Related
My girlfriend was asked the below question in an interview:
We trigger 5 independent APIs simultaneously. Once they have all completed, we want to trigger a function. How will you design a system to do this?
My girlfriend replied she will use a flag variable, but the interviewer was evidently not happy with it.
So, is there a good way in which this could be handled (in a distributed context)? Note that each of the 5 API calls are made by different servers and the function to be triggered is on a 6th server.
The other answers suggesting Promises seem to assume all these requests necessarily come from the same client. If the context here is distributed systems, as you said it is, then I don't think those are valid answers. If they were, then the interview question would have nothing to do with distributed systems, except to essay your girlfriend's ability to recognize something that isn't really a distributed systems problem.
And the question does have the shape of some classic problems in distributed systems. It sounds a lot like YouTube view counting: How do you achieve qualities like atomicity and consistency in a multi-threaded, multi-process, or multi-client environment? Failing to recognize this, thinking the answer could be as simple as "a flag", betrayed a lack of experience in distributed systems.
Another thing about that answer is that it leaves many ambiguities. Where does the flag live? As a variable in another (Java?) API? In a database? In a file? Even in a non-distributed context, these are important questions. And if she had gone on to address these questions, even being innocent of all the distributed systems complications, she might have happily fallen into a discussion of the kinds of D.S. problems that occur when you use, say, a file; and how using a ACID-compliant database might solve those problems, and what the tradeoffs might be there... And she might have corrected herself and said "counter" instead of "flag"!
If I were asked this, my first thought would be to use promises/futures. The idea behind them is that you can execute time-consuming operations asynchronously and they will somehow notify you when they've completed, either successfully or unsuccessfully, typically by calling a callback function. So the first step is to spawn five asynchronous tasks and get five promises.
Then I would join the five promises together, creating a unified promise that represents the five separate tasks. In JavaScript I might call Promise.all(); in Java I would use CompletableFuture.allOf().
I would want to make sure to handle both success and failure. The combined promise should succeed if all of the API calls succeed and fail if any of them fail. If any fail there should be appropriate error handling/reporting. What happens if multiple calls fail? How would a mix of successes and failures be reported? These would be design points to mention, though not necessarily solve during the interview.
Promises and futures typically have modular layering system that would allow edge cases like timeouts to be handled by chaining handlers together. If done right, timeouts could become just another error condition that would be naturally handled by the error handling already in place.
This solution would not require any state to be shared across threads, so I would not have to worry about mutexes or deadlocks or other thread synchronization problems.
She said she would use a flag variable to keep track of the number of API calls have returned.
One thing that makes great interviewees stand out is their ability to anticipate follow-up questions and explain details before they are asked. The best answers are fully fleshed out. They demonstrate that one has thought through one's answer in detail, and they have minimal handwaving.
When I read the above I have a slew of follow-up questions:
How will she know when each API call has returned? Is she waiting for a function call to return, a callback to be called, an event to be fired, or a promise to complete?
How is she causing all of the API calls to be executed concurrently? Is there multithreading, a fork-join pool, multiprocessing, or asynchronous execution?
Flag variables are booleans. Is she really using a flag, or does she mean a counter?
What is the variable tracking and what code is updating it?
What is monitoring the variable, what condition is it checking, and what's it doing when the condition is reached?
If using multithreading, how is she handling synchronization?
How will she handle edge cases such API calls failing, or timing out?
A flag variable might lead to a workable solution or it might lead nowhere. The only way an interviewer will know which it is is if she thinks about and proactively discusses these various questions. Otherwise, the interviewer will have to pepper her with follow-up questions, and will likely lower their evaluation of her.
When I interview people, my mental grades are something like:
S — Solution works and they addressed all issues without prompting.
A — Solution works, follow-up questions answered satisfactorily.
B — Solution works, explained well, but there's a better solution that more experienced devs would find.
C — What they said is okay, but their depth of knowledge is lacking.
F — Their answer is flat out incorrect, or getting them to explain their answer was like pulling teeth.
I already know when calling an asynchronous method e.g. myAsync(), the caller e.g. Caller() can continue executing without waiting for it to be finished. But on the other hand, myAsync() also is executing.
public void Caller(){
myAsync();---------------------running
dosometing();------------------running
}
The code next to myAsync() in Caller() will execute with myAsync() at the same time. So could this situation be considered as a kind of concurrency?
update
I prefer use javascript and c#
That very much depends on the concurrency model of your programming language.
If your language allows you to define methods that are "implicitly" running in parallel; then of course, calling myAsync() would use some kind of "concurrency mechanism" (for example a thread) to do whatever that method is supposed to do.
In that sense, the answer to your question is yes. But it might be important to point out that many "common" programming languages (such as Java) would only "work" in such a context when myAsync() would be creating some thread to then run "something" using that thread.
