How can an http request return the absolute number of the current request under high concurrency?

For example, my first access request returned json:
{"n": 1}

My 100th access request returned json:
{"n": 100}

The traditional method of writing to the database and then checking the database to return the results seems not guaranteed when the concurrency is large. What should I do? This should be the most simplified question

Reply content:

For example, my first access request returned json:
{"n": 1}

My 100th access request returned json:
{"n": 100}

The traditional method of writing to the database and then checking the database to return the results seems not guaranteed when the concurrency is large. What should I do? This should be the most simplified question

The simplest way is to build a mysql table with an auto-incremented primary key of id, then insert a record for each request, and then read the record. The id read out is the value you want.

Then you can easily handle high-concurrency scenarios based on the id value. For example, [Instant Kill] can use the rule that the id is less than 300 and divisible by 6 to consider the instant kill successful; [Lottery] can use the rule that the id is less than 300 and divisible by 6; [Lottery] can use the rule that the id is divisible by 100 (one hundredth) Probability) as winning etc.

If you implement it yourself, it is nothing more than a single-threaded infinite loop processing socket requests and maintaining a global variable. It is not as convenient and reliable as using ready-made mysql.

If it is java, a global AtomicLong can meet your needs, getAndIncrement atomic operation, plus volatile modification, if it is other languages, it will be similar

Using setnx(id) in redis, a single thread guarantees increment by 1 each time, and it is also an in-memory database, so it is super fast.

• Read operation: use cache

• Write operation: asynchronous writing using queue

In pure Java, the counter object can be made into a singleton, and all requests to the calculator can be intercepted by filter and incremented by 1 (synchronization is required). I don’t know what you mean by database, {n : 100}, n is taken from the database?

In fact, what you want to do is a permanent memory queue, which is queued and processed according to the request.
On a single machine, you can try reading and writing SQLite from /dev/shm on the Linux memory file system (tmpfs).
Reading files does not require going through the network , and there is no need to implement memory resident, lock, auto-increment and unique constraints by yourself.

<code><?php
header('Content-Type: text/plain; charset=utf-8');
// sudo mkdir -m 777 /dev/shm/app
$file = '/dev/shm/app/data.db3';
$ddl = "
BEGIN;
CREATE TABLE IF NOT EXISTS queue (
    id           INTEGER PRIMARY KEY AUTOINCREMENT,
    user_id      INTEGER
);
CREATE UNIQUE INDEX IF NOT EXISTS queue_user_id_idx ON queue(user_id);
COMMIT;
";
if(!file_exists($file)) {
    //多核下多进程并发时可能都会进入到这个判断分支,所以DDL中要用IF NOT EXISTS
    $db = new PDO('sqlite:'.$file);
    $db->exec($ddl); // pdo_sqlite 的 query 和 prepare 不支持一次执行多条SQL语句
} else {
    $db = new PDO('sqlite:'.$file);
}
$stmt = $db->prepare('INSERT INTO queue(user_id) VALUES(?)');
$stmt->execute(array(time())); //time()换成你的用户ID
echo $stmt->rowCount()."\n";   //查询中受影响(改动)的行数,插入失败时为0
echo $db->lastInsertId();      //插入的自增ID,插入失败时为0
// php -S 127.0.0.1:8080 -t /home/eechen/www >/dev/null 2>&1 &
// ab -c100 -n1000 http://127.0.0.1:8080/</code>

The simplest way is to use redis's zset for auto-increment, which is efficient and simple. If you are using a single machine, you can also consider using atomiclong (it will become invalid after a shutdown and restart)