Slideshow Top 5 Storage Pitfalls in Virtual Environments

  • August 22 2016, 6:30am EDT
6 Images Total

Top 5 Storage Pitfalls in Virtual Environments

The rapid growth in the amount of data that organizations need to store is putting a strain on many. Mark Lewis, chief executive officer at Formation Data Systems, discusses five “pitfalls” of storage in virtual environments.

Stranded Capacity in VMware Hosts

“A challenge faced by many organizations that deploy ESX hosts attached to networked storage configuration (SAN/NAS) is that the storage capacity within the host is completely unused,” Lewis explains. “Standard server configurations typically include as much as a terabyte of capacity, which for very large VMware environments, could potentially represent Petabytes of wasted capacity and capital expenses. To address this requires a software defined infrastructure that can intelligently discover and pool this stranded capacity across all ESX hypervisors and with the proper storage services applied can be provisioned to hosts for additional use cases.”

Content Continues Below

Storage Related Performance Issues

“In a mixed workload environment, delivering consistent performance across all VMs can be very challenging. Due to this, supporting high performance, low latency applications typically requires dedicating storage resources to applications in order to deliver the specified performance,” Lewis says. “This results in poor efficiency, lower utilization and performance degradation, especially during peak load periods. To address this, storage solutions that employ quality of service can prioritize workloads in the virtualized environment, delivering consistent throughout and I/O performance, regardless of the load on the storage platform.”

VM Snapshot/Backup performance issues

“When performing VM backup operations, many admins choose to utilize VM snapshots in order to create a mountable image for a backup,” Lewis notes. “The challenge with this approach is that the VM snapshot process consumes resources from the ESX host, resulting in VM performance issues and long backup windows and complex recovery, increasing the risk of data loss. When snapshots are created by the storage system, these performance issues can be completely avoided. Additionally, when the storage system utilizes journaled data protection architecture, the risk of data loss is greatly minimized.”

Storage agility for DevOps

“The inability to migrate data from test/dev to prod Scaling storage across environments is a particular challenge for VMware admins,” Lewis says. “The total solution cost and complexity of the solution often prevents advances in test and development systems and devops. What is needed is the ability to instantly create virtual clones from production that can be migrated to devops and test and developments environments. This also enables rapid merging of new releases into production helping to accelerate new products and services.”

Content Continues Below

Challenges with Storage vMotion

“vMotion was originally developed to address the shortcomings of traditional storage arrays especially when it comes to moving data between different storage systems. Modern storage systems now have automated movement of the most active data to the higher performance flash tier, but even that may fall short,” Lewis explains. “An advanced software defined storage system would accomplish this data movement dynamically place data and grow volumes in real time based on pre-defined policies. As the I/O utilization of the supported workloads shift, the software-defined system should make adjustments to data positioning so that the heaviest demand workloads continue to deliver the performance they expect.”