Interprocess Communications프로세스간 통신방법들

12 minute read

The Windows operating system provides mechanisms for facilitating communications and data sharing between applications. Collectively, the activities enabled by these mechanisms are called interprocess communications (IPC). Some forms of IPC facilitate the division of labor among several specialized processes. Other forms of IPC facilitate the division of labor among computers on a network.

윈도우는 어플리케이션간의 쉬운 커뮤니케이션과 데이터 공유를 위한 매커니즘을 제공한다. 이런 매커니즘을 통틀어서 IPC(interprocess communications)라고 부른다. 몇몇 IPC는 프로세스의 분업을 도와준다. 다른 IPC들은 네트워크의 컴퓨터 사이의 분업을 도와준다.

Typically, applications can use IPC categorized as clients or servers. A client is an application or a process that requests a service from some other application or process. A server is an application or a process that responds to a client request. Many applications act as both a client and a server, depending on the situation. For example, a word processing application might act as a client in requesting a summary table of manufacturing costs from a spreadsheet application acting as a server. The spreadsheet application, in turn, might act as a client in requesting the latest inventory levels from an automated inventory control application.

보통 IPC를 쓰는 어플리케이션은 client와 server로 분류된다. client는 서비스를 요청한다. server는 client의 요청에 응답한다.  대부분의 어플리케이션은 상황에 따라 client가 되기도 하고 server가 되기도 한다. 예를들어, 워드프로세서가 스프레드시트에 생산원가에 대한 표를 요청할 때 워드프로세서는 client가 되고 스트레드시트는 server가 된다. 스트레드시트가 재고관리 어플리케이션에 마지막 재고 상황을 요청할 때는 client가 된다.

After you decide that your application would benefit from IPC, you must decide which of the available IPC methods to use. It is likely that an application will use several IPC mechanisms. The answers to these questions determine whether an application can benefit by using one or more IPC mechanisms.

IPC를 쓰기로 결정했다면… 어떤 IPC방식을 사용할지도 결정해야한다. 어플리케이션은 보통 몇 가지 IPC 메커니즘을 사용한다. 다음 질문의 답변을 통해 IPC메커니즘을 사용하는 것이 어떤 잇점이 있는지 알 수 있을 것이다.

  • Should the application be able to communicate with other applications running on other computers on a network, or is it sufficient for the application to communicate only with applications on the local computer?
  • Should the application be able to communicate with applications running on other computers that may be running under different operating systems (such as 16-bit Windows or UNIX)?
  • Should the user of the application have to choose the other applications with which the application communicates, or can the application implicitly find its cooperating partners?
  • Should the application communicate with many different applications in a general way, such as allowing cut-and-paste operations with any other application, or should its communications requirements be limited to a restricted set of interactions with specific other applications?
  • Is performance a critical aspect of the application? All IPC mechanisms include some amount of overhead. 성능(속도?)이 중요한가? 모든 IPC 메커니즘은 오버헤드가 필요하다.
  • Should the application be a GUI application or a console application? Some IPC mechanisms require a GUI application. 어플리케이션이 GUI인가 Console인가? 몇몇 IPC는 GUI환경이 필요하다.

The following IPC mechanisms are supported by Windows:

윈도우에서 아래의 IPC메커니즘이 제공된다.

Using the Clipboard for IPC

The clipboard acts as a central depository for data sharing among applications. When a user performs a cut or copy operation in an application, the application puts the selected data on the clipboard in one or more standard or application-defined formats. Any other application can then retrieve the data from the clipboard, choosing from the available formats that it understands. The clipboard is a very loosely coupled exchange medium, where applications need only agree on the data format. The applications can reside on the same computer or on different computers on a network.

Key Point: All applications should support the clipboard for those data formats that they understand. For example, a text editor or word processor should at least be able to produce and accept clipboard data in pure text format. For more information, see Clipboard.

Using COM for IPC

Applications that use OLE manage compound documents— that is, documents made up of data from a variety of different applications. OLE provides services that make it easy for applications to call on other applications for data editing. For example, a word processor that uses OLE could embed a graph from a spreadsheet. The user could start the spreadsheet automatically from within the word processor by choosing the embedded chart for editing. OLE takes care of starting the spreadsheet and presenting the graph for editing. When the user quit the spreadsheet, the graph would be updated in the original word processor document. The spreadsheet appears to be an extension of the word processor.

