Optimization goals are fueling a trend across vertical markets among IT managers who are remapping their data centers. The concept eliminates islands of technology, such as groups of servers, storage systems and network resources formerly dedicated to a single department or enterprise operation.
Instead, these managers are taking a more holistic view of IT and creating pools of resources designed to boost utilization rates and enable on-the-fly resource provisioning.
Hardware consolidation, extensive virtualization and balanced mixes of on-premises and cloud implementations are important elements in data center optimization efforts; however, these ambitious strategies often go beyond the technologies themselves. IT managers also can more effectively use available physical space and fine-tune power and cooling systems for greater efficiencies and less cost.
IT managers start with virtualization. The widespread server virtualization efforts of the past few years have helped many enterprises reduce physical hardware requirements, increase the utilization rates of remaining devices and alleviate skyrocketing power demands.
Now organizations are looking beyond consolidation to the next big wave of virtualization, says Greg Schulz, founder of the Server and StorageIO consulting group. “The goal is to enable higher productivity.”
Schulz cites the hypothetical example of a server that supports a mission-critical database application.
“The conventional thinking in the consolidation game is that this resource can’t be virtualized,” he says. Why? Because a database’s performance demands require dedicated computing power. But that may be required only during regular office hours.
“Instead of having that server sit idle at night, an organization can use virtualization to move other processes, such as reporting, analytics or backups to the server after hours. Then during the daytime, the critical application goes back to the server and gets all the resources it needs,” Schulz explains. “The result is that you are using the server more effectively, and you are increasing overall productivity.”
In a similar way, organizations can use virtualization to dynamically shift loads across various servers to perform upgrades without incurring downtime, and enhance operations continuity and disaster recovery efforts, he adds.
Virtualization provides similar load shifting and uptime benefits for storage systems. In addition, optimization-minded IT managers are looking to unified storage technologies to provide added value.
Unified storage systems support both block and file storage and the relevant protocols in the same unit, often along with redundant arrays of independent disks (RAID), load balancing and other management capabilities. As a result, IT shops can take another step toward optimization by reducing the range of technology they have to purchase and support.
Networking technology is following suit with virtual, software-based switches that create opportunities for hardware consolidation and dynamic resource allocations.
Virtualization in all its forms isn’t the complete answer for data center optimization. IT departments also need tools to ease the management of resource pools. A number of options are becoming available. For example, Symantec’s ApplicationHA provides application monitoring to help maintain high availability and reduce service disruptions for applications in virtualized environments.
“These technologies are helping organizations take the risk out of virtualizing more of their critical applications,” says Mike Reynolds, product marketing manager for Symantec’s Storage and Availability Management group.
IT managers can use ApplicationHA to monitor applications and receive immediate alerts in consoles such as VMware vCenter if a program crashes for any reason. The solution will restart a failed application if possible or it works with native high-availability tools to ensure that it fails over to a healthy virtual machine.
To help optimize storage systems and facilitate rapid provisioning of storage resources, Symantec offers VirtualStore. The technology creates a standard image of Virtual Machine Disk (VMDK) files that can be re-used when new virtual machines are created. Because provisioning virtual machines is fast and easy, administrators can proactively or reactively create them based on business requirements; however, creating a VMDK for each virtual machine can quickly eat up storage space, Reynolds says. The golden image created by VirtualStore can help organizations reduce VMDK storage requirements by up to 90 percent, according to Reynolds.