Beyond Servers: Is Your Data Center Truly Virtualized?

5 comments

Nashua, NH-based Travis Vigil is executive director, Dell Storage.

Travis Vigil-DELLTRAVIS VIGIL
Dell

The concept of introducing virtualization into corporate data centers has moved well past the trial phase over the past few years into one of wide acceptance today. The latest virtualization technologies deliver on the promise of increased efficiency, reduced expenses and a smaller overall footprint.

But when it comes to a truly virtualized data center, server virtualization, which enables several applications to run independently on a single physical server, is only half of the solution.

Storage Demands Continue

CIOs today continue to struggle keeping up with the demands for more storage capacity as data continues to grow exponentially. That, in turn, has led them to add storage systems ad-hoc, responding to the needs of a specific deployment, a particular business function, or a geographic area. As such, many organizations have created a complex, inflexible storage infrastructure that has become both difficult and costly to manage.

To help address this complexity, leading storage providers have worked diligently and closely for years with server virtualization software providers – such as VMware, Citrix and Microsoft — to seamlessly link storage systems with highly virtualized server environments.

Through these efforts, and industry-wide projects such as VMware VAAI, storage arrays can now best optimize storage performance and flexibility when called upon by virtual servers.

Innovative Storage Systems Employ Virtualization

While this virtual server and storage array integration is extremely important for the success of a virtualized data center, innovative storage systems should be using virtualization technology themselves. In fact, when combined together, a truly virtualized server and storage environment paves the way for most effective cloud infrastructures.

This is where storage virtualization comes into play.

Think of storage virtualization as technology that masks the complexity of the storage environment by simplifying the cumbersome tasks that come with managing it – like performance tuning, load balancing and configuring RAID.

These tasks are still occurring, but they happen in the background of a virtualized storage environment, abstracted from not just the administrators but the servers, as well. In a nutshell, the storage capacities can be consolidated and presented as a single, highly-automated volume that looks like a normal disk to the virtualized servers.

It’s a simplified explanation for something that’s actually far more complex. But that simplified understanding is the heart of why storage virtualization systems can be as effective as they are now. They improve efficiency and productivity because they reduce the time, effort and, more importantly, the knowledge that’s needed to effectively manage storage.

Yes, it’s consolidated storage but presented as “pools” of storage in a system that allows loads to be balanced and volumes to be created and expanded quickly, in a manner that doesn’t impact or disrupt the users. In the process, RAID configuration, replication and snapshots are all integrated and transparent to the applications.

IT managers know that not all storage is created equal. Most enterprise SANs utilize a tiered environment that includes faster, more expensive storage with slower, more economical drives. Critical data is housed on the faster-performance storage while data that’s used infrequently gets stored on the lower-end drives. Some SANs require a manual process to move all this data around, while more modern, virtualized SANs handle the tiering automatically, which is why the feature is sometimes called automated tiered storage.

Real-life Example of Storage Needs

So, how does this play out in real-life scenarios? When the lawyers of Milwaukee-based law firm Quarles & Brady log-in to their workstations remotely, the power of storage virtualization is at work. The firm’s virtual desktops are accessing data from an automated storage array equipped with both higher-performance solid-state drives and more economical SAS drives.

Physically determining which types of data belong where and migrating it to the most cost-effective type of storage easily would become a cumbersome and labor-intensive – and unnecessarily expensive – process. The storage arrays used by Quarles & Brady handle the peak traffic with automated load balancing, shifting “hot data” to higher-performance drives during “boot storms,” such as the morning hours when employees all start logging into their virtual desktops.

The tiering and load-balancing are automated, the performance is uncompromised, and the firm has seen the time that it takes to provision a virtual desktop drop from 7 minutes to about 40 seconds.

As such, the firm has become an example for the increased efficiency, cost savings and, more importantly, the flexibility that can come with a true virtual data center, instead of just server virtualization.

Virtualization Reduces Demand for More Devices

This business and others have discovered that storage virtualization also reduces the need for more storage devices. By automating the management of existing storage, the data center not only takes advantage of all available space but also shifts data on the fly, without any interruptions. Shifting data to the most cost-effective tier of storage becomes an automated process and that not only reduces labor costs but also dynamically utilizes less-expensive storage products while freeing up valuable space on the higher-performance drives.

Thin provisioning also eliminates the need to pre-allocate drive capacity to a volume before that capacity is used, a long-time problem that sometimes involved a best-guess approach in determining the amount of storage that would be needed later. Managing storage needs the old way usually led to over-provisioning, a costly and wasteful approach.

The demands facing today’s IT environments are often large and complex. Data continues to grow at exponential rates. In many cases, there are compliance regulations and data governance policies that dictate which types of data must be retained and for how long – even if it’s not being accessed regularly. And the teams that are accessing the data have come to expect real-time information regardless of how or where they try to access it – whether from a remote office or a mobile device.

Transition in the Data Center

IT managers have come to enjoy the cost savings and increased efficiency that server virtualization has offered. Many are now discovering that a truly virtualized data center is more than just virtualized servers. This combination of virtualization from servers and storage also provides a natural transition to users looking to deploy various cloud and IT-as-a-Service models by setting the framework for a more easily shared and optimized environment.

Companies that are implementing or considering a server virtualization strategy need to know that those efforts are only part of the solution. Storage virtualization completes the puzzle by bringing the data centers of the past into the future, leaving behind the expenses and workloads that once went hand-in-hand with outdated approaches.

Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library.

Add Your Comments

  • (will not be published)

5 Comments

  1. me@home

    The title is a little bit misleading, but overall good discussion. Virtualizing servers is older technology but still very important. I think virtualizing storage is another 'cog' in the wheel of the true virtualized data center. The next step in the virtualized data center is the actual data center itself. Virtualized Data Center NIRVANA is having true Active/Active data centers and being able to load balance between the two so that users and applications never know the difference nor do they care about where their apps reside and which physical location they are truly 'connected' too! Network virtualization and global/geo load balancing provides the final 'cog' in a truly overall virtualized data center. Unfortuantely there is no single company that can provide this overall complete 'wheel' but a couple of you guys are getting close!

  2. I completely agree with many of your points about making data centers more cost-effective, compliant and easier to manage by virtualizing both applications and the growing amount of data that accompanies them. What are your go-to solutions for managing uptime and availability for the more critical applications and data? Disclosure, I'm in the business and asking to see what is top of mind to you - hope you don't mind.