The foundation of OLE is the Component Object Model (COM). A software component that uses COM can communicate with a wide variety of other components, even those that have not yet been written. The components interact as objects and clients. Distributed COM extends the COM programming model so that it works across a network.

Key Point: OLE supports compound documents and enables an application to include embedded or linked data that, when chosen, automatically starts another application for data editing. This enables the application to be extended by any other application that uses OLE. COM objects provide access to an object’s data through one or more sets of related functions, known as interfaces. For more information, see COM and ActiveX Object Services.

Using Data Copy for IPC

Data copy enables an application to send information to another application using the WM_COPYDATA message. This method requires cooperation between the sending application and the receiving application. The receiving application must know the format of the information and be able to identify the sender. The sending application cannot modify the memory referenced by any pointers.

WM_COPYDATA메세지를 통해 데이터를 전송한다. 이 방법을 쓰려면 두 어플리케이션이 같이 동작해야된다. 두 어플리케이션은 정해진 format을 사용해야한다. 보내는 어플리케이션은 메모리 레퍼런스나 포인터를 건드릴 수 없다.

Key Point: Data copy can be used to quickly send information to another application using Windows messaging. For more information, see Data Copy.

Data copy는 윈도우즈 메세지를 통해 빠르게 데이터를 전송할 수 있다.

Using DDE for IPC

DDE is a protocol that enables applications to exchange data in a variety of formats. Applications can use DDE for one-time data exchanges or for ongoing exchanges in which the applications update one another as new data becomes available.

The data formats used by DDE are the same as those used by the clipboard. DDE can be thought of as an extension of the clipboard mechanism. The clipboard is almost always used for a one-time response to a user command, such as choosing the Paste command from a menu. DDE is also usually initiated by a user command, but it often continues to function without further user interaction. You can also define custom DDE data formats for special-purpose IPC between applications with more tightly coupled communications requirements.

DDE exchanges can occur between applications running on the same computer or on different computers on a network.

Key Point: DDE is not as efficient as newer technologies. However, you can still use DDE if other IPC mechanisms are not suitable or if you must interface with an existing application that only supports DDE. For more information, see Dynamic Data Exchange and Dynamic Data Exchange Management Library.

Using a File Mapping for IPC

File mapping enables a process to treat the contents of a file as if they were a block of memory in the process’s address space. The process can use simple pointer operations to examine and modify the contents of the file. When two or more processes access the same file mapping, each process receives a pointer to memory in its own address space that it can use to read or modify the contents of the file. The processes must use a synchronization object, such as a semaphore, to prevent data corruption in a multitasking environment.

You can use a special case of file mapping to provide named shared memory between processes. If you specify the system swapping file when creating a file-mapping object, the file-mapping object is treated as a shared memory block. Other processes can access the same block of memory by opening the same file-mapping object.

File mapping is quite efficient and also provides operating-system–supported security attributes that can help prevent unauthorized data corruption. File mapping can be used only between processes on a local computer; it cannot be used over a network.

Key Point: File mapping is an efficient way for two or more processes on the same computer to share data, but you must provide synchronization between the processes. For more information, see File Mapping and Synchronization.

Using a Mailslot for IPC

Mailslots provide one-way communication. Any process that creates a mailslot is a mailslot server. Other processes, called mailslot clients, send messages to the mailslot server by writing a message to its mailslot. Incoming messages are always appended to the mailslot. The mailslot saves the messages until the mailslot server has read them. A process can be both a mailslot server and a mailslot client, so two-way communication is possible using multiple mailslots.

A mailslot client can send a message to a mailslot on its local computer, to a mailslot on another computer, or to all mailslots with the same name on all computers in a specified network domain. Messages broadcast to all mailslots on a domain can be no longer than 400 bytes, whereas messages sent to a single mailslot are limited only by the maximum message size specified by the mailslot server when it created the mailslot.

Key Point: Mailslots offer an easy way for applications to send and receive short messages. They also provide the ability to broadcast messages across all computers in a network domain. For more information, see Mailslots.

Using Pipes for IPC

There are two types of pipes for two-way communication: anonymous pipes and named pipes. Anonymous pipes enable related processes to transfer information to each other. Typically, an anonymous pipe is used for redirecting the standard input or output of a child process so that it can exchange data with its parent process. To exchange data in both directions (duplex operation), you must create two anonymous pipes. The parent process writes data to one pipe using its write handle, while the child process reads the data from that pipe using its read handle. Similarly, the child process writes data to the other pipe and the parent process reads from it. Anonymous pipes cannot be used over a network, nor can they be used between unrelated processes.

