Phases of File Virtualization

The tactical data migration today can be the strategic infrastructure of tomorrow. By Jack Norris, Senior Director, EMC

Organizations need to address the enormous amount of fast growing unstructured data. With file virtualization, administrative tasks that used to be disruptive and require a great deal of end-user coordination can now be executed in 1/10th the time with no impact on end-user access (TheInfoPro Wave 8 Storage Industry Study).

Virtualization drives dramatic short term benefits with uses ranging from consolidations, tech refreshes, archiving and tiered storage deployments. With virtualization, payback can be measured in weeks. Virtualization is also a strategic part of a long term infrastructure. But with many of the initial use cases for virtualization focused on specific tasks, where do you start? What are the critical criteria you need to look at it when evaluating file virtualization solutions?

It is important to realize that there are distinct phases for file virtualization. As organizations move from short-term tactical uses to longer term strategic, the architecture and integration requirements for virtualization change as well. When evaluating various virtualization solutions, it is important to acknowledge how you plan to use virtualization initially as well as how it will fit into your architecture and storage strategy in later phases. To a large degree, these phases are dictated by the state of the environment. Typically before virtualization is first used in an organization, storage management costs are high, data migrations are difficult, and utilization rates are relatively low. The first phase for file virtualization in these environments is as a migration service. Contrast that picture with the third phase of virtualization where data flows freely in response to changes in the business environment, flexibility is high, management costs are low, and automation is prevalent. In the third phase virtualization is an integrated part of the Information Infrastructure.

Phases of File Virtualization
Phase 1: Migration Service
Phase 2: Resource Optimization
Phase 3: Integrated into the Information Infrastructure

Phase I: File Virtualization as a Migration Service
Networked Attached Storage (NAS) management is relatively easy with the first device. As storage devices proliferate storage management becomes more difficult. As environments grow, analysts estimate that storage management costs can be 4-7 times the acquisition costs. One of the most difficult and expensive storage management tasks is storage consolidation:

• Consolidation can be extremely disruptive. IT tries not to interrupt users or processes. They do their best to stick with limited down time windows, but it’s not always possible to avoid impacting customer service levels.
• Consolidation can be extremely time-consuming. Many consolidation projects take months of intense planning, with the project itself taking even longer.
• Consolidation can be extremely expensive. Costs are driven by 1) Productivity impacts for IT professionals involved in the coordinating with end-users, planning, and conducting consolidations. 2) Productivity impacts for end-users due to disruptions. 3) Business downtime due to application and end-user data access disruptions, and 4) Under utilized storage resources due to the length of time and uncertain timing of consolidation projects.

With file virtualization the major obstacle to storage management --end-user disruption--is eliminated and storage consolidations are dramatically different. Non-disruptive, transparent data migration preserves customer service levels, takes a fraction of the time as standard consolidations, saves significant administrator time, and avoids the cost of over-allocation.

For example, a major manufacturer completed a 160TB consolidation project ten times faster with file virtualization. They were able to move data to the new storage devices during peak hours with no end-user disruption.

Without file virtualization, after a storage consolidation project is complete the whole process repeats itself. Storage devices fill at disparate rates. Storage administrators have to negotiate with end-users to find convenient windows to take the storage devices offline and complete a migration project. In other words, without file virtualization organizations remain locked in a continuing cycle of consolidation after consolidation.

Not all file virtualization solutions provide effective migration services. Many solutions require that all clients mount the virtualization solution. This approach requires a disruptive reconfiguration of existing mount points and transfers the management headaches and coordination from the consolidation process, to the virtualization installation process. Only those virtualization solutions that can be installed transparently in an environment and leverage the existing network connections and standard switches can effectively function as a migration service. For example, EMC’s Rainfinity does not require reconfiguration of mount points and is invisible to clients and IP-based storage devices. Rainfinity performs transparent data migrations and communicates with the existing IP network switches to selectively filter the traffic without requiring reconfigurations or mount point changes.

After completing a successful and shortened storage consolidation project by using file virtualization, organizations are in a perfect position to move on to Phase II.

