Skip to main content
Version: 3

Deployment Infrastructure - Setup Sizing and Scaling

Setup Sizing Estimator​

Initial Setup

  • Adeptia utilizes a proprietary sizing estimator calculator tool that is used for “capacity planning” and to recommend installation size based on specific customer usage scenarios
  • This recommends size of servers/VMs and number of cluster nodes for initial setup

Post-setup Analysis - based on a working and stable solution implementation

  • Perform baseline benchmarking for few selected combinations of file types and sizes for a given memory and CPU configuration on a single node. For best results, use a number of concurrent files (say 2 or 3) that will have the CPU and RAM utilization to safe limits not exceeding the current RAM settings and having enough margin.
  • Profile CPU and RAM usage and measure peak CPU and RAM utilization in percentage (%).
  • Now the Sizing Estimator will provide an estimated number of nodes for a given size of single node e.g. 16Gb or 32Gb for the peak loads in consideration. Actually, slightly lesser configuration may also be able to meet performance demands

Scaling Performance​

Performance capacity can be scaled both horizontally and vertically

Horizontal Scaling

  • To scale data volume processing capacity and system throughput, increase # of Kernel nodes in cluster
  • To scale APIs/WS capacity and system throughput, increase # of Kernel nodes & WebRunner in cluster
  • To scale user response times and concurrent user capacity, increase # of Connect nodes

Vertical Scaling

  • To handle large jobs and huge data files in a scalable manner, select larger servers for Kernel such as with 8-16 Cores and 32GB+ of RAM
  • To handle both large jobs and high expected data and transaction volumes, separate installations may be considered for different departments or processes