This was one of the best explanations of message queues I've seen. Wondering when would an MQ be overkill? Would love to see a video on caching patterns.
Excellent content! A very good example of a message queue is integrating systems. For example I have to bring orders over from an ecommerce system to our erp. But it rate limits me. So I let it chill in the message queue while the workers do thing
Great content. What I find useful is two data structures. One is a time based stream. This allows multiple consumers to access the single stream. The consumers access the stream based upon time stamps. The second is various summary tables for the stream. The asynchronous techniques which are mentioned above and i the video can be implemented on small systems or services using Redis
I have a service with a high response time the last few days. I tried inhibited put queue in mq explorer to reduce the response time, it turned out that there was a problem with another service. how do i solve it?
Why would you EVER use message queue model instead of pub/sub? If a pub/sub can have MULTIPLE subscribers? Also could a message queue broker/bus like RabbitMQ CLONE the quote and give it to MULTIPLE consumers as well? You could keep a copy of the clone which gets stored as a 'log', if this situation is possible with message queue. To me a message queue does not even sound decoupled, at this point why don't we just write directly from app1 to app2 since there is only going to be one producer/consumer.
Because Pub/Sub is when you share the same data to different suscribers. Point 2 Point is when you have specific data to the consumer. Ex, you want to send orders to the customer, each customer has different orders (id, products, price) you cant send the same order to all customers that need an order. And Pub/sub example is when you have newsletter, and many customers are subscribed to your news letter. Hope it makes sense.
Another thing, you cant send information from app 1 to app2 directly as you said. Imagine you have an antifraud system, that delays like 1 minute to check if the order is fraud or not. Are you going to let the customer wait for 1 minute until the anti fraud system decides if it's fraud or not? Of Course not. The queue is going to send that information async, and wait the response asyncronous. Another example. Customer purchases something, emails normally takes like 10s or even more to be sent. Customer is not going to wait on the checkout until it's email is sent, it can be on a queue running on background.
This is one of the best explanations of message queues I've seen
This was one of the best explanations of message queues I've seen. Wondering when would an MQ be overkill? Would love to see a video on caching patterns.
Excellent content! A very good example of a message queue is integrating systems. For example I have to bring orders over from an ecommerce system to our erp. But it rate limits me. So I let it chill in the message queue while the workers do thing
14:10 I think we have come a long way, having java for frontend and javascript for backend :). Anyway, great video as always.
Extremely well-presented. IBM is a great source of content for these things.
How is this guy writing backwards so accurately
Reverted the stream)
It’s mirrored. IBM shared some details on it.
this is a pretty commonly used technique. you'll never see it live though :)
@@johnyepthomi892that’s a secret information
Excellent! Thank you very much for breaking Message Queuing down into digestible bits 👏
I am learning distributed systems, and your video helps me a lot. Thank you for your effort. Keep going!
Great content. What I find useful is two data structures.
One is a time based stream. This allows multiple consumers to access the single stream. The consumers access the stream based upon time stamps.
The second is various summary tables for the stream.
The asynchronous techniques which are mentioned above and i the video can be implemented on small systems or services using Redis
This was the best explanation I’ve seen on MQuing
message queues are architecture agnostic. still a great video.
Excelent video! I'm following the whole series
Great content. Really helped to clear many of my assumptions. Thanks.
Very well explained
Simple to understand. Thank you for this!
Great explanation sir
I love your surname, keep it up!
Best explanation ever.
Thanks. Great content!
Useful. Thank you.
Great video, fantastic, exactly how it should be!!
great video
Great content!
I have a service with a high response time the last few days. I tried inhibited put queue in mq explorer to reduce the response time, it turned out that there was a problem with another service. how do i solve it?
Jamil!!! Thanks!
This is a way to chat with live stream videos?
nice
What are you writing on?
Search on "lightboard videos".
subbed
Is he... writing mirrored?
See ibm.biz/write-backwards
Yes😂
Why would you EVER use message queue model instead of pub/sub? If a pub/sub can have MULTIPLE subscribers?
Also could a message queue broker/bus like RabbitMQ CLONE the quote and give it to MULTIPLE consumers as well? You could keep a copy of the clone which gets stored as a 'log', if this situation is possible with message queue.
To me a message queue does not even sound decoupled, at this point why don't we just write directly from app1 to app2 since there is only going to be one producer/consumer.
Because Pub/Sub is when you share the same data to different suscribers. Point 2 Point is when you have specific data to the consumer. Ex, you want to send orders to the customer, each customer has different orders (id, products, price) you cant send the same order to all customers that need an order. And Pub/sub example is when you have newsletter, and many customers are subscribed to your news letter. Hope it makes sense.
Another thing, you cant send information from app 1 to app2 directly as you said. Imagine you have an antifraud system, that delays like 1 minute to check if the order is fraud or not. Are you going to let the customer wait for 1 minute until the anti fraud system decides if it's fraud or not? Of Course not. The queue is going to send that information async, and wait the response asyncronous. Another example. Customer purchases something, emails normally takes like 10s or even more to be sent. Customer is not going to wait on the checkout until it's email is sent, it can be on a queue running on background.
Would this be the solution to my 30s + long request time
so you have practiced to write backwards ?????? OMG