Your data center is at likely at 12-18% utilization according to numerous studies done by McKinsey, IBM, and others. For VDI, it is simple to understand why: (a) you have to provision for peak usage, likely once a day for a few hours (b) all your infrastructure is sitting idle most nights and weekends – an average work week is 40 hours out of 168 hours.
DaaS is a fundamentally different business model. We are all aware of the potential optimizations – start/stop, pause-resume, allocate/de-allocate – to manage cost. But without a feedback loop, it’s like driving a car in the dark with no lights.
We have built an optimized workflow that enable you to closely match your demand curve with cost in the cloud ensuring that you pay for what a user uses, never more, and sometimes less!
- Just-in-time Provisioning
- Real-time Cost Alerts
- Review Excess Provisioning
- AI-Generated Usage Forecast
- AI-Generated Just Enough Provisioning
Let’s show you how the workflow in action.
Just-in-time Provisioning
First, you can control the provisioning of desktops for just-in-time provisioning for both persistent and non-persistent scenarios:
- idle timeouts
- increase capacity and keep in paused mode or ready mode
- decrease capacity and keep in paused mode or de-allocate
Real-time Cost Alerts
Sometimes the IaaS provider does not co-operate and operations don’t succeed. This can have negative effects in multiple ways – your cost meter keeps going and future users may not be able to access that machine. We show you problems in real-time so you can resolve them faster.
Review Excess Provisioning
One critical component is to understand the balance between capacity provisioning and user connection times. You can optimize user connection times to <10 seconds, but you may have over-provisioned. See chart below. The red line shows the excess capacity, while the user connection times are fantastic, so there is room to optimize provisioning.
AI-Generated Usage Forecast
We help you understand future usage based on historical data. The blue line shows historical usage for the pool, the orange line is our prior forecast, and the green line is the future forecast generated by AI.
AI-Generated Just Enough Provisioning
Based on the forecast, we can recommend a just-enough warm up policy for the pool that will match provisioning to demand.
There are some obvious next steps to complete this workflow. Stay tuned.