I could not find detailed documentation about the #async macro. From the docs about parallelism I understand that there is only one system thread used inside a Julia process and there is explicit task switching going on by the help of the yieldto function - correct me if I am wrong about this.
For me it is difficult to understand when exactly these task switches happen just by looking at the code, and knowing when it happens seems crucial.
As I understand a yieldto somewhere in the code (or in some function called by the code) needs to be there to ensure that the system is not stuck with only one task.
For example when there is a read operation, inside the read there probably is a wait call and in the implementation of wait there probably is a yieldto call. I thought that without the yieldto call the code would stuck in one task; however running the following example seems to prove this hypothesis wrong.
#async begin # Task A
while true
println("A")
end
end
while true # Task B
println("B")
end
This code produces the following output
BA
BA
BA
...
It is very unclear to me where the task switching happens inside the task created by the #async macro in the code above.
How can I tell about looking at some code the points where task switching happens?
The task switch happens inside the call to println("A"), which at some point calls write(STDOUT, "A".data). Because isa(STDOUT, Base.AsyncStream) and there is no method that is more specialized, this resolves to:
write{T}(s::AsyncStream,a::Array{T}) at stream.jl:782
If you look at this method, you will notice that it calls stream_wait(ct) on the current task ct, which in turn calls wait().
(Also note that println is not atomic, because there is a potential wait between writing the arguments and the newline.)
You could of course determine when stuff like that happens by looking at all the code involved. But I don't see why you would need to know this exactly, because, when working with parallelism, you should not depend on processes not switching context anyway. If you depend on a certain execution order, synchronize explicitly.
(You already kind of noted this in your question, but let me restate it here: As a rule of thumb, when using green threads, you can expect potential context switches when doing IO, because blocking for IO is a textbook example of why green threads are useful in the first place.)
I've consolidated many of the useful answers and came up with my own answer below
For example, I am writing a an API Foo which needs explicit initialization and termination. (Should be language agnostic but I'm using C++ here)
class Foo
{
public:
static void InitLibrary(int someMagicInputRequiredAtRuntime);
static void TermLibrary(int someOtherInput);
};
Apparently, our library doesn't care about multi-threading, reentrancy or whatnot. Let's suppose our Init function should only be called once, calling it again with any other input would wreak havoc.
What's the best way to communicate this to my caller? I can think of two ways:
Inside InitLibrary, I assert some static variable which will blame my caller for init'ing twice.
Inside InitLibrary, I check some static variable and silently aborts if my lib has already been initialized.
Method #1 obviously is explicit, while method #2 makes it more user friendly. I am thinking that method #2 probably has the disadvantage that my caller wouldn't be aware of the fact that InitLibrary shouln't be called twice.
What would be the pros/cons of each approach? Is there a cleverer way to subvert all these?
Edit
I know that the example here is very contrived. As #daemon pointed out, I should initialized myself and not bother the caller. Practically however, there are places where I need more information to properly initialize myself (note the use of my variable name someMagicInputRequiredAtRuntime). This is not restricted to initialization/termination but other instances where the dilemma exists whether I should choose to be quote-and-quote "fault tolorent" or fail lousily.
I would definitely go for approach 1, along with an easy-to-understand exception and good documentation that explains why this fails. This will force the caller to be aware that this can happen, and the calling class can easily wrap the call in a try-catch statement if needed.
Failing silently, on the other hand, will lead your users to believe that the second call was successful (no error message, no exception) and thus they will expect that the new values are set. So when they try to do something else with Foo, they don't get the expected results. And it's darn near impossible to figure out why if they don't have access to your source code.
Serenity Prayer (modified for interfaces)
SA, grant me the assertions
to accept the things devs cannot change
the code to except the things they can,
and the conditionals to detect the difference
If the fault is in the environment, then you should try and make your code deal with it. If it is something that the developer can prevent by fixing their code, it should generate an exception.
A good approach would be to have a factory that creates an intialized library object (this would require you to wrap your library in a class). Multiple create-calls to the factory would create different objects. This way, the initialize-method would then not be a part of the public interface of the library, and the factory would manage initialization.
If there can be only one instance of the library active, make the factory check for existing instances. This would effectively make your library-object a singleton.
I would suggest that you should flag an exception if your routine cannot achieve the expected post-condition. If someone calls your init routine twice, and the system state after calling it the second time will be the same would be the same as if it had just been called once, then it is probably not necessary to throw an exception. If the system state after the second call would not match the caller's expectation, then an exception should be thrown.
In general, I think it's more helpful to think in terms of state than in terms of action. To use an analogy, an attempt to open as "write new" a file that is already open should either fail or result in a close-erase-reopen. It should not simply perform a no-op, since the program will be expecting to be writing into an empty file whose creation time matches the current time. On the other hand, trying to close a file that's already closed should generally not be considered an error, because the desire is that the file be closed.
