When it comes to mature message queues, there is no doubt that most people's first reaction must be Kafka.
As a star product, Kafka can fully meet the high throughput and high concurrency requirements in massive data scenarios. In just a few years, Kafka has been regarded as a top-tier enterprise such as Ali, Tencent, Baidu, ByteDance, Netflix, and Twitter. Technical core - it can be said that I haven't heard of Kafka today, which means I don't know that the earth is round!
However, in the world of technology, there has never been a saying "Thirty years in Hedong and thirty years in Hexi". It is an exaggerated metaphor: many technologies are often born today, explode tomorrow, and become obsolete the day after tomorrow . Is there really no other product on the market that is likely to replace Kafka? How long will Kafka, with its halo and crown on its head, remain popular?
01
The Fall of the King of Messaging Systems Is Soon
Kafka's biggest opponent is coming?
Known as the king of messaging systems, Kafka has all the features: high throughput, low latency, fault tolerance, persistence, scalability, especially the well-known high throughput, Kafka can produce about 250,000 messages per second (50 MB) at 550,000 messages (110 MB) per second ! Another huge advantage of Kafka is fault tolerance. It has an inherent function that can deal with node failures in the cluster by itself.
Kafka is of course very strong, but in the process of operation and maintenance and practice, there are always some thorny problems. For example:
Poor scalability, stripping Broker means that topic partitions and replicas must be copied, which is very inefficient;
Lack of consistency, once the API changes, problems are likely to occur;
The storage cost is very high, almost no one uses Kafka to store data for a long time;
Native multi-tenancy without complete isolation from tenants requires a self-configuring solution.
Overlooking the technical ecology, is there a platform that can not only have the advantages of Kafka, but also avoid its defects, and at the same time integrate a series of features of each MQ? Yes, the perfection you want, the next-generation cloud-native messaging platform, Apache's top project Pulsar can meet it!
02
Separation of storage and computing, layered and sharded cloud-native architecture
Apache Pulsar
Born to be a king!
Many companies often choose between Kafka and RabbitMQ or RocketMQ, because in real-time streaming architecture, messaging use cases can be divided into two categories: queues and streams. And the best part of Pulsar is here, it combines queues and streams into one, and unifies it into a messaging model (producer-topic-subscription-consumer)!
In other words, Pulsar can not only handle high-speed real-time scenarios like Kafka, but also support traditional standard message queue modes , such as multi-consumer and failover subscription.
A Pulsar system that directly allows you to have various features of RabbitMQ + RocketMQ + Kafka! What's even more amazing is that Pulsar adopts a cloud-native architecture. Its architecture design can make full use of distributed and elastically expand cloud resources, and permanently say goodbye to insufficient performance.
Architecture diagram of Pulsar message sending mode
At present, Pulsar has been widely practiced by giants in various fields such as Tencent, Zhaopin, Tuya Smart, China Mobile, China Telecom, 360, Dada Group (JD Daojia), Suning, Ping An Securities, Huawei, etc., and has the potential to catch up with Kafka in an all-round way. The trend is very similar to the situation when Ali promoted Flink to replace Spark!
In the second half of the mobile Internet in 2021, high concurrency and high throughput are no longer the primary pain points. It is almost certain that Pulsar will be quickly applied by more and more major manufacturers and become the next outlet!
In order to help you quickly learn, understand, master and practice the powerful Pulsar, and catch up with the next high-paying trend in the technology circle, we specially invited the co-founder of StreamNative, who is also a PMC member of the open source projects Apache Pulsar and Apache BookKeeper, and Zhai of Committer Teacher Jia , co-founder of Yonaixue Education, former senior big data architect of 58 Zhuanzhuan Company, and head of Hadoop platform, Mr. Li Xiyuan , launched the first domestic Pulsar course and created the latest issue of "Apache Pulsar - Trillion-level Cloud Native Distributed Information Platform" online column course !
The original price is ¥499, and you can learn for free at 0 yuan for a limited time!
Long press to scan the code????, lock the 0 yuan quota
9.8 Special price registration, full refund after 3 days of study!
Valid only for the first 50 people, first come first served!
The original price of the first set of Pulsar online column courses in China is 499 , and now you can get it for 9.8 half a cup of milk tea, and you can get a full refund after 3 days of study, which is equivalent to learning the course for free , which is definitely a good deal. 12 high-quality courses on the necessary technical knowledge of big data architects carefully polished by famous teachers, from May 24th to 26th, 3 days of fast training, 20:00-22:00 in the evening , to help you catch up with the next industry trend.
Important reminder: Once you sign up, you can watch forever with unlimited times and dates. At the same time, there are surprise prizes waiting for you to win for free during the course. Xiaomi body fat scale, Joyoung breakfast machine, 360 smart camera and other 100 yuan gifts are all free of charge. The event is real and effective!
Purchasing benefits
12 dry goods practical lessons, take you to quickly master:
Architecture selection of message system;
In-depth grasp of the architecture design of the new generation of trillion-level message system Pulsar;
Get the core features of Pulsar, a new generation of trillion-level message system;
Master the core design of Apache BookKeeper;
Grasp the latest technology of big data and first-hand industry dry goods trends and trends.
03
Why do I recommend this course to you
Go deep into the core of the Pulsar architecture from 0 to 1
Through special design and abstraction, Pulsar uniformly supports two message consumption modes of Streaming and Queueing, maintaining the high performance of Streaming mode and the flexibility of Queueing mode. While ensuring the performance and throughput of the big data message system, Pulsar provides more enterprise-level features, including convenient operation and maintenance and expansion, flexible message model, multi-language API, multi-tenant, remote multiple backup, and data The strong persistence consistency and so on solve many shortcomings of the existing open source message system.
In short, through the in-depth analysis from the "point" of big data core technology, to the "line" of Pulsar architecture design , and to the "surface" of enterprise application practice , it will help students master the core technology of Pulsar in all aspects.
Star teaching and research team
Quickly grasp with your hands
Sister Xuan, a famous teacher, received rave reviews
Teaching, research and teaching are more secure
Swipe left and right to see more
The first set of Pulsar courses in China
What Big Data Architects Must Know
Fully handle the Pulsar architecture design practice
Master the Next Generation Cloud-Native Messaging Platform
12 major nodes 3- day crash course
Original price 499 limited time special 9.8
Quickly get the necessary skills for big data architects!
After 3 days of study, full refund!
????????????
This official account is limited to the first 50 special purchases
Please be patient and wait for the course advisor to pass
suitable for whom to learn
Technology waves follow one after another, and there will always be new technologies emerging. As a technology practitioner, what I fear most is that I am still using old technology to solve problems that others have solved long ago. If you are a:
Distributed system development engineer
Big Data Platform Engineer
Big Data R&D Engineer
Big Data Architect
Data Warehouse Engineer
Back-end development engineer (language direction such as Java/Go/C++/PHP/Python)
Architect (big data/cloud native/platform/system/business, etc.)
Then, this practical excellent course is tailor-made for you!
Do education with your heart, Naixue will grow up with you!
About Nai Xue Education
Click to view "Read the original text" to learn more about the course content of Naixue Education!