Scalable, low cost computing
April 19, 2018
Scalable, low cost computing
Score 10 out of 10
Vetted Review
Verified User
Overall Satisfaction with AWS Lambda
We currently employ Lambda to do a number of event-driven tasks without our backend infrastructure. This ranges from API services to on-the-fly image manipulation. The reason for utilising Lambda in this fashion comes from a few distinct advantages: ease of integration with other AWS services (i.e. we can trigger it extremely easily), cost and scalability. A number of our services have either very low workloads, and thus it would be wasteful to run services 24/7, or very unpredictable demands - both of which Lambda help us with massively.
- Pay for only what you use. Because Lambda is billed by the 100ms of execution time, you can run low volume services extremely cheaply.
- Scalability. Lambda will spin up as many concurrent executions on demand as required to fulfil the triggers (up until a soft limit at least). This means for unpredictable workloads we get reliable execution with minimal costs.
- Ease of integration with other AWS services - Lambda can be plugged into just about everything and anything within the AWS ecosystem and also can be trigger via APIs from external systems making it very easy to integrate with.
- Language support is OK, but could be improved. In particular it would be nice to see native support for PHP, given its prevalence, and possibly Ruby.
- It would be great if there was a way of doing scheduling with a better granularity than 1 minute. For example, if you want to poll something every 15 seconds, it is not straight forward to do this using Lambda and the associated triggers as things stand.
- We've been able to replace some existing long-running services with Lambda-based services and turn the old ones off. Due to the fact that we now only pay for usage rather than having to keep services running 24/7 and scaling them ourselves, we've dramatically reduced costs (around 90%) as well as complexity and maintenance overheads.