Think of a mission-critical application outage because a system unexpectedly ran out capacity. With so many infrastructure elements supporting so many important applications, you need a way to clearly see all your systems’ capacity and where it’s headed.
AIOps software is a technical solution for this challenge. We’ve previously discussed how AIOps software lets you see your infrastructure’s overall health and the impact of performance anomalies, so you can prioritize your actions. We’ll now look into how Dell’s AIOps software (CloudIQ) predicts the future.
Seeing the Future with Capacity Prediction
A mission-critical application became in-operational because storage reached full capacity. You need to be aware of when a system will reach capacity by learning usage history and patterns, predicting what is likely to happen in the future and then providing actionable insights to act before you run out of capacity.
Technical Solution – Capacity Prediction
Let’s start with storage arrays. CloudIQ displays a summary of the data storage objects, such as LUNs, pools, volumes, file systems and more, approaching full capacity, an estimated time range remaining until each file system or pool is full and shows a forecast capacity chart. If full capacity is imminent, such as due to a file system or pool running out of space within 24 hours, CloudIQ displays the estimated time and a list of top five contributors. This allows you to make quick decisions about moving workloads to other storage systems, reclaim unused capacity or other action.
Three main factors that distinguish CloudIQ capacity prediction are:
- Response time of capacity prediction is 10X faster in CloudIQ compared to other solutions.¹ Overall, the method and process are patented to achieve such a response time. Algorithms scale the time series data in a unique way to use multiprocessing and multi-threading operations.
- Capacity prediction in CloudIQ considers multi-level storage objects while forecasting for the parent storage object. This significantly increases the quality of capacity predictions. Intelligent algorithms look at both higher-level as well as lower-level storage objects and aggregate the capacity forecast to achieve trustworthy predictions.
- The footprint of capacity prediction in CloudIQ is much larger compared to others. The scale of capacity prediction is enormous in CloudIQ. The prediction window enables both strategic, long-term capacity planning, up to a year, as well as urgent, short-term capacity planning, within 24 hours.
Beyond Storage Arrays – Capacity Prediction for Other Types of Systems
Today, CloudIQ also monitors and predicts full capacity for hyperconverged infrastructure systems. Capacity full protection for servers (CPU and memory), IP and SAN switches (ports) and data protection for back-up on premises and tiered in public clouds, are on the horizon for full infrastructure. Another recent innovation is…