Proactive VM Rebalancing

The rebalancing of workloads is essential to managing a virtual environment. Unlike tactical, reactive rebalancing performed by technology like VMware DRS, CiRBA enables proactive rebalancing, which is the strategic, long-term placement of resources based on detailed analysis of supply and demand. CiRBA enables organizations to proactively rebalance virtual infrastructure by combining placement analysis, capacity analysis, and forward-looking analysis to derive the optimal placements of workloads given recent and planned changes in an environment. This leaves the tactical, short-term balancing to happen only in exception situations.

Proactively Placing Workloads in Preparation for Changes in the Environment

Proactive Placements minimize tactical rebalancing, which may create volatility and disruption if it occurs in the middle of an operational cycle. CiRBA leverages historical utilization patterns and personalities, business rules and configuration constraints, combined with modeling of anticipated changes such as a business, event-driven spike in activity, to proactively recommend VM placements that best serve the coming utilization pattern. Proactive Placements are also required to balance resources when live migration of workloads between physical hosts is not supported or permitted. In these environments, rebalancing of workloads “on the fly” is not possible, and it is critical that placements be established ahead of time to ensure that application demands are met.

Placing Workloads to Accommodate Seasonal or Cyclical Changes

CiRBA’s time-windowed analysis focuses on a specific time range within historical data to help determine the optimal resourcing strategy for upcoming time windows. This is particularly useful when an organization has routine spikes or dips in utilization that need to be accommodated. In highly cyclical applications, such as financial services (where, for example, the purchase of 401Ks varies greatly from quarter to quarter), this can be used to right-size infrastructure and free up assets for other uses. This technique is not just applicable to virtual infrastructure, and can also be applied to physical infrastructure if the time windows are long enough to allow application components, such as databases, to be restored onto new infrastructure during defined outage windows.

Common forms of this analysis supported by CiRBA include:

• Cyclical right-sizing, including the analysis of daytime vs. overnight vs. weekend infrastructure requirements
• Seasonal right-sizing, such as per-month or per-quarter right-sizing of business services
• Time-dependent Cloud-instance-sizing, allowing Cloud applications to be right-sized to segments of the operational cycle, thus producing significant savings in instance fees

Enabling ITIL Compliant Management of Resources

Proactive management activities are instrumental in ensuring that virtual environments are successfully brought under ITIL service delivery and service management processes. Without proactive management practices, many virtual environments descend into a “fog” of activity, with no way to identify whether changes (such as VM movements) are occurring due to incidents, change activity, release activity, capacity management, etc. CiRBA provides proper forward-looking analysis, so that optimal workload placements are known in advance. CiRBA’s analysis enables placement decisions to be based on historical activity (plus risk buffers), so that they essentially reflect expected activity. If unplanned motioning occurs, it can then be interpreted as an ITIL incident, as it is caused by unexpected operational conditions. This separation of incidents from planned changes is key to gaining operational control over virtual environments that are hosting production workloads.