Problem 7 of chapter 8

72 views
Skip to first unread message

Hongfei Yan

unread,
May 11, 2010, 4:54:56 AM5/11/10
to cs501pku
8:Q: May distributed algorithms require the use of a coordinating process. To what extent can such algorithms actually be considered distributed? Discuss

我的理解是集中式算法coordinator是固定的。分布式算法也有coordinator,但是运行中选择出来的。所以有coordinator不影响分布.

--
您收到此邮件是因为您订阅了 Google 网上论坛的“cs501pku”论坛。
要向此网上论坛发帖,请发送电子邮件至 cs50...@googlegroups.com
要取消订阅此网上论坛,请发送电子邮件至 cs501pku+u...@googlegroups.com
若有更多问题,请通过 http://groups.google.com/group/cs501pku?hl=zh-CN 访问此网上论坛。

suzhiyang

unread,
May 11, 2010, 7:41:58 AM5/11/10
to cs501pku
这个是problem 8 of chapter 6吧:-)
 
problem 4 of chapter 6:
4. Q: When a node synchronizes its clock to that of another node, it is 
generally a good idea to take previous measurements into account as well. 
Why? Also, give an example of how such past readings could be taken into 
account. 
A: The obvious reason is that there may be an error in the current reading. 
Assuming that clocks need only be gradually adjusted, one possibility is to 
consider the last N values and compute a median or average. If the measured 
value falls outside a current interval, it is not taken into account (but is 
added to the list). Likewise, a new value can be computed by taking a 
weighted average, or an aging algorithm.
 
这里面的意思是不是说本次的同步可能是错误的,所以需要考虑前N次的同步结果,在同步之后先进行判断是否误差较大。若误差太大,则本次结果不予考虑,但加到一个特殊的错误LIST??若在正常范围内,则取加权平均或其他什么老化算法进行计算即可。
 
 
2010-05-11

suzhiyang

发件人: Hongfei Yan
发送时间: 2010-05-11  16:55:09
收件人: cs501pku
抄送:
主题: Problem 7 of chapter 8

Hongfei Yan

unread,
May 11, 2010, 8:24:58 AM5/11/10
to cs501pku
我理解是加入last N列表,不是特殊的错误列表。

2010/5/11 suzhiyang <suzh...@gmail.com>
Reply all
Reply to author
Forward
0 new messages