- Print
- DarkLight
- PDF
Internal Nodes Usage is based predominantly on the amount of time each execution takes.
Execution time is measured in GB-seconds where GB is the amount of memory. The amount of memory for an Internal Node may change, as EchoStream continually monitors each Node for execution efficiency and may dynamically increase or reduce the memory in order to optimize cost and execution time.
Memory (MB) | Node Processing Usage per 1ms |
---|---|
512 | 0.000000083 |
1024 | 0.000000167 |
1536 | 0.000000250 |
2048 | 0.000000333 |
3072 | 0.000000500 |
4096 | 0.000000667 |
Each Internal Node starts life with 1536MB of memory. Each Internal Node also has an Alarm created to monitor it 24/7 for errors during Node execution. Each Alarm generates Monitoring Usage of 1.00 per month, prorated for Internal Nodes by how long during the month the Internal Node exists.
Example
Let's assume that your Tenant contains 5 Internal Nodes; a Processor Node (P1) that receives the initial messages and preprocesses them, a Bitmap Router Node (BR1) that determines which route(s) the messages take, and three post-routing Processor Nodes (P2, P3, P4) that receive the messages from the Bitmap Router Node.
This Tenant receives 1,000,000 messages to process each month. Each message is routed to exactly one ProcessorNode. Each Node in the system can process 10 messages at a time (resulting in 100,000 executions per Node layer), with the following execution times per Node:
Node | Per execution time | Node Processing Usage per Execution | Executions | Total Usage |
---|---|---|---|---|
P1 | 150ms | 0.0000375 | 100,000 | 3.75 |
BR1 | 50ms | 0.0000125 | 100,000 | 1.25 |
P2 | 500ms | 0.000125 | 33,333 | 4.166625 |
P3 | 100ms | 0.000025 | 56,667 | 1.416675 |
P4 | 15,000ms | 0.00375 | 10,000 | 37.5 |
48.0833 |
( 48.0833 (Node Processing Usage) + 5.00 (Monitoring Usage) ) * $0.50(Cost per Usage) = $26.54165/month, or $0.00002654165/message