According to Wikipedia [1], "the serverless compute is a cloud computing execution mode which the cloud provider dynamically manages the allocation of machine resources. ". Serverless compute has been around for some time. Amazon pioneered in the serverless with AWS Lambda. The current popular serverless databases includes Amazon DynamoDB, Microsoft Azure Data Lake, Microsoft Azure CosmosDB, Google Firebase and FaunaDB. Amazon Aurora Serverless was announced in AWS re:Invent 2017 [3], and it is now ready for preview. This blog briefly looks at the new serverless architecture.
Serverless introduces a new software architecture. Serverless database means that database storage, scaling, high availability are now managed by the cloud provider. Users can just use WebService APIs for database services. The new architecture has the following benefits [2]:
- Less operational cost. In the long run, a fully serverless solution means zero system administration.
- Quick time to market
- Vendor lock-in
- Lacking optimizations to customized workload
- Multitenancy on robustness, security and performance
Resources
- Serverless Computing, Wikipedia
- Mike Roberts, Serverless Architectures, 08/06/2016
- AWS re:Invent 2017 Keynote: Andy Jassy 11/30/2017