Virtualization brings the advantage of being able to run a multitude of workloads on the same physical infrastructure. In theory, virtualization promises IT professionals better infrastructure utilization and simplified management. The reality that many are finding, however, is in fact to the contrary.
Virtual machines (VMs) are effectively an abstracted aggregation of data files and metadata. The challenge lies in the fact that storage devices typically read aggregated storage volumes or logical units, which can be comprised of many VMs. Storage devices do not have visibility into the composition of individual VMs, or into what those VMs’ requirements are in areas such as performance. As a result, the storage device cannot account for or respond to VM-specific requirements. All VMs are treated the same, and the storage device makes a best effort attempt at reacting on the level of volumes and logical units.
What ends up happening is that some VMs squeeze out others when it comes to capacity and performance, because those resources have not been allocated in an optimal manner. This is called the “noisy neighbor” problem. Some infrastructure resources are maxed out, while others are underutilized. This flies in the face of a core objective of virtualization, which is to help organizations to get the most out of their capex investments in on-premises infrastructure.
One of the major pain points associated with serving storage to VMs lies in the need to manually provision storage logical unit number (LUN) groupings or volumes on an ongoing basis. The storage manager is left making their best attempt at allocating storage in a way that provides enough I/O performance to meet application service level agreements (SLAs), while at the same time avoiding saturating storage queues. This adds substantial management complexity and leaves a degree of storage capacity unused.
The LUN-level visibility furthermore adds challenges around backup and disaster recovery. Snapshots are taken of the entire volume, and the VM must be extracted from that volume. The same is true when it comes to replication and recovery.
Arguably even more strategically, application performance suffers, and some applications run out of capacity while business operations and competitiveness are quickly impacted. Again, human beings need to intervene, in this case by attempting to identify and resolve the bottleneck issue.
Data Direct Networks (DDN) recently joined Storage Switzerland for a Lightboard Video discussion regarding the value of a VM-specific storage management platform. For more information, access the video on demand.
Senior Analyst, Krista Macomber produces analyst commentary and contributes to a range of client deliverables including white papers, webinars and videos for Storage Switzerland. She has a decade of experience covering all things storage, data center and cloud infrastructure, including: technology and vendor portfolio developments; customer buying behavior trends; and vendor ecosystems, go-to-market positioning, and business models. Her previous experience includes leading the IT infrastructure practice of analyst firm Technology Business Research, and leading market intelligence initiatives for media company TechTarget.
StorageSwiss.com - The Home of Storage Switzerland 