Many data center managers find management of physical server storage resources to be a daunting challenge. With physical storage management already a challenge, the thought of managing storage in virtual environments, to some, is nearly as scary as any of the Saw movies.
Of course, data center managers can't simply hide in a closet. Instead, they have to face the beast head
- Capacity planning
- Asset management
- Fault management
- Configuration management
- Accounting management
- Security management
Server virtualization relies heavily on shared storage resources as nearly all data centers running production systems in virtualized environments leverage store VM files on storage area networks (SANs). Direct attached storage (DAS) is primarily out of the question for VM storage because shared storage is needed for use with high availability solutions such as VMware HA, Virtual Iron LiveRecovery, or a SWsoft Virtuozzo cluster. In production environments, high availability and its dependent shared storage is required because it mitigates the impact of a physical server failure. Without a high availability solution, for example, failure of a single physical server could result in ten or more virtual systems being unavailable. Many servers that previously relied on DAS will likely be virtualized and make use of shared storage on a SAN. In turn, data center management now includes need to be more keenly aware of SAN storage in order to ensure that adequate storage resources remain available as new VMs are added to the IT infrastructure.
In the physical world, assets are often logged and inventoried using barcodes and system serial numbers for tracking. In the virtual world, asset management isn't so easy. A physical hard drive may have its contents cloned, but the hard drive itself remains unique. Virtual hard disks can be easily duplicated without distinction, making tracking storage assets in the virtual realm much more challenging than with physical systems. Instead of visible components requiring tracking and inventory control, the files that make up virtual systems now must be tracked and inventoried. Vendors such as Dunes Technologies (acquired by VMware) and embotics are working to expand asset management intelligence for virtual environments.
While physical storage faults should still be managed by existing storage management productions, virtual storage failures are a new concern. Virtual storage may be deployed as virtual hard disks (such as .vhd or .vmdk files) or as raw device mappings (RDMs) where a VM's virtual hard disk is directly connected to a SAN LUN or local storage device. Virtual storage failures result from virtual hard disk corruption as well as SAN failures or configuration errors.
Numerous vendors including Veeam, Configuresoft, VMware, HP, IBM, Symantec, and CA are expanding their solutions in the configuration management space to further assist in tracking virtual machine configuration changes across an enterprise. Configuration changes to virtual storage must be tracked alongside configuration changes to SAN storage. Technologies such as thin provisioning allow organizations to get the most out of their storage investments, but ample tracking and auditing is necessary to ensure that storage remains available as VM storage requirements increase.
Accounting and chargeback also presents new challenges for virtual resources. While many organizations have well established chargeback policies for physical resources, the same cannot be said for virtual resource management. Vendors such as IBM, PlateSpin, Vizioncore, and VKernel now offer virtualization chargeback tools, with several additional vendors expected to follow suit. In virtualized environments, applications no longer run on dedicated hardware resources, and servers are no longer deployed on dedicated hardware elements. Storage resources are often shared across departments within an organization. Heavy reliance on shared physical resources is requiring organizations to rethink their existing chargeback policies, and several vendors are now looking to provide products that help organizations manage chargeback across virtual environments.
The challenges facing storage security have increased significantly with the rise of server virtualization. Storage security management is no longer about securing physical storage assets. Virtual hard disks, for example, could easily be copied to removable media such as a USB thumb drive and removed from a building without detection. Removal of physical storage devices such as hard disks or backup tapes has always been more difficult than theft of a virtual hard disk that can be duplicated to removable media. Organizations must be vigilant in tracking and auditing access to virtual hard disks. In addition, it's imperative that organizations restrict IT staff rights to virtual resources to ensure that staff members are only provided with virtual resource assess that is essential to their work responsibilities.
Business value is what matters most
Numerous organizations follow the Information Technology Information Library (ITIL) framework as their blueprint for IT operations management. ITIL v3, released in June 2007, moved ITIL's focus toward leveraging IT to increase business value. If you missed Cisco CEO John Chambers VMworld 2007 keynote, it is definitely worth a few minutes of your time. At this year's VMworld conference, Chambers discussed his vision of IT's next evolutionary step. Sure it included virtualization, but it also included management automation as a means to leverage IT to further increase business productivity. For example, if front end web server workload suddenly scales beyond server capacity, additional resources should be able to be dynamically allocated to meet the new workload demands. So as business needs change at the drop of a hat, technology is able to immediately keep pace. Of course, Cisco is now trying to get the word out about its VFrame data center automation product, but they're not alone. Several other vendors, including Scalent Systems, BMC, and Opsware offer data center automation tools as well. Many believe (myself included), that data center management is moving quickly toward service automation. As workload changes, the data center as a whole must add or remove resources (server, storage, network) as needed. Leveraging virtualization is the first step toward complete data center automation, as virtualization provides the system portability and hardware independence needed to make automation a reality.
If it ain't broke, fix it
Virtualization offers far too many benefits (consolidation, improved availability, and system portability, to name a few) to be ignored. Even if IT operations are running well in a purely physical world, virtualization's benefits are far too valuable to be overlooked. Organizations that add agility through virtualization will see a competitive advantage over those that don't. While I've outlined enough management concerns to make you want to run for some Rolaids, keep in mind that solutions exist for each virtualization management challenge. Planning for and taking on each challenge one-at-a-time is a surefire way to realize the benefits of the virtual infrastructure, improve IT agility, and retain your sanity.
About the author:Chris Wolf is a Microsoft MVP for Windows Server – File System/Storage and is a MCSE, MCT, and CCNA. He's a Senior Analyst for Burton Group who specializes in the areas of virtualization solutions, high availability, enterprise storage, and network infrastructure management. Virtualization: From the Desktop to the Enterprise (Apress), Troubleshooting Microsoft Technologies (Addison Wesley) and a contributor to the Windows Server 2003 Deployment Kit (Microsoft Press).
Reach him at firstname.lastname@example.org
This was first published in October 2007