About the title

For the first time, t-io proposed a billion-level long connection in public . The author built the "non-perceptive" cluster capability at the t-io layer as early as 2020. After more than a year of self-testing, he also experienced benign feedback from cluster customers in 2021 and 2022. , Today is the time to announce that t-io has a long connection capability of 100 million. For more information about t-io clusters, please go to: "t-io technical white paper" , which has a description of the cluster architecture.

Modify point

PR merge
pom upgrade
Individual local performance optimization

POM coordinates

<dependency>
    <groupId>org.t-io</groupId>
    <artifactId>tio-core</artifactId>
    <version>3.8.1.v20220401-RELEASE</version></dependency>

follow-up plan

With the help of netty, t-io is used to implement the mqtt protocol, so that everyone can use t-io to build their own IoT platform (in order to better take care of netty users, the t-io version of mqtt will try to use the original netty-mqtt version in the naming)
Launched the web test terminal of mqtt-client

Users with more needs can leave a message for feedback, and we will choose the best

Network programming is bitter, but it will be sweet after using t-io

t-io is a high-performance network programming framework developed based on java. Its explosive performance, long battery life stability, extremely easy-to-use API, and rich and sufficient built-in functions make it popular among engineers as soon as it debuts. , and is passed on by continuous word of mouth

source code repository

Why develop t-io

Instead of learning the thick "xxx authoritative guide", it is better to stand up and develop independently and create programming APIs that more people can understand at a glance

t-io's past and present

In 2010, talent-tan took over the communication module of the network management system in a large communication factory. The old code uses traditional IO, and a client needs at least 3 threads to guard, which often causes memory overflow and downtime. Talent-tan was commissioned to rewrite the communication module to address all the concerns of the leadership
In 2012, the talent-nio framework was developed based on nio
In 2013, a transparent transmission module in mycat was written with talent-nio
In 2014, the IM module of heatwave live broadcast was realized with talent-nio
In 2015, talent-tan began to pay attention to aio technology, and at the same time further abstracted the thread pool, lock processing, and concurrent data structure in talent-nio, so that these "Wang Xietang Qianyan" flew into "the vast number of code farmers"
In 2016, talent-nio was rewritten based on aio technology, named talent-aio, the code settled in the code cloud, low-key open source
In 2017, talent-aio changed its name to t-io, and began to spread t-io in the form of news in open source China. In the same year, t-io became a GVP project, and t-io gained a large number of users.
In 2018, peripheral products such as tio-http-server, tio-websocket-server, and tio-webpack were implemented based on t-io. These products not only verified the excellence of t-io, but also promoted the progress of t-io.
In 2019, a Huawei team conducted a 3-6 month copying limit stress test for a smart product based on t-io. t-io withstood the test and laid a solid foundation for the official entry into Huawei's open source library.
In 2020, t-io officially entered Huawei's open source library, the first commercial IM Tan Chat developed based on t-io was officially launched, and the cluster version of t-io was officially developed and passed the stress test.
In 2021, customer feedback based on t-io research and development of Tan Chat is very stable
In 2022, Tan Chat, a cluster version based on t-io's research and development, won the ultimate reputation from users, and also brought a long-term financial guarantee for t-io's continuous investment.

Pain points solved by t-io

The starting point of t-io is to solve the user pain points of network programming, and its mission is to make the world no longer difficult to develop network programs, and look at the surprises t-io brings to users

Easy to learn and use, talent-tan created t-io because the learning cost of similar products on the market is high, so when designing APIs, special attention is paid to user acceptance. The first batch of t-io users mastered t-io just by watching the demonstration project officially provided by t-io
Rolling over the data monitoring capabilities of all well-known similar products----not only provides comprehensive monitoring data, but also ensures excellent performance
Built-in heartbeat timeout check, heartbeat timing sending ability
Extremely polished underlying cluster capabilities, which can seamlessly solve the cluster needs of large-scale products such as IM and the Internet of Things
Dropped automatic reconnect capability
t-io measured performance 1: 1.9G memory stably supports 300,000 TCP long connections: https://www.tiocloud.com/61
t-io measured performance 2: use t-io to run 10.51 million chat messages per second: https://www.tiocloud.com/41
t-io measured performance three: netty and t-io comparison test results: https://www.tiocloud.com/154
Built-in ack message capability
Built-in half-pack sticky pack handling
Self-created tool libraries such as synchronization locks, synchronization safe thread pools, and synchronization data structures provide rich out-of-the-box APIs for business applications
Built-in slow attack defense mechanism to help applications automatically block suspected IPs
Rich ecology, http, websocket, mqtt and a large number of private protocols have been implemented with t-io
Low requirements for development engineers, saving labor costs for enterprises
Excellent performance, saving the cost of hardware deployment for enterprises