2014-05-06 00:45
原题:
What would happen if you have only one server for a web cache (a web browser cache whose key is url and value is the loaded content of the webpage) but huge numbers of clients? And how would you solve it? Assume the cache is implemented with a hashmap and a linkedlist.
题目:你建了个网站,只有一台缓存服务器,服务器cache用的是拉链式的哈希表。如果客户端的并发量很大,怎么办呢?
解法:既然并发量很大,那就LRU cache吧。不过这个也治标不治本,因为大并发量的时候,cache会被不断刷新,等于没有。如果静态内容很多的话,浏览器端缓存能够减轻一部分服务器端缓存的压力。最简单直接的解决办法,就是掏钱给集群扩容了,多台机器好办事。
代码:
1 // http://www.careercup.com/question?id=4807591515389952 2 // What would happen if you have only one server for a web cache (a web browser cache whose key is url and value is the loaded content of the webpage) but huge numbers of clients? And how would you solve it? Assume the cache is implemented with a hashmap and a linkedlist. 3 // Answer: Well, I guess the cache can be considered a hash table, using chaining to handle collisions. LRU cache will be better. Despite the cache, with huge number of requests, the cache will be constantly flushed. It would still cause the server to crash. If the system is not so "dynamic", perhaps browser cache will help to reduce a little stress for the server-side cache. 4 int main() 5 { 6 return 0; 7 }