Sign up to be invited when a spot opens up
Bigger is not always better. Selecting an optimal compute instance size that accounts for the app's resource utilization behavior and load characteristics can improve performance. When the probable capacity of compute instance types are known, they can be selected based on deployment strategy and cost.
Resource utilization is dependent on load characteristics, which are a mix of different workloads with each class requiring different resources. Load characteristics change over time and can be cyclical. Matching the optimal instance types to changing load characteristics results in better performance.
Load characteristics change over time and may also be cyclical. Since cloud deployments can be elastic, rather than the cluster always using the same compute instance type, it can instead be configured to use fewer compute instances by using instance types that are optimal to specific load characteristics.
Understanding app resource behavior enables Operations to manage deployment and costs better. When you know the maximum capacity of different compute instance types it is easier to deploy strategies focused on minimizing footprint.
Cloud IaaS vendors provide dozens of compute instance types. Selecting an instance type has impact both on the performance of your apps as well as the cost. IntelliSense™ IaaS Workbench makes it easy to select the right instance type for your SLA thresholds - taking into consideration the app's resource utilization and load. The result? You save time and money.
Selecting the right compute instance normally requires Subject Matter Experts (SME) that have the expertize in analyzing, analyze app performance, resource utilization and IaaS vendor specific instance types. IntelliSense™ IaaS Workbench utilizes proprietary machine learning algorithms that embed domain expertise when providing recommendations.
Traditional techniques typically require SMEs, manual analysis and the use of custom build frameworks that require maintenance and upkeep. IntelliSense™ IaaS Workbench does away with many of these manual measures, making it faster and easier to select optimal instance types. The effects are less cost, optimization based on demand, and becoming part of all Continuous Delivery (CD) workflows.
Machine learning algorithms classify load mixes and detect changes in load characteristics. Understanding the load characteristics makes it easier to determine the optimal compute for when an app receives the load.
Machine learning algorithms understand the resource utilization and behavior of the app. They use ranking and forecasting techniques to select instance type and forecast the probable capacity within required SLO threshold.
Performance tuning needs to be specific to the load characteristics that the app is undergoing. But load characteristics are dynamic and periodic.
AskLytics Machine Learning detects when load profiles change and when to apply the appropriate JVM settings.
Apps receive different classes of work. Some could be short running while others are long-running. The system resources consumed are based on the mix of workloads that affect the JVM settings. AskLytics Machine.
Machine learning at your service
Recommendations based on the profile of the load and the workload classification.
Recommendations based on the compute instance choices available from the IaaS vendor.
Sign up to be invited when a spot opens up