4 ms·

Most of the things you do for a user are not blocking, that's the whole point of node. Instead of blocking, almost everything is done with asynchronous callback

by boucher 12y ago

Most of the things you do for a user are not blocking, that's the whole point of node. Instead of blocking, almost everything is done with asynchronous callbacks.

amelius 12y ago
And while you are servicing a nonblocking request, you are effectively blocking other users, because the CPU can do one thing at a time. That is the main pitfall of asynchronous programming as opposed to multithreaded programming.
- mcs 12y ago
  In thread-per-connection server applications, the concurrency is exactly the number of threads spun up for the pool of connections. In node, it achieves that concurrency without a thread pool for the connections themselves, and thus you get less resource usage (memory mostly) for the same number of connections as the daemon scales up. Node's real draw, aside from pretty easy concurrency, is the package ecosystem.
  - tracker1 12y ago
    Not to mention the overhead of switching threads... Moving context in and out of secondary memory (RAM, disk) is pretty big. And when you get to thousands of threads per core, the context switching becomes very costly. Some modern systems use thread pools, and manage state switching internally to avoid this overhead at the CPU layer... just the same it is costly. I worked on a simulator a few years ago that was being written in many threads (each virtual character in play would have its' own worker thread)... this lead to a much lower number of users/system than should have been needed... switching to an event-loop and message bus reduced the overhead significantly.
- jkrems 12y ago
  You are not blocking other users - the CPU is busy and couldn't do other stuff even if it tried. With using one process per core (plus a spare core for "other stuff") you get full CPU utilization. You are right that node is not the right tool for doing large CPU-bound tasks as part of HTTP request handling. If that was your point: yes, that's true and that won't change. But if most of what you're doing is forking out to other services and some light assembly (JS is a scripting language after all), then node is a very nice fit.
  - amelius 12y ago
    But what if your tasks need to talk to other systems, that may use mutexes to control concurrent access. In that case, your HTTP request handling logic will just block on those mutexes, and all is lost, since the event loop will be stuck. Now, you could of course rewrite those 3rd-party systems to use your event-loop instead of plain mutexes, but the point is that threads are a more natural fit for multiprocessing, from a software-engineering point of view, because they allow one to compose systems more easily. In other words, threads have been invented for a reason. Otherwise, we might as well just go back to the Windows 3.11 era, with its cooperative multitasking model.
    - tracker1 12y ago
      Your calls to third party systems should be through a callback interface.. they should not be blocking inline. The event loop will continue. There are plenty of ways you can process out of band requests... you could use a generic-pool with a size of one instance if you really need to... However, if you are really blocking access to a single client at a time, it's likely node won't be your bottleneck.
- vertex-four 12y ago
  The CPU can only do one thing at a time even with threads. It just hides that fact from you better. In an app that grabs a page from a database and returns it, the node.js app will receive a request from a user, make a request to the database, receive a request from another user, make a request to the database, then wait for either database request to return, then return it to the user, then when the other returns, return that to the user. Your program doesn't deal with one whole request at a time, it chops it up into lots of little "wait for this event" parts and node.js has an event loop that listens for events, which can happen in an arbitrary order. So you can be waiting for hundreds of events to happen for hundreds of requests, and as node.js doesn't make you wait for the first one to process the second, nobody's holding up anyone else. Threading does nearly exactly this automatically... except that you need memory per thread in order to maintain the illusion that each thread has complete control over the CPU (its own stack, etc), and it can also split up CPU-bound workloads automatically. node.js needs much less memory per event it's waiting for.

by boucher 12y ago

Most of the things you do for a user are not blocking, that's the whole point of node. Instead of blocking, almost everything is done with asynchronous callbacks.

amelius 12y ago

And while you are servicing a nonblocking request, you are effectively blocking other users, because the CPU can do one thing at a time. That is the main pitfall of asynchronous programming as opposed to multithreaded programming.

mcs 12y ago
In thread-per-connection server applications, the concurrency is exactly the number of threads spun up for the pool of connections. In node, it achieves that concurrency without a thread pool for the connections themselves, and thus you get less resource usage (memory mostly) for the same number of connections as the daemon scales up. Node's real draw, aside from pretty easy concurrency, is the package ecosystem.
- tracker1 12y ago
  Not to mention the overhead of switching threads... Moving context in and out of secondary memory (RAM, disk) is pretty big. And when you get to thousands of threads per core, the context switching becomes very costly. Some modern systems use thread pools, and manage state switching internally to avoid this overhead at the CPU layer... just the same it is costly. I worked on a simulator a few years ago that was being written in many threads (each virtual character in play would have its' own worker thread)... this lead to a much lower number of users/system than should have been needed... switching to an event-loop and message bus reduced the overhead significantly.
jkrems 12y ago
You are not blocking other users - the CPU is busy and couldn't do other stuff even if it tried. With using one process per core (plus a spare core for "other stuff") you get full CPU utilization. You are right that node is not the right tool for doing large CPU-bound tasks as part of HTTP request handling. If that was your point: yes, that's true and that won't change. But if most of what you're doing is forking out to other services and some light assembly (JS is a scripting language after all), then node is a very nice fit.
- amelius 12y ago
  But what if your tasks need to talk to other systems, that may use mutexes to control concurrent access. In that case, your HTTP request handling logic will just block on those mutexes, and all is lost, since the event loop will be stuck. Now, you could of course rewrite those 3rd-party systems to use your event-loop instead of plain mutexes, but the point is that threads are a more natural fit for multiprocessing, from a software-engineering point of view, because they allow one to compose systems more easily. In other words, threads have been invented for a reason. Otherwise, we might as well just go back to the Windows 3.11 era, with its cooperative multitasking model.
  - tracker1 12y ago
    Your calls to third party systems should be through a callback interface.. they should not be blocking inline. The event loop will continue. There are plenty of ways you can process out of band requests... you could use a generic-pool with a size of one instance if you really need to... However, if you are really blocking access to a single client at a time, it's likely node won't be your bottleneck.
vertex-four 12y ago
The CPU can only do one thing at a time even with threads. It just hides that fact from you better. In an app that grabs a page from a database and returns it, the node.js app will receive a request from a user, make a request to the database, receive a request from another user, make a request to the database, then wait for either database request to return, then return it to the user, then when the other returns, return that to the user. Your program doesn't deal with one whole request at a time, it chops it up into lots of little "wait for this event" parts and node.js has an event loop that listens for events, which can happen in an arbitrary order. So you can be waiting for hundreds of events to happen for hundreds of requests, and as node.js doesn't make you wait for the first one to process the second, nobody's holding up anyone else. Threading does nearly exactly this automatically... except that you need memory per thread in order to maintain the illusion that each thread has complete control over the CPU (its own stack, etc), and it can also split up CPU-bound workloads automatically. node.js needs much less memory per event it's waiting for.