Happy Anniversary!!!🎉 FIFO is always a pain. In order to solve it I think we should have a two stsge process. One is to collect the messages in an out of order dtyle quickly. At the second pahse when we are able to understand that we have a number of messages that makes sense, then we order them and lrocrss them in a batch mode. Any other ideas?
Thank you for the video. My question is: How can we make sure about the order of events in the consumer side in case the consumers will consume same topic from multiple clusters?
Happy your 6th anniversary. Which framework / tools do you recommend to implement multi broker pattern? Because I googled it briefly and couldnt find much.
Hi Mark, Fantastic content as always! It's impressive to see the depth you covered in Lesson 72 on the Multi Broker Pattern (ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-iXvC-ZyNZ8k.htmlsi=Cl6XMhEO6ITmopJl). I'm thoroughly enjoying the insights. I'd love to suggest a few topics for future discussions: 1. BFF (Backend for Frontend) 2. BFF vs ESB (Enterprise Service Bus) 3. BFF vs Orchestration Pattern 4. Feature Toggle, inspired by Martin Fowler's insights 5. API Gateway, Authentication & Authorization, Open ID 6. Serverless Architecture 7. Frontend Architecture Technologies, with a focus on React... Looking forward to your expert insights on these subjects! Keep up the excellent work.
Thanks for the suggestions! I'm always looking for more content and will consider these for future lessons. Very astute of you to recognize I already did a lesson on this pattern (oops!)
Hey Mark thank you again 🙏. What book do you recommend reading first and is easily digestible ❤ I think I should go for head first architecture. What is your recommendation?
Thanks, Mark, for all of these. I'm a junior software engineer and learning a lot from your videos. I just started from the first ones, however, wanted to comment here as an appreciation
Hi, Mark. Happy anniversary. Can you please provide an example of constraint that can lead to not being able to use kafka instead of messaging solution like rabbit in order address throughput?
Sure; one might be as simple as the company doesn't want to support it 🙂 - technically though, if I require a strict point-to-point solution, messaging is more appropriate. Also, if I need transactions or a chain-of-persistence guarantee (think client-ack mode here) then messaging would be more appropriate. Kafka is best suited for smaller messages, so large message payloads would be another reason.
I have nearly finished studying your book “Fundamentals of software Architecture”. As a college student, it helped me a lot. I would insist on uploading these kind of videos more frequently.
I started out 6 years ago every week, bit the time commitment was too much for my busy schedule, which is why I moved to every 2 weeks. But I appreciate the encouragement!
