Amazon Kinesis vs. Google Cloud Dataflow

Perfect for real-time data processing and streaming. Also, there's no need for any specific setup - you just start using it immediately and it easily integrates with the rest of AWS capabilities (like Redshift), although integration with Lambda could be better. You can make your overall analytics landscape way simpler with Kineses even if you have non-Amazon solutions like Tableau. It all integrates really well!

Incentivized

Based on my experience, streaming / real time / machine learning / AI type of processing and batch processing which needs less transformation are very well suited. Work load that needs complex transformation / multiple hops gets very complicated to implement. New feature like Dataflow SQL option will come in handy for sql heavy users.

• Integrating with other Amazon services
• Scaling requests
• Totally serverless platform
• Simple management

Incentivized

• Streaming, Real time work load
• Batch processing
• Auto scaling
• flexible pricing

• Improve integration with AWS Lambda
• Some duplicate records coming from the stream

Incentivized

• inbuild template options can be expanded
• more data connector options
• easy of use

The documentation was confusing and lacked examples. The streams suddenly stopped working with no explanation and there was no information in the logs. All these were more difficult when dealing with enhanced fan-out. In fact, we were about to abort the usage of Kinesis due to a misunderstanding with enhanced fan-out.

Incentivized

Kinesis is oriented to streaming in a scalable way large volumes of information in real-time. Glue is more an ETL so it is not well suited for real-time applications while Beanstalk is more a simple container platform. Lambda could do the job but it would require a lot of programming to accomplish the same as Kinesis. In fact, our solution employed the four elements for different tasks but using Kinesis as the message bus.

Incentivized

Google Cloud Dataproc Cloud Datafusion

• Caused us to need to re-engineer some basic re-try logic
• Caused us to drop some content without knowing it
• Made monitoring much more difficult
• We eventually switched back to SQS because Kinesis is not the same as a Queue system

Incentivized

• cost saving from managing our own data center for ETL servers
• consumption based pricing
• with auto scaling feature, we were able to expand components to support work load