This would include the collector , which collects resource information , and the negotiator , which checks the collector for resources and queries the job queue of the scheduler daemons to match the resource requests with available resources 这包括搜集器和协商器，前者搜集资源的信息，后者检查搜集器的资源，并查询调度器守护进程的任务队列对资源请求和可用资源进行匹配。
If each grid resource is sent the same number of jobs , a long job queue could build up in some slower machines , or a long - running job could block others from starting if not carefully monitored 如果每一个网格资源上都发送到相同数目的任务，那么在速度较慢的机器上会形成较长的任务队列，而一个长时间运行的任务在不受到细心监视的情况下可能阻塞其他的任务，使之根本无法启动。
Based on the above schemes , this paper proposes the design of embedded realtime linux os used in terabit router , which combines low - latency scheme with preemption scheme to solve the non - preemptive kernel and replace the single running queue with multiple running queues in order to reduce the delay of scheduling . finally , it puts the schemes into reality and tests the performance 在对上述方案分析的基础上，结合t比特路由器对嵌入式实时操作系统的特殊要求，提出了t比特路由器嵌入式实时linux操作系统设计方案? ?低延迟与抢占方案相结合解决内核的非抢占机制，并且以多可运行任务队列来代替原来的单可运行任务队列的方法降低调度延迟。
Lastly , in order to improve the overall performance of the application server , a task - queue - based and multithread - manageable rmi framework is presented in this thesis . the rmi logic is separated from the application logic . the result is that thread overload problem in java rmi is solved and the timed request can be produced 最后，为进一步提高web应用服务器的系统性能，我们提出基于任务队列可控多线程的rmi服务器，把rmi连接活动从应用程序处理中分离出来，解决了javarmi的不能计时和线程过载等问题。
Make clear the procedure of sending and receiving packet of kernel and the implementation of traffic control . the special mechanism in linux such as wait queue , task queue , time interrupt is also explained . after analyzing the file system of linux at large , we bring forward the design of modifying of file system . since many little file is created during the service of email and file service , we introduce the non volatile random access ram , by modifying the arithmetic of file system , changing the data flow of file system 本文详细介绍了针对网络专用服务器的专用linux系统的设计，对linux文件系统进行了详尽的解析，包括vfs的实现， buffercache的作用等；并对块设备驱动程序在内核中各个接口进行了分析；解析了内核接收和发送数据包的全过程以及流量控制在内核中的全过程；并对linux系统的一些特殊的机制如等待队列，任务队列，时钟中断等进行了详细的解释。
