ECS 189e

Sam King

Processes, threads, and event-driven programming: general concepts

Administrative

• HW4 due on Saturday
• Quiz 3 is out
• Sprint planning due today — please pay close attention to what we’re asking for and see the course website for details
• Project proposals due today
• Please double check your groups in canvas and let me know if there are any issues

​

We created your groups so all project assignments are submitted as a group

Administrative

Last time: API calls

This time: The history of threads and event driven programming

Next time: Concurrency in iOS

Next next time: Synchronization in iOS

Threads and concurrency

• Motivation
• OSes getting complex
• Multiple users, programs, I/O devices, etc.
• How to manage this complexity?

​

• Decompose or separate hard problems into simpler ones

• Processes decompose mix of activities running on a processor into several parallel tasks (columns)

​

​

​

​

​

• Each job can work independently of the others
• For any area of OS, ask:
• What interface does the hardware provide?
• What interface does the OS provide?

What’s in a process?

• Definition of a process
• (informal) a program in execution. A running piece of code along with all the things the program can read/write
• Note: process != program
• (formal) one of more threads in their own address space
• Play analogy

• Thread
• Sequence of executing instructions from a program (i.e., the running computation)
• Active
• Play analogy: Actors (threads) reading from a script (the program)

​

• Address space
• All the data in the process uses as it runs
• Passive (acted upon by the thread)
• Play analogy: all the object on the stage in a play

Multiple threads

• Can have several threads in a single address space
• Play analogy: several actors on a single set. Sometimes interact (e.g., dance together), sometimes do independent tasks

​

• Private state for a thread vs. global state shared between threads

A little bit of history

• Computer systems circa 2000 were uniprocessor, I/O bound
• Web servers were the research problem of the day

General flow for handling a web server request

handleWebRequest() {

socket = newClientConnection(serverSock)

request = readHTTPRequest(socket)

object = accessDatabase(request)

response = accessFilesystem(request)

sendResponse(socket, response)

}

• Handle one request at a time
• Easy to program, slow
• No overlapping disk requests with computation or with network receive

​

​

​

Request 1

Start disk I/O req 1

Disk I/O ret req 1

Reply to req 1

Request 2

Process req 1

Process req 2

• Event-driven with async I/O
• Need to keep track of outstanding requests

​

Request 1

Proc req 1

Start I/O 1a

Request 2

Proc req 2

Start I/O 2a

Request 3

Proc req 3

Start I/O 3a

Disk I/O 1a ret

Reply req 1

activeFds.add(serverSock)

handleWebRequest() {

activeFd = select(activeFds)

// giant state machine to track pending

// requests (yuck)

switch (activeFd) {

case serverSock:

// new request

case isHttpRequest(activeFd):

// read request, store state when done

case isDatabaseFd(activeFd):

// handle response from database

…

}

​

Web server using threads

• Each thread handles one request

Request 1

Request 2

Request 3

handleWebRequest() {

while (true) {

socket = newClientConnection(serverSock)

createNewThread(webRequestThread, socket)

}

}

​

webRequestThread(socket) {

request = readHTTPRequest(socket)

object = accessDatabase(request)

response = accessFilesystem(request)

sendResponse(socket, response)

}

​

Is Swift on iOS multi-threaded or event driven?

Sketch out from a high level what your iOS Swift-based web server would look like

A Swift based server

Socket.acceptRequest() { request in

Request.lookupData(request) { data in

Request.fetchTemplate(data) { template in

request.respond(data, template)

}

}

}

​

​

Web server programming model advantages and disadvantages

• Advantages of thread example?

​

​

• Advantages of event-driven example?

​

​

• Advantages of closure-based example?

​

Benefits and uses of threads

• Thread system in operating system manages the sharing of the single CPU among several threads
• Applications get a simpler programming interface

​

• Typical domains that use multiple threads
• Physical control
• Slow component?
• Window system (1 thread per window)

​

• Network server

​

• Parallel programming (for using multiple CPUs)

​

​

​