《Cracking the Coding Interview》——第10章：可扩展性和存储空间限制—

本文探讨了如何设计一个能够为最多1000个客户端提供实时股票价格查询的服务，涉及使用SQL数据库、K-V数据库或XML文件进行数据存储，并讨论了每种方式的优缺点。

2014-04-24 00:38

题目：加入你要设计一个服务，供至多1000个client查询每日股票的最高、最低、开盘、收盘价，请描述你要如何设计这个服务。

解法：可以用SQL数据库组织，也可以用K-V数据库来提供更高性能。如果股票数量相对不大，还可以用XML文件组织数据。在百度里搜“BIDU”，或在Google里搜“GOOG”，就能看见例子了。

代码：

 1 // 10.1 Suppose you're to design the backend part of a system, which provides queries of real-time stock prices. How would you do it?
 2 // Answer:
 3 //    Go and search "GOOG" on Google or Baidu, you'll get the real-time stock price of Google. The data transferring is best done with AJAX to the frontend, which can greatly save the amount of data transferred.
 4 //    As for the backend part, you have to use some techniques to store the data:
 5 //        1. SQL database, easy for various kinds of structured queries. You won't expose the direct SQL API to external users because it is not safe. Write a wrapper with the web programming language.
 6 //            Strength: Ability to provide more information than you expected.
 7 //            Weakness: Not so agile, if the data structure changes. Altering a SQL table is always expensive.
 8 //        2. XML files, a good way to store structured data. Good performance when the data scale is small or medium.
 9 //            Strength: Agile, easy to modify the schema. Regenerating the XML file is usually easy with script languages such as python.
10 //            Weakness: Not suitable for large data scale, as you have to parse everything in the XML only to find an item inside.
11 //        3. No-SQL database
12 //            Strength: suitable for large scale, high concurrency, flexible with data schema changes.
13 //            Weakness: Don't know yet, maybe more space usage.
14 int main()
15 {
16     return 0;
17 }