Redis: the key technology for building a real-time ranking system

Redis is an open source high-performance key-value database system. It is widely used in real-time rankings due to its fast read and write speed, support for multiple data types, rich data structures and other characteristics. in the system. The real-time ranking system refers to a system that sorts data according to certain conditions, such as points rankings in games, sales rankings in e-commerce, etc.

This article will introduce the key technologies used by Redis in building a real-time ranking system, as well as specific code examples. The content includes the following parts:

1. Data type of Redis
2. Sort algorithm
3. Leaderboard implementation in Redis
4. Scalability
5. Data types of Redis

Redis supports multiple data types, including strings, hash tables, lists, sets, and ordered sets.

Ordered set is the key data type to implement the ranking list. It can easily sort the data according to the value of a certain field. Each element in the sorted set has a score and is sorted according to the score. When the scores are the same, sort them lexicographically. Each element in an ordered set has a unique member value that uniquely identifies the element.

Specific ordered set related commands include: ZADD, ZREM, ZRANGE, etc.

2. Sort Algorithm

The real-time ranking system needs to perform fast and accurate sorting, so it is necessary to choose an appropriate sorting algorithm. Redis uses the skip list algorithm to implement ordered collections.

The skip list is a randomized data structure, similar to a linked list, but each node has multiple pointers, making the search more efficient. The nodes in the jump table are arranged in increasing order, and each node has a random "level number", and each level has a pointer to the node in the next level. This "number of layers" is randomly generated and can be adjusted according to needs.

The time complexity of the jump table is O(log n) and the space complexity is O(n), which can well meet the needs of the real-time ranking system.

3. Implementation of rankings in Redis

Using Redis to implement rankings requires the following steps:

1) Create an ordered set

Use the ZADD command to create an ordered set and add elements (members and scores) to it. Each member has a unique identifier, for example, a user ID can be used in a game, or an item number can be used in e-commerce.

2) Get ranking data

Get the elements in the ordered set according to the ranking. Use the ZRANGE command to perform interval query on the ordered set. For example, to get the top 10 user information, you can use the ZRANGE command 0 9 WITHSCORES command.

3) Update Score

When the user’s score changes, the corresponding score in the ordered set needs to be updated. Updates can be made using the ZADD command.

4) Get the ranking

Get the ranking on the leaderboard based on the user ID. You can use the ZRANK command to get the ranking corresponding to the user ID.

The following is a sample code for a Redis-based ranking list implementation, which uses the skip table algorithm:

import redis

# 连接 Redis 数据库
r = redis.StrictRedis(host='localhost', port=6379, db=0)

# 创建排行榜
def create_leaderboard():
    r.zadd('leaderboard', {'Tom': 100, 'Jerry': 90, 'Peter': 80, 'Lucy': 70})

# 获取排行榜前 N 名的数据
def get_topN(n):
    data = r.zrevrange('leaderboard', 0, n - 1, withscores=True)
    return data

# 更新用户积分
def update_score(username, score):
    r.zadd('leaderboard', {username: score})

# 获取指定用户在排行榜中的排名
def get_rank(username):
    rank = r.zrank('leaderboard', username)
    return rank

# 测试代码
if __name__ == '__main__':
    create_leaderboard()
    print(get_topN(3))  # 输出前 3 名的数据
    update_score('Tom', 95)  # Tom 的积分变为 95
    print(get_topN(3))  # 再次输出前 3 名的数据，应该会有变化
    print(get_rank('Tom'))  # Tom 目前的排名是第 2 名

4. Scalability

With As the amount of data in the real-time ranking system increases, it may encounter system performance bottlenecks. In order to ensure the scalability of the system, Redis cluster can be used to horizontally expand the real-time ranking system.

Redis cluster refers to a Redis instance running distributedly on multiple servers. It stores a large amount of data on different nodes to achieve high availability and load balancing of data. Redis cluster can be implemented using Redis Cluster or Redis Sentinel.

You need to pay attention to the following points when implementing a Redis cluster:

1) Data partitioning: Scattered storage of data on different nodes can effectively reduce the load pressure on a single node.

2) Read and write separation: Use the master-slave architecture to achieve read and write separation, which can allocate read operations to multiple nodes and improve the reading efficiency of the system.

3) Fault-tolerance mechanism: Use Redis Sentinel or other fault-tolerance mechanisms to implement automatic failover to ensure high availability of the system.

Summary:

Redis is a powerful tool for implementing real-time ranking systems. It supports multiple data types and rich data structures, and can effectively implement data sorting and query. Efficient sorting can be achieved using the skip table algorithm, and coupled with the horizontal expansion of the Redis cluster, the real-time ranking system can handle large amounts of data and ensure high system availability. The code examples provided in this article can be used as basic components to implement a real-time ranking system, and readers can modify and optimize them according to actual needs.

The above is the detailed content of Redis: the key technology for building a real-time ranking system. For more information, please follow other related articles on the PHP Chinese website!