Apache Flink vs Google Cloud Dataflow comparison

"With Flink, it provides out-of-the-box checkpointing and state management. It helps us in that way. When Storm used to restart, sometimes we would lose messages. With Flink, it provides guaranteed message processing, which helped us. It also helped us with maintenance or restarts."

"Apache Flink is meant for low latency applications. You take one event opposite if you want to maintain a certain state. When another event comes and you want to associate those events together, in-memory state management was a key feature for us."

"The setup was not too difficult."

"The product helps us to create both simple and complex data processing tasks. Over time, it has facilitated integration and navigation across multiple data sources tailored to each client's needs. We use Apache Flink to control our clients' installations."

"Apache Flink's best feature is its data streaming tool."

"It is user-friendly and the reporting is good."

"Easy to deploy and manage."

"This is truly a real-time solution."

"Google Cloud Dataflow is useful for streaming and data pipelines."

"The solution allows us to program in any language we desire."

"The most valuable features of Google Cloud Dataflow are scalability and connectivity."

"It is a scalable solution."

"The support team is good and it's easy to use."

"Google's support team is good at resolving issues, especially with large data."

"I don't need a server running all the time while using the tool. It is also easy to setup. The product offers a pay-as-you-go service."

"The best feature of Google Cloud Dataflow is its practical connectedness."

"In terms of stability with Flink, it is something that you have to deal with every time. Stability is the number one problem that we have seen with Flink, and it really depends on the kind of problem that you're trying to solve."

"PyFlink is not as fully featured as Python itself, so there are some limitations to what you can do with it."

"The TimeWindow feature is a bit tricky. The timing of the content and the windowing is a bit changed in 1.11. They have introduced watermarks. A watermark is basically associating every data with a timestamp. The timestamp could be anything, and we can provide the timestamp. So, whenever I receive a tweet, I can actually assign a timestamp, like what time did I get that tweet. The watermark helps us to uniquely identify the data. Watermarks are tricky if you use multiple events in the pipeline. For example, you have three resources from different locations, and you want to combine all those inputs and also perform some kind of logic. When you have more than one input screen and you want to collect all the information together, you have to apply TimeWindow all. That means that all the events from the upstream or from the up sources should be in that TimeWindow, and they were coming back. Internally, it is a batch of events that may be getting collected every five minutes or whatever timing is given. Sometimes, the use case for TimeWindow is a bit tricky. It depends on the application as well as on how people have given this TimeWindow. This kind of documentation is not updated. Even the test case documentation is a bit wrong. It doesn't work. Flink has updated the version of Apache Flink, but they have not updated the testing documentation. Therefore, I have to manually understand it. We have also been exploring failure handling. I was looking into changelogs for which they have posted the future plans and what are they going to deliver. We have two concerns regarding this, which have been noted down. I hope in the future that they will provide this functionality. Integration of Apache Flink with other metric services or failure handling data tools needs some kind of update or its in-depth knowledge is required in the documentation. We have a use case where we want to actually analyze or get analytics about how much data we process and how many failures we have. For that, we need to use Tomcat, which is an analytics tool for implementing counters. We can manage reports in the analyzer. This kind of integration is pretty much straightforward. They say that people must be well familiar with all the things before using this type of integration. They have given this complete file, which you can update, but it took some time. There is a learning curve with it, which consumed a lot of time. It is evolving to a newer version, but the documentation is not demonstrating that update. The documentation is not well incorporated. Hopefully, these things will get resolved now that they are implementing it. Failure is another area where it is a bit rigid or not that flexible. We never use this for scaling because complexity is very high in case of a failure. Processing and providing the scaled data back to Apache Flink is a bit challenging. They have this concept of offsetting, which could be simplified."

"One way to improve Flink would be to enhance integration between different ecosystems. For example, there could be more integration with other big data vendors and platforms similar in scope to how Apache Flink works with Cloudera. Apache Flink is a part of the same ecosystem as Cloudera, and for batch processing it's actually very useful but for real-time processing there could be more development with regards to the big data capabilities amongst the various ecosystems out there."

"Amazon's CloudFormation templates don't allow for direct deployment in the private subnet."

"In a future release, they could improve on making the error descriptions more clear."

"Apache Flink's documentation should be available in more languages."

"There is a learning curve. It takes time to learn."

"When I deploy the product in local errors, a lot of errors pop up which are not always caught. The solution's error logging is bad. It can take a lot of time to debug the errors. It needs to have better logs."

"The technical support has slight room for improvement."

"There are certain challenges regarding the Google Cloud Composer which can be improved."

"I would like to see improvements in consistency and flexibility for schema design for NoSQL data stored in wide columns."

"Google Cloud Dataflow should include a little cost optimization."

"I would like Google Cloud Dataflow to be integrated with IT data flow and other related services to make it easier to use as it is a complex tool."

"The deployment time could also be reduced."

"Occasionally, dealing with a huge volume of data causes failure due to array size."

"It's an open source."

"The solution is open-source, which is free."

"This is an open-source platform that can be used free of charge."

"Apache Flink is open source so we pay no licensing for the use of the software."

"It's an open-source solution."

"The solution is cost-effective."

"The solution is not very expensive."

"The price of the solution depends on many factors, such as how they pay for tools in the company and its size."

"Google Cloud Dataflow is a cheap solution."

"Google Cloud is slightly cheaper than AWS."

"On a scale from one to ten, where one is cheap, and ten is expensive, I rate the solution's pricing a seven to eight out of ten."

"The tool is cheap."

"On a scale from one to ten, where one is cheap, and ten is expensive, I rate Google Cloud Dataflow's pricing a four out of ten."