두 가지 타입의 pipe의 2웨이 통신 : anonymous pipe와 named pipe. Anonymous pipe는 관련 프로세스가 서로 정보를 주고받을 수 있다. 보통 anonymous pipe는 표준 input output을 child 프로세스에 redirecting 한다??.

Named pipes are used to transfer data between processes that are not related processes and between processes on different computers. Typically, a named-pipe server process creates a named pipe with a well-known name or a name that is to be communicated to its clients. A named-pipe client process that knows the name of the pipe can open its other end, subject to access restrictions specified by named-pipe server process. After both the server and client have connected to the pipe, they can exchange data by performing read and write operations on the pipe.


Key Point: Anonymous pipes provide an efficient way to redirect standard input or output to child processes on the same computer. Named pipes provide a simple programming interface for transferring data between two processes, whether they reside on the same computer or over a network. For more information, see Pipes.

Anonymous pipe는 같은 컴퓨터에서 실행되는 child 프로세스에 효율적으로 표준 input output을 redirect 해 준다. Named pipe는 같은 컴퓨터 또는 같은 네트워크 안의 두 프로세스간에 간편하게 데이터 전달 통로를 제공 해 준다.

Using RPC for IPC

RPC enables applications to call functions remotely. Therefore, RPC makes IPC as easy as calling a function. RPC operates between processes on a single computer or on different computers on a network.

The RPC provided by Windows is compliant with the Open Software Foundation (OSF) Distributed Computing Environment (DCE). This means that applications that use RPC are able to communicate with applications running with other operating systems that support DCE. RPC automatically supports data conversion to account for different hardware architectures and for byte-ordering between dissimilar environments.

RPC clients and servers are tightly coupled but still maintain high performance. The system makes extensive use of RPC to facilitate a client/server relationship between different parts of the operating system.

Key Point: RPC is a function-level interface, with support for automatic data conversion and for communications with other operating systems. Using RPC, you can create high-performance, tightly coupled distributed applications. For more information, see Microsoft RPC Components.

Using Windows Sockets for IPC

Windows Sockets is a protocol-independent interface. It takes advantage of the communication capabilities of the underlying protocols. In Windows Sockets 2, a socket handle can optionally be used as a file handle with the standard file I/O functions.

Windows Sockets are based on the sockets first popularized by Berkeley Software Distribution (BSD). An application that uses Windows Sockets can communicate with other socket implementation on other types of systems. However, not all transport service providers support all available options.

Key Point: Windows Sockets is a protocol-independent interface capable of supporting current and emerging networking capabilities. For more information, see Windows Sockets 2.



Send comments about this topic to Microsoft

Build date: 1/27/2011

Community Content Add

[Annotations]( en-us VS.85 “Annotations”) FAQ

Sockets for IPC

A negative for using sockets for IPC may be having to assign a port for the local host server. Is this a configuration hassle to find a unused port and make sure the firewall does not block the port for someone just wanting to create a service used by local clients. Named pipes and COM are probably the best for local IPC. Anyone at Microsoft or the developer community care to comment???

[SamB 2011-03-21]

The firewall does not appear to actually block local (loopback) traffic, at least not on XP. I have, however, seen the dialog appear when an application started listening on a port for this purpose, even though (if I remember correctly) the application appeared to be listening specifically on (except it probably wasn’t actually listening on port 1234).


Maximum Throughput for IPC?

Hello. Is there a maximum throughput when using IPC? We are using a 32-bit programming interface which could greatly benefit from 64-bit addressing space. Our thought is to create a separate 32-bit and 64-bit application, and use IPC to communicate between the two. The only potential pitfall is that we need to stream up to 480 MB of data per second. Are there any benchmarks on the maximum throughput of IPC? Here is our system: Windows 7 x64, 8 GB RAM, 2.3 GHz dual-core AMD Turion 64 X2 processor (upgrading to Intel i7 architecture soon), data is coming from x4 PCIe link.

[tfl - 15 02 10] Hi - and thanks for your post. You should post questions like this to the MSDN Forums at or the MSDN Newsgroups at You are much more likely get a quicker response using the forums than through the Community Content. For specific help about: .NET Framework : PowerShell : SQL Server : Visual Studio : Windows : All Public :