or a virtual environment

System imaging technologies—application and data images to be taken and migrated to another system. Recovery times need to be rapid and the recovery process needs to be complete so that users experience minimum disruption. An automated process will be more timely to complete than a manual process subject to human error. Examples of how these technologies can be applied include:

• Application and system images being restored to the same or different servers
• Applications can migrate easily between physical and virtual systems
• Systems can be recovered or rebuilt on ‘bare metal’
• Images can be recovered from different types of storage, including direct attach disk, NAS, SAN, and tape
• Applications can be migrated to fewer servers overnight, powering down the unused servers and saving in energy costs

Reducing volumes of duplicated data—de-duplication addresses the need to reduce the volumes of duplicate data that are stored. De-duplication identifies the same files, and sets up a process to keep one copy and provide a tag in all other occurrences of this data directing the system to the master image. In the case of in-band (as opposed to postprocess) data deduplication, this results in less disk space required to store the data, faster back-ups, and a lower number of duplicated copies of data.

Virtual Tape Libraries (VTLs)—Virtual Tape Libraries write data to disk as if it were writing to tape, with the primary benefits being speed and the completion of back-up within normal working hours. Initial and multiple back-up copies can be held on the VTL for fast recovery. Virtual tape functions are aligned with off-loading the data to tape. Tape becomes a secondary back-up to enable copies to be moved offsite or archived for longer periods. By introducing VTL into the back-up process, back-up windows are dramatically reduced.

Data management costs versus physical hardware— managing the data resident on the disks and tapes costs up to five times the cost of the hardware. Management is necessary to allocate resources, to complete data protection practices such as backup and recovery, track and report activities, and where possible, categorize data for long-term storage, regulatory, or disposal purposes. Virtualization builds on these foundations and enables CIOs to address the issue of data management before it becomes a major business problem.

Data migration and tiered storage - moving and migrating data to the most appropriate locations is a fundamental element of effective data management. This arises for cost and system management reasons. Examples for data migration include:

• Moving data that has not been accessed for a long time to a lower cost storage device
• System testing, where new application software can be tested with real data sets before it is deployed in the production environment
• Data protection and additional system images that can be maintained on lower cost SATA, MAID, or tape solutions

File systems and storage architecture—Establishing a single file system, with a global namespace, enables the system to scale and grow without users noticing the physical changes. Attention can be focused on providing a service to users, whether they are supported by one system at one location or supported by physically separate disk arrays or file servers. Data protection, data security, and de-duplication can be implemented consistently across the system architecture with centralized management policies.

Businesses are demanding a better return from their investments in IT.

As a result virtualization technologies are being implemented across all layers of the system architecture as this enables the consolidation and sharing of IT resources through a single, easy-to-deploy management interface. Virtualized services operate and offer benefits at many levels—sharing servers or storage devices, supporting secure access to systems, enabling safe failover, and restoring data to a trusted state. Virtualization is the way to the next stage of consolidation.