Phase II: Improving Utilization across Tiers
Now that storage is consolidated, real-time capacity balancing saves even more money and time. The use of file virtualization in Phase I has paved the way to eliminating the need for time-consuming sporadic storage consolidation efforts. In Phase II, data can be continually moved across storage devices in response to changes in the environment and the business. Virtualization balances capacity, performance and leverages tiered storage saving money, resources, and headaches by improving resource utilization and allowing intelligent allocation. For example, a leading ISP desperately needed a way to efficiently allocate capacity to its constantly changing data stores. With file virtualization, they were able to move data in real-time significantly increasing their filer farm’s utilization and performance, and lowering their storage capital expenditures.

During this second phase, file virtualization drives significant benefits on a day to day basis. File virtualization can automate the placement of individual files across tiers. Policy driven file archiving can also be used in conjunction with the movement of entire directories or volumes across tiers to free up large amounts of space on primary tiers. A leading semiconductor company was able to reduce CAPEX by 50% through moving content transparently across primary and secondary storage resources.

In this second phase, file virtualization delivers significant benefits:

• Uninterrupted access. Data movement occurs in the background with no user or application disruption. This allows storage administration tasks to run safely and continuously without waiting for batch processing windows.
• High performance. Data can be relocated based on performance requirements, ensuring that the performance is delivered where it’s needed most.
• High Utilization. With transparent capacity management, utilization rates of 90% are not atypical. This is due in large part because the major obstacle to performing capacity adjustments – end-user disruption has been completely eliminated.
• Lower Storage CAPEX. Capacity management can extend across tiers of storage to shift demand to lower cost storage devices..
• Smooth Integration. Global file virtualization eases into the existing network without requiring configuration changes to IP-based storage servers, file servers, clients, application servers or storage management tools and utilities. It supports all devices using industry standard protocols like NFS or CIFS.

A Phase II organization also has another benefit: an infrastructure that is ready to adopt next-generation automated data movement.

Phase 3 – Information Infrastructure
File virtualization is an enabling layer for automated storage strategies like ILM, Utility Computing and Grid Computing. In the final phase, file virtualization enables the non-disruptive movement of data as part of a larger information infrastructure. This infrastructure includes a Global Unified Namespace that provides an abstraction layer for end-users and applications that works in conjunction with a series of services including applications. In phase II, these services functioned more like separate applications initiated by administrators.

In Phase III virtualization services are governed by higher level orchestration. Policies extend across the environment and are leveraging an intelligent classification engine. Administrators define required service levels, data protection classes, and recovery objectives for various categories or classes of data. Phase III is the natural evolution of the use of file virtualization. Incremental steps proceeding from the first use of virtualization, such as a migration service, lead seamlessly to the third phase and further lower the TCO of the unstructured storage environment.

When evaluating a file virtualization solution you must consider how that solution will eventually fit into your larger infrastructure. In the third phase, a proprietary virtualization approach will fail to scale to meet the demands of an enterprise infrastructure. The requirements for file virtualization at this phase include the ability to drive virtualization externally through a standard API, the ability to integrate into the existing IP network without introducing a proprietary switch. ILM and service oriented architectures will include the intelligent placement of data and the storage infrastructure to respond flexibly to changing business conditions. This requires the file virtualization layer to seamlessly integrate into a large information infrastructure.

An important part of the information infrastructure is the global namespace. This provides a logical abstraction for the physical location of the file-based data and works similarly to DNS. End-users use a logical folder name, not the actual physical path to access data. A global namespace can be fully implemented in Phase II and lead seamlessly to Phase III. However, file virtualization solutions that work with multiple namespaces do not require the deployment of a global namespace. EMC Rainfinity provides a Global Unified Namespace that federates across multiple shared namespaces including login scripts. With Rainfinity organizations can benefit from file virtualization with the flexibility to deploy a global namespace during any phase of virtualization.

Conclusion
Today’s storage management projects should be both tactical and strategic: yielding valuable efficiencies and savings in the short-term, and preparing the IP network infrastructure to adopt the advanced technologies of the future.

File virtualization is not a point solution. It is an important enabling part of the infrastructure. The most important criteria in selecting a file virtualization is how does it evolve. How does a file virtualization solution efficiently link each phase to the next? How does it move an organization smoothly through consolidation, to tiered storage, capacity management and into next-generation storage computing?

By focusing on Phase I and the ability of a virtualization solution to serve as a migration service, you will reveal a great deal about a virtualization solution’s ability to meet your needs in Phase III. A file virtualization solution that is architected correctly drives benefits throughout each phase and delivers progressively higher cost savings.