Leading Insurer Minimizes Volatility & DRS Moves with Predictive Analytics
Summary
- Organization Size
- 116,000 employees
- Private Cloud Size
- 35,000 servers
- VMware VDI environment size
- 15,000
Customer Challenge
A global insurer was experiencing lots of VM placement volatility within their VDI environment due to frequent issues with resource contention. They were relying on VMware DRS®, a reactive load balancer, to place VMs and were experiencing a very high number of VM moves during the business day. DRS looks at short windows of time, typically 5 minutes of data) and moves VMs when it detects contention for CPU and memory resources. The customer wanted to proactively rebalance the VDI environment considering storage (Disk) and network IO in addition to CPU and memory to reduce the number of intra-day moves and avoid host level imbalances. DRS could not do this.
Densify Solution
Densify was being used in the customer’s VMware environment, but not on their VDI servers. The team consulted with their Densification Advisor™ and decided to apply the analytics to the VDI environment. The predictive analytics examined the historical workload patterns to model future workload behavior and get an accurate view of what the environment will look like over the next 30 days. Based on this analysis, Densify was able to optimally dove-tail workload patterns and strategically VM placements in order to reduce future resource contention and intra-day VM moves.
The Results
Densify’s placement plan worked and reduced reactive DRS moves by over 76% during business hours. The insurer also leveraged Densify’s integration with VMware® DRS to program Densify’s placement policies into DRS to guide its actions.
Combining Densify’s proactive automated placements with DRS’s reactive placements provide the insurer with lower resource contention and volatility, an optimized stacking ratio and a safety net for realtime issues that arise, thus improving the overall performance and minimizing risks.