|
Distributed operating systemsWhen the communication and processors are all perfect, reaching such agreement is often straightforward, but when they are not, problems arise. In this section we will look at some of the problems and their solutions (or lack thereof). The general goal of distributed agreement algorithms is to have all the non-faulty processors reach consensus on some issue, and do that within a finite number of steps. Different cases are possible depending on system parameters, including: 1. Are messages delivered reliably all the time? 2. Can processes crash, and if so, fail-silent or Byzantine? 3. Is the system synchronous or asynchronous? Before considering the case of faulty processors, let us look at the "easy" case of perfect processors but communication lines that can lose messages. There is a famous problem, known as the two-army problem, which illustrates the difficulty of getting even two perfect processors to reach agreement about 1 bit of information. The red army, with 5000 troops, is encamped in a valley. Two blue armies, each 3000 strong, are encamped on the surrounding hillsides overlooking the valley ...» |
Код для вставки книги в блог HTML
phpBB
текст
|
|