Hello everyone, I am Xiaomi! Today I want to talk to you about a question that is often asked in technical interviews-high concurrency current limiting algorithm! This topic is very interesting and one of the challenges we often encounter in our daily work. In this article, I will introduce in detail some common high-concurrency current limiting algorithms and the different scenarios they are suitable for.
What is high concurrency current limiting?
Before we start discussing high-concurrency current-limiting algorithms, let us first clarify what high-concurrency current-limiting is. High concurrency refers to a large number of requests flooding into the system at the same time, which may affect the stability and performance of the system. To protect the system from excessive request pressure, we need to implement a few measures, one of which is throttling.
Current limiting is a strategy to control the request access rate to ensure that the system can still run normally under high concurrency conditions. High-concurrency current-limiting algorithms are key tools used to implement this strategy. They can help us control the number of requests and prevent the system from being overwhelmed by too many requests.
Common high concurrency current limiting algorithms
1. Token Bucket algorithm (Token Bucket): The Token Bucket algorithm is a classic current limiting algorithm. Its working principle is similar to a bucket. There are a certain number of tokens in the bucket, and each token represents a request. license. The request needs to obtain a token before it can be processed. If there are not enough tokens in the bucket, the request will be temporarily blocked or discarded.
The advantage of this algorithm is that it can handle requests smoothly and can flexibly handle sudden high concurrency even in a short period of time. It is suitable for scenarios where the request rate needs to be strictly controlled, such as network requests or message queues.
2. Leaky Bucket: The leaky bucket algorithm is another common current limiting algorithm. It works like a bucket with a fixed capacity. Requests are put into the bucket and flow out at a fixed rate. If the bucket is full, excess requests will be rejected or dropped.
The characteristic of this algorithm is that it can process requests at a constant rate without sudden high concurrency. The leaky bucket algorithm is suitable for scenarios that require a fixed rate to process requests, such as flow control.
3. Sliding Window Counter: The sliding window counter is a current limiting algorithm based on time windows. It divides the request counter into multiple time windows, and each time window has a fixed request limit. Over time, old time windows are removed and new time windows are added.
This algorithm allows bursts of requests over a period of time, but limits the total number of requests within the time window. It is suitable for scenarios where the average rate of requests needs to be controlled, such as API interface current limiting.
4. Token bucket algorithm based on leaky bucket: The token bucket algorithm based on leaky bucket is an algorithm that combines the leaky bucket algorithm and the token bucket algorithm. It uses leaky buckets to control the rate of requests and token buckets to handle burst requests. This algorithm combines the advantages of both and is suitable for scenarios that require flexible control of the request rate.
Scenarios where different algorithms are applicable
Now that we have learned about some common high-concurrency current limiting algorithms, let's look at the scenarios in which they are applicable.
1. Token bucket algorithm: The token bucket algorithm is suitable for scenarios where the request rate needs to be strictly controlled. For example:
- Network request throttling: Protect the server from excessive network requests.
- Message queue current limiting: Ensure that the message queue is not congested by too many messages.
- Interface flow control: Control the access rate of the API interface to prevent excessive requests from causing excessive server load.
2. Leaky bucket algorithm: The leaky bucket algorithm is suitable for scenarios that require a fixed rate to process requests. For example:
- Traffic shaping: Limit egress bandwidth to ensure smooth network traffic.
- Data transmission rate control: Limit the data sending rate to prevent data from entering the receiving end too quickly.
- Request queue control: Limit the processing rate of requests in the request queue.
3. Sliding window counter: The sliding window counter is suitable for scenarios where the average rate of requests needs to be controlled. For example:
- API interface current limit: control the number of requests per second or per minute and allocate resources evenly.
- Ad click limit: Control the click rate of ads to prevent clicks.
- Scheduled task current limiting: Limit the execution rate of scheduled tasks to avoid excessive resource usage.
4. Token bucket algorithm based on leaky bucket: The token bucket algorithm based on leaky bucket is suitable for scenarios that require flexible control of the request rate. For example:
- CDN cache refresh current limiting: Control the CDN cache refresh rate to prevent too many refresh requests.
- Batch task current limiting: Limit the execution rate of batch tasks to avoid impact on the back-end system.
- Sending barrage messages: Control the sending rate of barrage messages to prevent abuse.
Summarize
High concurrency current limiting algorithm is an important tool to protect the system from excessive request pressure. Different algorithms are suitable for different scenarios. The token bucket algorithm is suitable for strictly controlling the request rate, the leaky bucket algorithm is suitable for processing requests at a fixed rate, the sliding window counter is suitable for average rate control, and the token bucket algorithm based on the leaky bucket can flexibly control the request rate .
In practical applications, we can choose appropriate algorithms according to specific needs to protect the stability and performance of the system. I hope this article can help you better understand the high-concurrency current-limiting algorithm and be helpful in interviews. If you have any questions or want to learn more about an algorithm, please leave a message in the comments and I will try my best to answer it!
END
Finally, thank you all for reading. I hope you will go further and further on the road of technology and make progress together! If you like this article, please like and share it so that more people can benefit!
If you have any questions or more technical sharing, please follow my WeChat public account " Know what it is and why "!