BTW, it's often helpful to have available a "Try" version of a method that might throw an exception. It would be nice, for example, to have a Control.TryBeginInvoke available for things like update routines (if a thread-safe control property changes, the property handler would like the control to be updated if it still exists, but won't really mind if the control gets disposed; it's a little irksome not being able to avoid a first-chance exception if a control gets closed when its property is being updated).
Have a private static counter variable in your class. If it is 0 then do the logic in Init and increment the counter, If it is more than 0 then simply increment the counter. In Term do the opposite, decrement until it is 0 then do the logic.
Another way is to use a Singleton pattern, here is a sample in C++.
I guess one way to subvert this dilemma is to fulfill both camps. Ruby has the -w warning switch, it is custom for gcc users to -Wall or even -Weffc++ and Perl has taint mode. By default, these "just work," but the more careful programmer can turn on these strict settings themselves.
One example against the "always complain the slightest error" approach is HTML. Imagine how frustrated the world would be if all browsers would bark at any CSS hacks (such as drawing elements at negative coordinates).
After considering many excellent answers, I've come to this conclusion for myself: When someone sits down, my API should ideally "just work." Of course, for anyone to be involved in any domain, he needs to work at one or two level of abstractions lower than the problem he is trying to solve, which means my user must learn about my internals sooner or later. If he uses my API for long enough, he will begin to stretch the limits and too much efforts to "hide" or "encapsulate" the inner workings will only become nuisance.
I guess fault tolerance is most of the time a good thing, it's just that it's difficult to get right when the API user is stretching corner cases. I could say the best of both worlds is to provide some kind of "strict mode" so that when things don't "just work," the user can easily dissect the problem.
Of course, doing this is a lot of extra work, so I may be just talking ideals here. Practically it all comes down to the specific case and the programmer's decision.
If your language doesn't allow this error to surface statically, chances are good the error will surface only at runtime. Depending on the use of your library, this means the error won't surface until much later in development. Possibly only when shipped (again, depends on alot).
If there's no danger in silently eating an error (which isn't a real error anyway, since you catch it before anything dangerous happens), then I'd say you should silently eat it. This makes it more user friendly.
If however someMagicInputRequiredAtRuntime varies from calling to calling, I'd raise the error whenever possible, or presumably the library will not function as expected ("I init'ed the lib with value 42, but it's behaving as if I initted with 11!?").
If this Library is a static class, (a library type with no state), why not put the call to Init in the type initializer? If it is an instantiatable type, then put the call in the constructor, or in the factory method that handles instantiation.
Don;t allow public access to the Init function at all.
I think your interface is a bit too technical. No programmer want to learn what concept you have used while designing the API. Programmers want solutions for their actual problems and don't want to learn how to use an API. Nobody wants to init your API, that is something that the API should handle in the background as far as possible. Find a good abstraction that shields the developer from as much low-level technical stuff as possible. That implies, that the API should be fault tolerant.
I have some high performance file transfer code which I wrote in C# using the Async Programming Model (APM) idiom (eg, BeginRead/EndRead). This code reads a file from a local disk and writes it to a socket.
For best performance on modern hardware, it's important to keep more than one outstanding I/O operation in flight whenever possible. Thus, I post several BeginRead operations on the file, then when one completes, I call a BeginSend on the socket, and when that completes I do another BeginRead on the file. The details are a bit more complicated than that but at the high level that's the idea.
I've got the APM-based code working, but it's very hard to follow and probably has subtle concurrency bugs. I'd love to use TPL for this instead. I figured Task.Factory.FromAsync would just about do it, but there's a catch.
All of the I/O samples I've seen (most particularly the StreamExtensions class in the Parallel Extensions Extras) assume one read followed by one write. This won't perform the way I need.
I can't use something simple like Parallel.ForEach or the Extras extension Task.Factory.Iterate because the async I/O tasks don't spend much time on a worker thread, so Parallel just starts up another task, resulting in potentially dozens or hundreds of pending I/O operations; way too much! You can work around that by Waiting on your tasks, but that causes creation of an event handle (a kernel object), and a blocking wait on a task wait handle, which ties up a worker thread. My APM-based implementation avoids both of those things.
I've been playing around with different ways to keep multiple read/write operations in flight, and I've managed to do so using continuations that call a method that creates another task, but it feels awkward, and definitely doesn't feel like idiomatic TPL.
Has anyone else grappled with an issue like this with the TPL? Any suggestions?
If you're worried about too many threads, you can just set ParallelOptions.MaxDegreeOfParallelism to an acceptable number in your call to Parallel.ForEach.