4 ms·
What's interesting about this is the conscious decision to throw money at the government IT problem. Meaning, services like AWS Lambda get very, very expensive
by solatic 5y ago
What's interesting about this is the conscious decision to throw money at the government IT problem. Meaning, services like AWS Lambda get very, very expensive at scale. Nationwide government-scale is probably pretty high scale, and I shudder to think what the AWS bill would be.
Yet, most government IT projects run very high costs to begin with anyway. Indeed, for all we know, they probably saved money, even with the high costs of running this at scale on Lambda. And it'll be worth it if the service is actually functional.
Just interesting to think about the cost of introducing stability as a standard before eventually the pendulum will swing back and get people to think about cost optimization.
- dclowd9901 5y agoGiven what they’d probably spend to a contractor like General Dynamics to completely flub this, I’m glad for them to foot a massive AWS bill instead.
- koblas 5y agoIf you go with the authors original description that the lambda function for handling orders is small and fast. Probably a bunch of server side validation before stuffing in DynamoDB or maybe into a queue (Message Queues / Kafka). We can assume that the program is small and fast. If you saw 300,000,000 requests in the month on a 256MB lambda function and 500ms of runtime. It would only cost $678 for the order capture. Maybe my math is wrong here, but not huge in my book.