Queue, production consumption model. html

A. Python queue:

• Overview:
• Import queue
• is created in Python memory and the Python process exits, the queue is emptied
• Ways to view these queue classes through Pycharm

• FIFO queue:
• Q = Queun. Queue ([10]) # Create a "FIFO" queue, accept up to 10 data, put the 11th card master, no parameter defaults to 0, that is, an unlimited number of
• Q.qsize () # View the number of elements in the current queue
• Q.maxsize # Maximum number of queues

• Q.put ([timeout=2]) # wait 2 seconds, if there is no vacant position after 2 seconds error
• Q.put ([Block=false]) # Default is blocked, and always wait for the card master, if False there is no location when the direct error

• Q.get () # Default Fetch data, if there are no elements in the queue, it will block the card master
• Q.get ([Block=false]) # does not block,
• Q.get (timeout=2) # timeout time

• Q.join () # If the tasks in the queue (get and insert are tasks) are not completed (not empty, and the counter is zeroed), the main process does not end, wait
• Q.task_done () # This code is often written after each queue method executes, indicating that the method has completed and the counter is zeroed

• Last-in-first-out queue (stack), which inherits the queue, can use the methods in its parent class:
• Q = queue. Lifoqueue () # Last on first out

• Priority queue (Inherits queue):
• Q = queue. Priorityqueue ()
• Q.put (One, ' Alex ') # When the element is placed, a priority is attached, and the queue is reordered in priority order

• Bidirectional queue
• Q = Queue.deque ()

Ii. Producer, consumption model (queue)

Import Queue

Import

Time

q = queue. Queue (20) # Create a FIFO queue

 

# "producer" each chef has been creating buns until the queue is full of blocking,

def productor (ARG):

While

True:

q.put (str (ARG) + ': Bun ') # Create a bun into the queue, cook number: Bun

# "consumer" every customer has been consuming buns until the queue is empty and blocked

def Consumer (ARG):

While

True:

Print (ARG, q.get ()) # take out a bun from the queue, which chef: Bun

Time.sleep (2)

For

I in range (3): # Create 3 chefs

t = Threading. Thread (Target=productor, args= (i,))

T.start ()

For

J in Range (20): # 20 people created

t = Threading. Thread (Target=consumer, args= (,))

T.start ()

• Use Queue benefits:
• Conclusion:
• Using queues, you can increase the concurrency of the client so that client requests are not acceptable (denial of service)
• To really improve processing power, you have to add back-end worker threads
• Explain:
• Scenario: Server concurrency is assumed to be 4, and the client is a dynamic request (insert INTO.) ), each INSERT statement executes an event if it is 20s

• Do not use queues:

• Process:
• The server receives a request, immediately takes a worker thread to process it, and remains connected to the client during processing (the connection is in a pending state)
• Concurrency capability: 20s, up to 4 concurrent
• Request Processing: 20s,4 worker threads can handle 4 requests, 1 worker threads can handle only one (incoming request, no one is processed and no egg is used)

• Using queues

• Process:
1. When the server receives a user request, it is dropped directly into the queue and then disconnected from the client.
2. Back-end working threads (producers) are constantly fetching requests from the queue for processing
3. When client queries the result of the last request execution again,
• If the request is still in the queue, the return queue location "Request is processing, you have n people in front of you"
• If the request is not in the queue, the server needs only select down DB to return (much faster than insert)

• Concurrency: Unlimited because you can disconnect after you drop in a queue, that is, there are always 4 connections available
• Request processing: Within 20s, 4 worker threads can handle 4 requests and 1 worker threads can handle only 1 requests

Queue, production consumption model. html