Apache Kafka is an open-source stream processing platform developed by the Apache Software Foundation written in Scala and Java. The Kafka event streaming platform is used by thousands of companies for high-performance data pipelines, streaming analytics, data integration, and mission-critical applications.
N/A
Azure Service Bus
Score 8.6 out of 10
N/A
Microsoft offers Azure Service Bus as a reliable cloud messaging as a service (MaaS) and simple hybrid integration solution.
Apache Kafka is well-suited for most data-streaming use cases. Amazon Kinesis and Azure EventHubs, unless you have a specific use case where using those cloud PaAS for your data lakes, once set up well, Apache Kafka will take care of everything else in the background. Azure EventHubs, is good for cross-cloud use cases, and Amazon Kinesis - I have no real-world experience. But I believe it is the same.
If you need a cloud-based service bus or a simple to use queue/topic/routing/pub-sub service, then Azure Service Bus is a very good choice at a reasonable price and performance. Typically on-premise we'd use RabbitMQ because it "just works", but if you're building a "cloud-first" application, then this is the one to go with. It's especially easy to integrate with if you're already embedded in the Microsoft ecosystem.
Really easy to configure. I've used other message brokers such as RabbitMQ and compared to them, Kafka's configurations are very easy to understand and tweak.
Very scalable: easily configured to run on multiple nodes allowing for ease of parallelism (assuming your queues/topics don't have to be consumed in the exact same order the messages were delivered)
Not exactly a feature, but I trust Kafka will be around for at least another decade because active development has continued to be strong and there's a lot of financial backing from Confluent and LinkedIn, and probably many other companies who are using it (which, anecdotally, is many).
Acting as a basic queuing service it works very well.
One of the best parts is that Azure Service Bus can work over HTTPS which helps in strict firewall situations. There is a performance hit if you choose to use HTTPS.
The routing capabilities are quite good when using topics and subscriptions. You can apply filters using a pseudo-SQL-like language though the correlation filters are quick and easy options.
Costs are very reasonable at low-ish volumes. If you're processing 10's of millions of messages a month... it may be a different story.
Sometimes it becomes difficult to monitor our Kafka deployments. We've been able to overcome it largely using AWS MSK, a managed service for Apache Kafka, but a separate monitoring dashboard would have been great.
Simplify the process for local deployment of Kafka and provide a user interface to get visibility into the different topics and the messages being processed.
Learning curve around creation of broker and topics could be simplified
Apache Kafka is highly recommended to develop loosely coupled, real-time processing applications. Also, Apache Kafka provides property based configuration. Producer, Consumer and broker contain their own separate property file
Support for Apache Kafka (if willing to pay) is available from Confluent that includes the same time that created Kafka at Linkedin so they know this software in and out. Moreover, Apache Kafka is well known and best practices documents and deployment scenarios are easily available for download. For example, from eBay, Linkedin, Uber, and NYTimes.
I used other messaging/queue solutions that are a lot more basic than Confluent Kafka, as well as another solution that is no longer in the market called Xively, which was bought and "buried" by Google. In comparison, these solutions offer way fewer functionalities and respond to other needs.
RabbitMQ is simple and awesome... but so is Azure Service Bus. Both accomplish the same thing but in different environments. If you're building a cloud-native application - especially one that is serverless by design - Azure Service Bus is the only real choice in Azure. It works well, it's performance, and it's reasonably priced in the Standard tier. From our testing, RMQ is more performant, but it's hard to compare service-based implementations vs RMQ installed on VMs.
Positive: Get a quick and reliable pub/sub model implemented - data across components flows easily.
Positive: it's scalable so we can develop small and scale for real-world scenarios
Negative: it's easy to get into a confusing situation if you are not experienced yet or something strange has happened (rare, but it does). Troubleshooting such situations can take time and effort.
