公开(公告)号：EP2206294A1

公开(公告)日：2010-07-14

申请号：EP08804222.1

申请日：2008-09-15

Applicant: Thomson Licensing

Inventor： SIEMENS, EDUARD ,  QIU, XIAOPENG

IPC: H04L12/56 ,  H04L1/18 ,  G06F5/12

CPC classification number: H04L1/1874 ,  G06F5/12 ,  G06F2205/062 ,  H04L1/1835 ,  H04L47/6245 ,  H04L49/90 ,  H04L49/901 ,  H04L49/9031 ,  H04L49/9047

Abstract: This invention relates to a design of an efficient buffer management model in order to increase the efficiency of data exchange between two process threads, - e.g. when implementing a network transport protocol stack. This invention proposes to use an interconnected system of different kinds of memory buffers (100, 101, 102), implemented as asynchronous read/write ring buffers ARWRB. These buffers are organized in a way, in which data can be stored into the buffer or fetched from the buffer essentially avoiding synchronization means like mutexes or semaphores. In contrast to the conventional buffer management model, three ring buffers, namely send ring buffer (100), send token ring buffer (101) and receive ring buffer (102), are used within the transport protocol stack.

Abstract translation: 本发明涉及一种高效的缓存器管理模型的设计，以增加两个处理线程之间的数据交换的效率 - E. G. 当实现一个网络传输协议栈。 本发明提出了作为异步读/写环形缓冲器ARWRB在不同种类的存储器缓冲器（100，101，102），来实现的互连的系统中使用。 这些缓冲器的方式，在该数据可被存储到从基本上避免同步缓存器中的缓冲或取出组织等手段或互斥信号量。 与此相反的常规缓冲器管理模式，三个环形缓冲器，即发送环形缓冲器（100），发送令牌环缓冲器（101）和接收环形缓冲区（102）的传送协议栈内被使用。